ATP

00:00:00 ◼   ► Marco, I have to ask you, what is the temperature in the building that you're in right now?

00:00:05 ◼   ► Because all I know is I see in the pre-show, in our internal show notes, Marco's HVAC update, and I'm scared for you.

00:00:13 ◼   ► Oh, no.

00:00:15 ◼   ► So I mentioned my HVAC or HVAC, I never know which one it is, I've heard both.

00:00:21 ◼   ► You can say both.

00:00:22 ◼   ► Problems in past shows, because basically all of last summer, I didn't have reliable air conditioning at the beach.

00:00:30 ◼   ► That is also our heat all winter long in that house.

00:00:32 ◼   ► We don't drain the water because we go there all the time in the winter, so it has to be kept from freezing.

00:00:39 ◼   ► So my solution last winter, when it first broke last February, like a year ago February, my solution was I'm just going to bring a bunch of space heaters in and just run them on low all the time.

00:00:51 ◼   ► I had like the nice safe oil radiator kind.

00:00:55 ◼   ► So it's not like a safety risk.

00:00:57 ◼   ► It's certainly not an efficient use of electricity or money.

00:01:00 ◼   ► Well, you got the solar panels.

00:01:01 ◼   ► Doesn't that help?

00:01:02 ◼   ► It helps, but I mean, it's a lot of power.

00:01:04 ◼   ► Like that's like it offsets it, but it doesn't cover the whole thing by any means.

00:01:08 ◼   ► Especially if you have snow covering the panels.

00:01:10 ◼   ► Yeah, exactly.

00:01:11 ◼   ► So over the course of the last few months, we occasionally get listeners asking, hey, whatever happened to that project?

00:01:18 ◼   ► Because I believe the last update I gave you was probably in the fall.

00:01:23 ◼   ► Basically all spring, summer, and fall, I had a parade of different service people come out and try to figure out what the heck was wrong with my systems.

00:01:30 ◼   ► What ended up ended up happening was, in short, nothing good.

00:01:35 ◼   ► The service person who spent the most time with it, and these are people, they were like calling tech support from the manufacturer and trying to decode all these codes.

00:01:45 ◼   ► And they couldn't, they're hooking up all sorts of diagnostic equipment and nobody could figure out what the problem was.

00:01:49 ◼   ► Eventually, this person recommended a major kind of servicing, reinstallation, like disassembling a bunch of things, cleaning them out, testing all the lines again, doing the vacuum, all this, like a ton of work.

00:02:03 ◼   ► And listeners, I don't know how much a new Honda Accord costs today, but I think I probably could have bought one for the amount of money that I had to spend on this.

00:02:15 ◼   ► And then at the end of it, he said, I still can't get it working right.

00:02:21 ◼   ► I recommend full replacement.

00:02:23 ◼   ► And I'm like, you know, gobsmacked, but I'm like, well, I mean, what else can I do?

00:02:30 ◼   ► And I called, I called like some of the previous people and I, and I kind of ran it by.

00:02:34 ◼   ► I'm like, Hey, I have this guy telling me this.

00:02:36 ◼   ► Does that sound right to you?

00:02:37 ◼   ► Cause you've also seen the system.

00:02:39 ◼   ► And they also said, I hate to say it, but yeah.

00:02:42 ◼   ► So I'm like, crap.

00:02:44 ◼   ► So I told the guy, all right, fine.

00:02:51 ◼   ► Make me a proposal.

00:02:52 ◼   ► Let's, let's see what you got for full replacement.

00:02:55 ◼   ► And then he goes to me completely.

00:02:58 ◼   ► Oh, that was October.

00:03:01 ◼   ► And he, and I called back a couple of times and I talked to his secretary.

00:03:04 ◼   ► Oh, he's working on it.

00:03:05 ◼   ► He's working on it.

00:03:05 ◼   ► Ghosted me.

00:03:06 ◼   ► Never called me back.

00:03:08 ◼   ► Never got me a proposal.

00:03:09 ◼   ► I needless to say, I'm not thrilled with having worked with this guy at this point, having given

00:03:17 ◼   ► him a lot of money for a system that was still broken.

00:03:20 ◼   ► But fortunately one of the other service providers I'd work with at the beach, who was very busy

00:03:25 ◼   ► in the summer, it was a lot less busy in the winter.

00:03:28 ◼   ► And so I got him on the job now and he is, he has replaced my upstairs unit.

00:03:35 ◼   ► I still, we're still waiting on the downstairs unit to come in from the manufacturer.

00:03:38 ◼   ► The upstairs one is replaced.

00:03:40 ◼   ► Downstairs is still space heaters.

00:03:42 ◼   ► Which brings us to this past few weeks.

00:03:45 ◼   ► The upstairs unit, again, it works fine.

00:03:48 ◼   ► You know, it's holding temperature.

00:03:49 ◼   ► It's a brand new unit.

00:03:50 ◼   ► And I said, you know what?

00:03:51 ◼   ► I'm tired of all these, get me a simple unit.

00:03:54 ◼   ► So we have a Daikin or Daikin, I don't know how that brand's pronounced.

00:03:57 ◼   ► We have a Daikin system that's just like the most cheap, simple thing because everything

00:04:03 ◼   ► out there is cheap, simple Daikin systems.

00:04:05 ◼   ► And when they break, everyone out there knows how to service them and they're easily serviced

00:04:09 ◼   ► and replaced.

00:04:10 ◼   ► Is this a Northeastern thing?

00:04:11 ◼   ► Because I don't think I've ever seen them around here.

00:04:13 ◼   ► It's all like train and a couple other things around here.

00:04:15 ◼   ► They're very popular for split unit systems.

00:04:18 ◼   ► They're not that common for central systems yet.

00:04:20 ◼   ► This is a central system.

00:04:21 ◼   ► But how do you spell the name of the company?

00:04:23 ◼   ► D-A-I-K-I-N.

00:04:25 ◼   ► As far as I can tell, they seem to have a really good sales network because every HVAC tech out

00:04:31 ◼   ► here wears Daikin shirts.

00:04:33 ◼   ► They have a Daikin sticker on their car.

00:04:35 ◼   ► They have the best parties.

00:04:36 ◼   ► Right.

00:04:37 ◼   ► They're obviously doing a lot of sales outreach.

00:04:39 ◼   ► But, you know, their stuff is everywhere.

00:04:43 ◼   ► Everyone has had to service it.

00:04:44 ◼   ► It's simple.

00:04:44 ◼   ► It doesn't seem like it's super high quality, but it's also not that expensive.

00:04:47 ◼   ► And at the beach, that's kind of what you want.

00:04:49 ◼   ► Because it's such a harsh environment.

00:04:51 ◼   ► Nothing out there lasts very long.

00:04:53 ◼   ► It's a consumable.

00:04:53 ◼   ► It's going to rot out anyway in a few years.

00:04:55 ◼   ► Right.

00:04:56 ◼   ► So if you're only going to get five to ten years out of something like this, it's best to have

00:05:00 ◼   ► it be reasonably serviceable and replaceable.

00:05:03 ◼   ► So anyway, so I've been using space heaters downstairs, new system upstairs.

00:05:07 ◼   ► And then we have this massive cold freeze this winter.

00:05:11 ◼   ► And so I've been relying on my Yolink sensors.

00:05:15 ◼   ► Yes.

00:05:16 ◼   ► Yes.

00:05:17 ◼   ► My house is filled with Yolink sensors.

00:05:19 ◼   ► I can see the temperatures everywhere.

00:05:21 ◼   ► I have one outside so I can correlate what's going on inside to what's going on outside so

00:05:26 ◼   ► I can tell like when the system is working right.

00:05:27 ◼   ► I have water sensors everywhere for water leaks at every, under every sink trap, under every

00:05:34 ◼   ► toilet valve, on a couple of, you know, next to the fridge, like in case that leaks, like

00:05:38 ◼   ► next to the ice maker.

00:05:40 ◼   ► And, you know, a couple of like floor locations, water leak sensors or water leak sensors everywhere.

00:05:45 ◼   ► So I'm going through this freeze and I'm watching my, and we, you know, we have like the house

00:05:49 ◼   ► is lifted because it's the beak.

00:05:52 ◼   ► So it's like, it's on stilts and the water main, you know, there's a challenge when you

00:05:56 ◼   ► have a lifted house with no basement under it.

00:05:57 ◼   ► What do you, how do you insulate the water main and the sewer main that have to come in

00:06:03 ◼   ► and out of the ground to go up to the house?

00:06:05 ◼   ► So we have this like boxed out closet that goes around them and there's a heater inside

00:06:10 ◼   ► that closet that keeps it, you know, 50 degrees or whatever.

00:06:12 ◼   ► And of course, sensor in there.

00:06:15 ◼   ► So I've been monitoring that.

00:06:17 ◼   ► And like during the worst of the freezes, that water main closet got down to like 36 degrees.

00:06:23 ◼   ► I was watching it.

00:06:24 ◼   ► I'm like, oh God.

00:06:25 ◼   ► And meanwhile, the bay is frozen.

00:06:28 ◼   ► There's no ferry service.

00:06:30 ◼   ► So the only, and I don't have a driving permit anymore, so I don't live there anymore.

00:06:33 ◼   ► So the only way I can get there if I need to do something is I can like bum a ride from

00:06:37 ◼   ► someone during the day if they happen to be driving their truck on for, you know, if they're

00:06:41 ◼   ► a contractor or something like, and I know a few people, so I can like, I can ask for rides,

00:06:45 ◼   ► but it's hard to get there.

00:06:47 ◼   ► And it's, it's not fast to get there.

00:06:49 ◼   ► And the answer often, you know, the answer, if I, if I want to go there on Saturday, the

00:06:52 ◼   ► answer is too bad.

00:06:53 ◼   ► You're going on Monday.

00:06:54 ◼   ► Anyway, I'm watching these sensors go down on these like, you know, super freezing Saturday

00:07:00 ◼   ► nights.

00:07:00 ◼   ► I'm like, oh God, getting tense.

00:07:02 ◼   ► And like, I know a couple of people out there that if I really had to, I could call them to

00:07:06 ◼   ► go over there and like change something if I really had to.

00:07:08 ◼   ► But like, you know, you also don't want to call a friend to go over to your house when

00:07:11 ◼   ► it's five degrees with 40 mile an hour winds.

00:07:14 ◼   ► If you really don't have to, like it's ideally, you don't need that.

00:07:20 ◼   ► Anyway, so I made it through all the deep freezes.

00:07:22 ◼   ► I thought yesterday I get a water leaks alert from Yolink in an outdoor utility closet that

00:07:31 ◼   ► houses the, the water heater and like the water filter and a couple of air handlers.

00:07:35 ◼   ► I know there's, there's a water filter attachment there that leaks sometimes.

00:07:39 ◼   ► So an occasional false alarm there is common.

00:07:42 ◼   ► So I'm like, ah, it's probably fine.

00:07:43 ◼   ► And then I got the water leak alert again.

00:07:47 ◼   ► And I'm like, hmm, that's weird.

00:07:51 ◼   ► Let me, let me check.

00:07:52 ◼   ► And I have like, I have one camera that is mounted right below that, like looking out on my front

00:07:57 ◼   ► stoop, like to see when teenagers are, you know, walking up to my stairs.

00:08:00 ◼   ► And peeing.

00:08:01 ◼   ► And peeing.

00:08:02 ◼   ► Yeah.

00:08:02 ◼   ► Well, they don't pee on those.

00:08:03 ◼   ► They pee at the other side.

00:08:04 ◼   ► Oh, my bad.

00:08:05 ◼   ► This is where teenagers get off the beach and use my foot wash.

00:08:08 ◼   ► Oh, lovely.

00:08:09 ◼   ► And just flood my entire front patio with water and sand.

00:08:13 ◼   ► It's delightful.

00:08:14 ◼   ► You're really selling, you know, Fire Island.

00:08:17 ◼   ► Well, it's uninhabitable, uninhabitable in the winter.

00:08:20 ◼   ► Even if you wanted to be there, you can't get there.

00:08:22 ◼   ► People treat private property like their own public utilities.

00:08:26 ◼   ► Like, you're really doing a bang up job of selling me on Fire Island.

00:08:28 ◼   ► That's part of its charm, too.

00:08:30 ◼   ► Yeah.

00:08:30 ◼   ► Uh-huh.

00:08:30 ◼   ► I'm sure.

00:08:31 ◼   ► So anyway, after the couple of water leak alerts, I check my camera that's there.

00:08:36 ◼   ► And it's raining.

00:08:39 ◼   ► I'm like, oh, hmm.

00:08:41 ◼   ► That's not because it's not raining, but it's raining at that camera.

00:08:44 ◼   ► Oh, I misunderstood.

00:08:47 ◼   ► Oh, no.

00:08:47 ◼   ► I was going to guess if your water heater is in there.

00:08:49 ◼   ► That's probably rotted out and it's just failed.

00:08:51 ◼   ► So I called a friend.

00:08:54 ◼   ► I'm like, I can't get there until Friday, really.

00:08:56 ◼   ► So I'm like, all right.

00:08:57 ◼   ► I called a friend.

00:08:57 ◼   ► He went up and opened the closet.

00:08:59 ◼   ► And sure enough, there's some pipe that ruptured.

00:09:01 ◼   ► And this has never been a problem before.

00:09:03 ◼   ► But because the downstairs HVAC system is not running, that closet, which has the air

00:09:11 ◼   ► handlers in it, is not being warmed up by it.

00:09:14 ◼   ► So a pipe froze in there and ruptured.

00:09:17 ◼   ► He turned it off.

00:09:19 ◼   ► I'll know in a couple of days quite how bad it is.

00:09:22 ◼   ► I'm going to send a plumber over there in the morning.

00:09:23 ◼   ► But I have a leak.

00:09:27 ◼   ► It rained all over my front patio all day long.

00:09:31 ◼   ► Thanks to Yolink for alerting me to it.

00:09:33 ◼   ► My next step is, I think I'm going to look into their various options for water main shutoffs.

00:09:39 ◼   ► I really want to get one for our house, just on principle.

00:09:43 ◼   ► But they're surprisingly expensive.

00:09:45 ◼   ► Yeah, they're about $300.

00:09:46 ◼   ► Well, the fanciest ones actually stop the water.

00:09:49 ◼   ► The less fancy ones turn a valve that you already have that stops the water.

00:09:53 ◼   ► Those are still a couple hundred bucks.

00:09:54 ◼   ► That's the $300 one, yeah.

00:09:55 ◼   ► Yeah, exactly.

00:09:56 ◼   ► I'm saying the very fancy ones literally have the valve inside them and will turn off the water.

00:09:59 ◼   ► Yeah, I almost ordered it today, but I wanted to go there and measure and make sure my valve.

00:10:03 ◼   ► Because it says your valve has to move fairly freely because it doesn't have that much pressure it can apply.

00:10:10 ◼   ► That's why you need the, what do you call it, the ball cock valve.

00:10:12 ◼   ► We get a lot of leverage with the big lever.

00:10:13 ◼   ► Yeah, so anyway, thanks, Yolink.

00:10:16 ◼   ► Yolink has been awesome for this.

00:10:19 ◼   ► Because we have the restaurant now, the restaurant has an apartment above it, and so I've also been able to monitor that apartment.

00:10:27 ◼   ► Like, you know, staff stays there in the summertime, but that's yet another property to manage, and I've been able to monitor that and make sure that's been not freezing during these coldest nights.

00:10:36 ◼   ► And so it's been remarkable having Yolink in my life.

00:10:41 ◼   ► So thank you, John, for bringing Yolink to us.

00:10:44 ◼   ► Yes, very much so.

00:10:45 ◼   ► You know, for anybody who missed the whole deal with Yolink, it's this family of these sensors that use a radio protocol called LoRa that is much, much, much lower power and higher range than Wi-Fi.

00:10:58 ◼   ► And it's not based on thread, it's a whole different thing.

00:11:00 ◼   ► But, so, basically, you have these very inexpensive sensors that run on, like, AA batteries or a button battery and last, like, two years or a year on that battery life.

00:11:11 ◼   ► And because it's super low frequency, it goes through walls, fridges, you know, like, lots of things that other things might have signal problems going through, it goes right through them.

00:11:24 ◼   ► And it also has extremely long range.

00:11:26 ◼   ► Now, the canonical example for me is I have a contact sensor on my mailbox, as we've spoken about many times.

00:11:33 ◼   ► And granted, my mailbox is, I don't know, 10, 15 meters from the house or yards, if you will.

00:11:38 ◼   ► It's not that far, but I have one of these sensors, you know, inside a metal box.

00:11:44 ◼   ► And it's beaming the open and close signal all the way into the house.

00:11:49 ◼   ► And I've understood that in perfect conditions, like, basically, almost outdoors or outdoors, these things can go upwards of, like, a mile.

00:11:57 ◼   ► They can work astonishingly long ways in really good conditions.

00:12:02 ◼   ► Yeah.

00:12:02 ◼   ► So they're great.

00:12:03 ◼   ► And so I have Yoling sensors all over the restaurant.

00:12:06 ◼   ► I have them all over my house.

00:12:08 ◼   ► And I can monitor things, get alerted for things like, hey, like, this fridge is above temperature.

00:12:13 ◼   ► Like, something's wrong.

00:12:14 ◼   ► Go check it.

00:12:15 ◼   ► And there's the WaterLeak sensors are now everywhere.

00:12:18 ◼   ► And they work.

00:12:19 ◼   ► They work really well.

00:12:20 ◼   ► And so – and they're cheap.

00:12:22 ◼   ► Like, a Yoling sensor is maybe $20 or $30, depending on what kind of sensor it is.

00:12:26 ◼   ► So you can have – and there's no, like, subscription or anything.

00:12:28 ◼   ► So you can have a whole bunch of them.

00:12:29 ◼   ► And, you know, their app is what you'd expect.

00:12:32 ◼   ► It's not a good app.

00:12:34 ◼   ► No, it's not great, but it works.

00:12:35 ◼   ► However, there's really robust home assistant integrations if you're one of those dorks like me.

00:12:40 ◼   ► And you do need to get a hub, which is, like, $40 or $50.

00:12:44 ◼   ► $20, $30, yeah, you can get a set with, like, a hub and a couple of temperature monitors for, like, $50 or $60.

00:12:48 ◼   ► Like, it's not – it's not anything, you know, burdensome, really, in this world.

00:12:53 ◼   ► No.

00:12:53 ◼   ► And compared to, like – you know, I've had, like, the EVE home sensors before that, the EVE Airs.

00:12:59 ◼   ► Those are, like, $70 each, and they don't work as well.

00:13:03 ◼   ► I can tell you that from experience, they don't work as well.

00:13:06 ◼   ► And, you know, they have, like, smart outlets and stuff, too.

00:13:09 ◼   ► Like, you know, they have a whole family of smart stuff.

00:13:12 ◼   ► And it all is, like, cheap but functional.

00:13:16 ◼   ► And reliable.

00:13:16 ◼   ► Thanks for Yolink for sponsoring this show.

00:13:19 ◼   ► No.

00:13:19 ◼   ► They didn't sponsor, but they –

00:13:21 ◼   ► They should.

00:13:22 ◼   ► At this point, I'm happy to sponsor them because they have saved me a lot of anguish.

00:13:26 ◼   ► Well, that's the thing is I was moaning just a moment ago about how I don't want to put a $300 valve on my water supply.

00:13:32 ◼   ► The flip side of that is I've lived having part of your house underwater, and that's a lot more expensive than $300.

00:13:40 ◼   ► So, it is on the to-do list at some point to get either the, you know, have a plumber come out and weld in or whatever, you know, one of the fancy, fancy versions.

00:13:48 ◼   ► Or maybe one of the ones that just sits on the valve and, you know, twists it shut on its own.

00:13:52 ◼   ► But I really do need to look into this one way or another.

00:13:54 ◼   ► There's also –

00:13:55 ◼   ► So, and the way it works is, you know, you can obviously remote control your water main when you have one of these.

00:14:00 ◼   ► But also, they have integrations where, like, if you have a leak detected, it can automatically turn it off.

00:14:05 ◼   ► And that kind of thing also has insurance implications.

00:14:09 ◼   ► Some insurance companies will require that you have those in certain high-risk situations.

00:14:14 ◼   ► Or sometimes you might be able to get a discount on something if you have that.

00:14:17 ◼   ► So, that's worth looking into if that's relevant to you.

00:14:20 ◼   ► All right.

00:14:22 ◼   ► We have some news.

00:14:23 ◼   ► We recorded and released a new member special.

00:14:26 ◼   ► And perhaps more exciting than anything else, I remembered to talk about it.

00:14:30 ◼   ► So, gold star in cookie for me.

00:14:33 ◼   ► We have a new member special, ATP Insider After Apple.

00:14:37 ◼   ► John, what does that mean?

00:14:38 ◼   ► It's a little bit difficult to explain.

00:14:40 ◼   ► But we were kind of musing on a world without Apple in a couple of different hypothetical scenarios.

00:14:45 ◼   ► In the beginning, we tried to explain, well, why are we talking about this topic now?

00:14:48 ◼   ► You know, is this a declaration that we're abandoning Apple?

00:14:50 ◼   ► No, that's not what it is.

00:14:51 ◼   ► Except for Marco.

00:14:53 ◼   ► Maybe for Marco.

00:14:53 ◼   ► Well, no spoilers.

00:14:55 ◼   ► But anyway, we did a couple of hypotheticals.

00:14:57 ◼   ► Like, what if Apple had never existed, alternate history type of thing?

00:15:00 ◼   ► Or what if Apple, at some point in our future when we're still alive, ceases to exist?

00:15:03 ◼   ► And we're exploring what would we do in terms of tech?

00:15:07 ◼   ► What platforms would we use?

00:15:09 ◼   ► What things would we miss?

00:15:10 ◼   ► What things would we be looking for?

00:15:12 ◼   ► And obviously, in that discussion, we inevitably end up talking about our history and how we got where we are.

00:15:16 ◼   ► Especially since both Marco and Casey are fairly new to the Apple platform from my perspective.

00:15:22 ◼   ► So we get to hear a lot of backstory on that as well.

00:15:24 ◼   ► So there you go.

00:15:25 ◼   ► ATP Insider after Apple.

00:15:27 ◼   ► Chosen mostly for the alliteration, let's be honest.

00:15:30 ◼   ► But I think it's a pretty good title.

00:15:31 ◼   ► Agreed.

00:15:31 ◼   ► Yeah, this one was fun.

00:15:33 ◼   ► We went into a bunch of random places, as the three of us are wont to do.

00:15:36 ◼   ► But I really enjoyed it.

00:15:38 ◼   ► There's a lot of history there, both personal history and general history.

00:15:41 ◼   ► And I think you'll really enjoy it.

00:15:43 ◼   ► If you are not a member and this or any of the other, many other member specials are wetting your whistle or striking your fancy,

00:15:50 ◼   ► you can go to atp.fm slash join.

00:15:52 ◼   ► And for not an altogether terribly large amount of money, you can get all of the member specials.

00:15:58 ◼   ► You can do what John suggests, which Marco and I do not agree with.

00:16:02 ◼   ► And you can be a member for a month and then download everything and then cancel your membership.

00:16:06 ◼   ► We do make that very easy.

00:16:07 ◼   ► But you're not going to do that.

00:16:09 ◼   ► You're going to become a member and stick with us because you like us.

00:16:11 ◼   ► And so you can go to atp.fm slash join.

00:16:14 ◼   ► John, remind me for the 800th time, if people wanted to see what member specials were available,

00:16:19 ◼   ► what is the URL for that? Specials?

00:16:20 ◼   ► atp.fm slash specials.

00:16:22 ◼   ► There you go.

00:16:23 ◼   ► And so you can check out all of the different member specials.

00:16:26 ◼   ► There's a bunch of them at this point.

00:16:27 ◼   ► I don't remember the count, but there's been a lot.

00:16:29 ◼   ► And there's a separate feed for them, too.

00:16:32 ◼   ► Like when you become a member, you get a dedicated feed that just has the specials.

00:16:35 ◼   ► So if you just want to go back through the specials one by one, you don't have to hunt them down on the other feeds.

00:16:38 ◼   ► They're in all the other feeds, too.

00:16:39 ◼   ► But there's a dedicated feed just for specials, if you want.

00:16:42 ◼   ► So we appreciate all of the members that make this show possible.

00:16:46 ◼   ► And we really would love it if you'd check out these member specials.

00:16:49 ◼   ► We have a lot of fun doing them, and we get to go a little bit off the beaten path.

00:16:53 ◼   ► And I really enjoy that, too.

00:16:54 ◼   ► All right.

00:16:55 ◼   ► Let's do some follow-up, and let's start with Units Corner.

00:16:57 ◼   ► We were talking last week about Robert Tate's micro-LED, quote-unquote, TV review video,

00:17:02 ◼   ► where he basically installed a stadium monitor in his house.

00:17:06 ◼   ► And during that, the display had a pixel pitch of 0.9 millimeters.

00:17:12 ◼   ► And in the video, Robert said, and this was transcribed by John, the general rule for pixel density is that if you multiply the pixel pitch by 10,

00:17:21 ◼   ► that will give you the viewing distance in feet where individual pixels are indistinguishable.

00:17:25 ◼   ► John, apparently, you said the general rule is if you multiply the pixel pitch by 10, you get the distance where you can't see the pixels anymore.

00:17:34 ◼   ► So 0.9 millimeters means about 9 feet away.

00:17:36 ◼   ► John, where did you go wrong there?

00:17:38 ◼   ► I don't think I really went that wrong, but some people were confused by it and thought it was hilarious that I was mixing these units together.

00:17:45 ◼   ► You know, like, people are going to think my pronunciation of hilarious is also hilarious.

00:17:49 ◼   ► It is.

00:17:50 ◼   ► It's millimeters in feet, and then 0.9 millimeters times 10 does not equal 9 feet.

00:17:55 ◼   ► It doesn't equal 9 meters.

00:17:56 ◼   ► It doesn't even, I mean, you know, what are you talking about?

00:17:59 ◼   ► I omitted the words ind feet.

00:18:01 ◼   ► What I should have said is the words ind feet after distance.

00:18:03 ◼   ► If you multiply the pixel pitch by 10, you get the distance ind feet where you can't see the pixels anymore.

00:18:08 ◼   ► This is not my rule.

00:18:09 ◼   ► This is directly from the video, which is just a rule of thumb for theater people.

00:18:13 ◼   ► Apparently, that math works out.

00:18:14 ◼   ► I think it was perfectly clear because I gave the example.

00:18:16 ◼   ► So for 0.9 millimeters, that means 9 feet away.

00:18:19 ◼   ► But my bad for not putting the words ind feet in there.

00:18:22 ◼   ► And as I say here in the notes, a lot of listeners are now very unit sensitive.

00:18:26 ◼   ► Thanks to Casey.

00:18:28 ◼   ► Here it is.

00:18:30 ◼   ► I'm trying to be inclusive, and no good deed goes unpunished.

00:18:33 ◼   ► Speaking of member specials, a little while ago, I think it was a couple months ago, we did HP Movie Club War Games.

00:18:39 ◼   ► And this is kind of apropos because the whole of the internet, and that is not a complaint.

00:18:44 ◼   ► I'm glad that you all did.

00:18:45 ◼   ► The whole of the internet sent us a YouTube video from Dave's Garage where he bought basically one of the display units.

00:18:52 ◼   ► I don't know how to describe this.

00:18:54 ◼   ► That generated those screens in the bunker.

00:18:58 ◼   ► Those, like, those, I don't know.

00:19:01 ◼   ► How can I verbalize this?

00:19:02 ◼   ► Well, so, I mean, what we're talking about is in the movie War Games in the big room with all the screens that show, like, the lines with the missiles exploding and stuff.

00:19:08 ◼   ► Those are the screens we're talking about.

00:19:10 ◼   ► It's like, how did they do that back in 1980, whatever, when this was made?

00:19:13 ◼   ► Well, they did it with the thing you're going to see on Dave's Garage channel.

00:19:16 ◼   ► It's a vector display.

00:19:17 ◼   ► And, of course, it's very tiny.

00:19:19 ◼   ► And it's monochrome.

00:19:20 ◼   ► So how did they do those big things?

00:19:21 ◼   ► Basically, this is not really about how they made the movie.

00:19:23 ◼   ► But just my understanding is that they got these little vector displays.

00:19:26 ◼   ► They filmed them with a film camera.

00:19:28 ◼   ► And they filmed different images with different color filters in front of the monochrome screen.

00:19:33 ◼   ► So they got a red pass and a blue pass and a green pass, whatever, whatever colors they use.

00:19:37 ◼   ► Like, you see colors in the movie.

00:19:38 ◼   ► That's just because they put different images.

00:19:39 ◼   ► And they filmed it with a film camera.

00:19:40 ◼   ► And then they put that film together.

00:19:43 ◼   ► And then they projected onto projection screens.

00:19:45 ◼   ► And then they filmed the projection screens with the movie camera.

00:19:48 ◼   ► So it's quite a ways to go.

00:19:49 ◼   ► But, you know, without computers, they can do graphics that big.

00:19:52 ◼   ► But anyway, the vector display itself is, like, I don't know, it was, like, seven inches diagonal or something.

00:19:56 ◼   ► It's a tiny.

00:19:56 ◼   ► It looks like an oscilloscope screen, practically.

00:19:58 ◼   ► But vector displays are really cool.

00:20:00 ◼   ► And getting it up and running is really cool.

00:20:01 ◼   ► That's really what the video is about, is.

00:20:03 ◼   ► So I bought one of these War Games displays on YouTube.

00:20:06 ◼   ► Again, they're not War Games displays, but they're used for the movie.

00:20:08 ◼   ► How do I get it to do literally anything?

00:20:11 ◼   ► And it was a fun video.

00:20:12 ◼   ► It was very well done.

00:20:14 ◼   ► And, you know, the Dave fellow was like, I'm not a developer.

00:20:16 ◼   ► But let me tell you about the low-level programming I had to do in the communication protocol I had to work through in order to get this to work.

00:20:23 ◼   ► It was very impressive.

00:20:24 ◼   ► We are sponsored this episode by Squarespace, the all-in-one website platform designed to help you stand out and succeed online.

00:20:32 ◼   ► Whether you're just starting out or scaling your growing business, Squarespace gives you everything you need to claim your domain name, showcase your offerings with a professional website, grow your brand, and get paid all in one place.

00:20:44 ◼   ► So Squarespace is an amazing website hosting platform to host your business.

00:20:49 ◼   ► And they've, for years, offered amazing features to host businesses that sell physical or digital goods.

00:20:55 ◼   ► Well, they've also, in recent years, broadened their offering to also give you everything you need to offer services and get paid for your services all in one place.

00:21:04 ◼   ► From consultations to events or experiences, you can showcase your offerings with a customizable website designed to attract new clients and grow your business.

00:21:13 ◼   ► And then you can get paid on time and easily with professional on-brand invoices and online payments.

00:21:19 ◼   ► Plus, you can streamline your workflow with built-in appointment scheduling and email marketing tools and so much more.

00:21:25 ◼   ► They have everything you will need to make your business succeed.

00:21:28 ◼   ► Things like SEO tools, analytics, email campaigns, and, of course, this scales to so many different kinds of businesses.

00:21:34 ◼   ► Again, physical goods, digital goods, things like services, and if you're a consultant, as they mentioned earlier.

00:21:40 ◼   ► But even things like, you know, if you sell content, if you sell a private podcast or newsletters or e-books, you can do all of that with Squarespace and so much more.

00:21:48 ◼   ► And you don't need to be a nerd.

00:21:50 ◼   ► There's no coding.

00:21:50 ◼   ► There's nothing like that.

00:21:51 ◼   ► It's super easy.

00:21:52 ◼   ► So if somebody comes to you and you are a nerd and they're like, hey, can you help me with my website?

00:21:56 ◼   ► You can send them to Squarespace and they don't need your help.

00:21:59 ◼   ► It's amazing.

00:22:00 ◼   ► See for yourself by starting a free trial at squarespace.com slash ATP.

00:22:05 ◼   ► You can build the whole thing in trial mode and see how it works for you.

00:22:08 ◼   ► When you're ready to sign up, go back there, squarespace.com slash ATP, and use offer code ATP to save 10% off your first purchase of a website or domain.

00:22:16 ◼   ► Once again, squarespace.com slash ATP.

00:22:18 ◼   ► Code ATP for 10% off.

00:22:20 ◼   ► Thank you so much to Squarespace for sponsoring our show.

00:22:23 ◼   ► All right, we're going to talk about AI for a few minutes.

00:22:29 ◼   ► This is still part of follow-up.

00:22:30 ◼   ► We're going to do this a little bit in a funky way.

00:22:32 ◼   ► What we're going to do is I'm going to introduce a couple things, and John, I think you would like to say a couple things.

00:22:37 ◼   ► But then I'm going to plow through a whole bunch of stuff, and there will be a bunch of links in the show notes.

00:22:43 ◼   ► And we're going to try to get through all of the stuff that I need to get through to set the stage.

00:22:51 ◼   ► And then I'll let you two loose to have a conversation about it.

00:22:55 ◼   ► But bear with us for a minute because it really requires a whole bunch of foundation before we can really have a conversation about it.

00:23:00 ◼   ► So first and foremost, on behalf of all three of us, we all apologize.

00:23:05 ◼   ► Genuinely, we all apologize for not having disclosed past or future sponsorships from AI companies.

00:23:11 ◼   ► I like to think we do a pretty good job of this.

00:23:13 ◼   ► We try really hard to do a good job of this.

00:23:15 ◼   ► But we 100% whiffed on this one.

00:23:17 ◼   ► And so that is our bad, and the offending parties have been sacked.

00:23:21 ◼   ► But we should note that Google Gemini has been a past sponsor.

00:23:25 ◼   ► Anthropic has been a past sponsor, will be a future sponsor.

00:23:28 ◼   ► And genuinely, I hope I don't come across as sarcastic.

00:23:31 ◼   ► We really are sorry that we didn't say that.

00:23:33 ◼   ► So our bad.

00:23:33 ◼   ► And again, I think our batting average is really high.

00:23:36 ◼   ► And I, Casey, is like 99% of the time is always remembering to do it.

00:23:39 ◼   ► I think part of the reason that it didn't even occur to me for these sponsors is kind of like if we had to say, by the way, Apple past sponsor of the show.

00:23:46 ◼   ► But of course, Apple has never sponsored the show.

00:23:48 ◼   ► But we talk about Apple so much.

00:23:49 ◼   ► Like these big companies, like Google obviously is in the league with Apple.

00:23:52 ◼   ► And now like Anthropic and OpenAI, like they're such big companies.

00:23:55 ◼   ► It's like, well, they're not.

00:23:57 ◼   ► They don't sponsor.

00:23:58 ◼   ► Our sponsors are small companies that want to get their message out.

00:24:01 ◼   ► We're not doing ads for Google, but we did.

00:24:03 ◼   ► We did do it for Google.

00:24:04 ◼   ► And the same thing for Anthropic.

00:24:06 ◼   ► Now these companies have hojillions of dollars and they are advertising.

00:24:09 ◼   ► So obviously, you listen to the show, you would hear that they're sponsors.

00:24:11 ◼   ► And even if you're a member, you can look in our show notes and you will see in the show notes who our sponsors are.

00:24:15 ◼   ► But we do try very, very hard.

00:24:16 ◼   ► Anytime we bring up a sponsor, Casey will say, and blah, blah, blah, former sponsor, blah, blah, blah, former future sponsor.

00:24:22 ◼   ► We try to do that pretty much all the time.

00:24:24 ◼   ► Obviously, we're not going to say that every time we mention them within an episode.

00:24:26 ◼   ► But we do try to, like the first time they're brought up or whatever, mention it.

00:24:30 ◼   ► And we didn't do it for Google Anthropic.

00:24:32 ◼   ► And I have to say, going forward, it's going to be difficult to remember to do it every time we bring up AI type things.

00:24:38 ◼   ► This is like Google.

00:24:39 ◼   ► We talk about Google all the time.

00:24:40 ◼   ► But anyway, we will we continue to try just to remind you because not everyone has listened to every episode.

00:24:46 ◼   ► And why do we tell you this?

00:24:47 ◼   ► We're telling you this so you can add whatever grains of salt you want to add to our opinions.

00:24:50 ◼   ► Some people will believe, oh, you had a Google, one Google gem on an ad on a past episode.

00:24:54 ◼   ► Therefore, everything you say about Google is bought and paid for by Google.

00:24:56 ◼   ► If you want to believe that, you should believe that.

00:24:58 ◼   ► Other people will say, I'm going to knock off 0.02% from everything you say because Anthropic brought ads on your show.

00:25:04 ◼   ► Whatever you want, whatever, how many or any grains of salt you want to add to our opinion, you do that.

00:25:08 ◼   ► But we have to tell you so that you know to either add grains of salt or not add grains of salt.

00:25:13 ◼   ► Obviously, we all think as humans all think I'm not influenced by them having ads or whatever.

00:25:16 ◼   ► And I think that's mostly true.

00:25:19 ◼   ► But, you know, it's very always very difficult to say that's why we want the audience to know.

00:25:23 ◼   ► Indeed.

00:25:24 ◼   ► So, with that in mind, I wanted to first highlight some stuff from front of the show, Steve Trouton-Smith.

00:25:32 ◼   ► Steve has been on a bit of a journey with regard to Codex and Xcode and has a bunch of different tweets on Mastodon about it.

00:25:41 ◼   ► And there's a thread about it.

00:25:43 ◼   ► John has pulled out some highlights that I'm going to read.

00:25:45 ◼   ► Again, we tried to cut this down to the bare minimum, but there's a lot here.

00:25:49 ◼   ► And it is worth looking into if you haven't seen these tweets.

00:25:51 ◼   ► It is fascinating.

00:25:52 ◼   ► And the short answer before we even get into this is Steve Trouton-Smith is excited by and doing things that he thinks is good with AI products.

00:26:00 ◼   ► So, with that in mind, Steve says,

00:26:02 ◼   ► I had long since rewritten an app that he was working on in Swift.

00:26:06 ◼   ► So, this was only a contrived test of Xcode's new agentic programming support.

00:26:10 ◼   ► But it did in five minutes what took me months of on-again, off-again effort and preparation.

00:26:14 ◼   ► Objective-C to Swift is too easy, all things considered.

00:26:18 ◼   ► What about porting the app to another platform like Android?

00:26:21 ◼   ► Hold my beer.

00:26:22 ◼   ► So, then later.

00:26:24 ◼   ► So, now I have a one-to-one recreation of classic same game, the app in question, in Android's Java slash XML.

00:26:29 ◼   ► Later, I passed it the Swift version of the project that it had created from the original Objective-C project and instructed it to turn it into an Android project that I could just open in Android Studio and install.

00:26:41 ◼   ► So, now it just exists on Android.

00:26:44 ◼   ► Steve continues,

00:26:46 ◼   ► I've spent some time cleaning up the project and prepping it for Google Play.

00:26:48 ◼   ► I've tested Google Play distribution internally, at least.

00:26:51 ◼   ► I'm still reeling from the fact that this is possible, never mind viable.

00:26:54 ◼   ► I'd tried this conversion process before, perhaps a year ago, with the agent feature on ChatGPT and got nowhere.

00:26:59 ◼   ► There's clearly been a lot of progress made here in a short amount of time.

00:27:03 ◼   ► Sticking with Steve, but kind of, you know, on a different subject, Steve writes,

00:27:08 ◼   ► Three apps, from zero to functional in a day.

00:27:12 ◼   ► This is the biggest change to Xcode and Project Builder in its entire history, dropped on a random Tuesday when the 26.3 beta was released.

00:27:19 ◼   ► So, Steve went from zero apps to three that were functional in the span of a day.

00:27:23 ◼   ► And then finally from Steve,

00:27:25 ◼   ► Building one app used 7% of my weekly codex usage limit.

00:27:30 ◼   ► Compare that to a single awful slideshow and keynote using 47% of my monthly Apple Creator Studio usage limit.

00:27:37 ◼   ► Yikes.

00:27:39 ◼   ► Before Steve had embarked on this whole, like, I'm going to write a bunch of new apps and important things,

00:27:44 ◼   ► because he had been exploring the AI features in Apple's Creator Studio and complaining bitterly that even just experimenting with it to try out the features was burning through the allocation of free, you know, free AI things that you get to do with, like, iWork or whatever, like, you know, give you an outline, generate a slideshow from it or whatever.

00:28:01 ◼   ► Those limits seem really, really low.

00:28:03 ◼   ► I mean, you're not paying a, I mean, you are paying a subscription to Creator Studio, but you're not paying a separate subscription to, like, whatever models it's using behind the scenes.

00:28:12 ◼   ► So on one hand, I kind of understand how, like, okay, well, Creator Studio, you pay a subscription for the software, and we give you some piddling amount of, like, you know, server-side AI-type things for free, but if you want more than that, you're going to hook up to an API key or something.

00:28:24 ◼   ► On the other hand, comparing this, like, I made three entire apps and used 7% of my $20 a month usage, like, you know, I don't know if Apple's AI usage prices are going to be, like, their RAM prices, or maybe that's a bad example these days.

00:28:36 ◼   ► But, yeah, it seems like Apple should make an adjustment here or charge more money for their Creator Studio.

00:28:41 ◼   ► So we used Steve Trout and Smith as an example of, yay, AI.

00:28:45 ◼   ► There's been a couple of posts that have been making the rounds recently, and that several of you had pointed out to us.

00:28:50 ◼   ► That's kind of about, like, AI dread and fatalism.

00:28:53 ◼   ► We're not going to really talk about them right this second, but I wanted to call them to your attention.

00:28:56 ◼   ► Again, links in the show notes.

00:28:57 ◼   ► We Mourn Our Craft by Nolan Lawson and The Tipping Point by Thomas Ricard.

00:29:02 ◼   ► I probably pronounced that wrong.

00:29:03 ◼   ► I'm sorry.

00:29:04 ◼   ► But there's a couple of, like, more pessimistic takes on AI and what that means for development and, you know, software engineering and so on, which I think are worth your time.

00:29:12 ◼   ► Whether or not you agree with them, I think it's interesting.

00:29:15 ◼   ► And then one or more anonymous people wrote in with regard to AI and productivity and the promise of, you know, oh, all of these people will 10x their productivity versus, you know, scientific studies about it.

00:29:29 ◼   ► So, again, I'm just going to kind of hit the high level real fast.

00:29:31 ◼   ► The scientific evidence doesn't actually back the idea that AI provides a huge productivity boost.

00:29:35 ◼   ► A recent study from Anthropic found that for junior developers on a set task, the time taken to complete a task was not statistically significantly faster, just two minutes, than those who did not use AI tools.

00:29:46 ◼   ► And their understanding of the output and ability to answer questions on it was actually much worse.

00:29:50 ◼   ► I'll put a link in the show notes.

00:29:51 ◼   ► Separately, a randomized controlled trial from mid-2025 found that experienced developers were slower with AI.

00:29:58 ◼   ► Separately, a Microsoft study from 2024 found that for trivial tasks, the users were able to easily describe AI could save some time.

00:30:08 ◼   ► But for complex tasks, it was of minimal benefit.

00:30:11 ◼   ► So, trivial tasks that you can describe to AI quickly, yeah, you save time.

00:30:14 ◼   ► But anything kind of complex, eh, not so great.

00:30:16 ◼   ► Then a survey from Atlassian found that only 4% of companies are seeing return on investment on their integration of AI tools.

00:30:23 ◼   ► Separately, a study from Berkeley at the California Review Management found that,

00:30:28 ◼   ► there is no statistically significant relationship between AI adoption and productivity gains.

00:30:32 ◼   ► At Harvard Business Review, there's a study that describes how AI-generated slop is actually harming productivity, even when it comes to be, when it seems to be producing work.

00:30:41 ◼   ► Think, an increase in the number of PRs to an open-source project, then in actuality just wastes the maintainer's time.

00:30:46 ◼   ► Then finally, MIT's State of AI in Business finds that 95% of companies are seeing zero returns.

00:30:52 ◼   ► Cool, cool, cool, cool, cool, cool, cool, cool, cool.

00:30:54 ◼   ► John, I am happy to go through these links that you compiled, or would you prefer to?

00:30:58 ◼   ► How would you like to proceed?

00:30:59 ◼   ► No, you can.

00:30:59 ◼   ► It's just some more in the same vein.

00:31:01 ◼   ► I had some of these links you just read I had also had in the notes ready to go, but we got many of them sent by listeners as well.

00:31:06 ◼   ► This is all, I think, from one listener that we just read this list, by the way.

00:31:09 ◼   ► All right, and so some more from John.

00:31:11 ◼   ► AI-generated code quality and the challenges we all face by Mark Leveson.

00:31:14 ◼   ► There were findings, the key findings from five different studies.

00:31:17 ◼   ► AI-assisted poll requests have 1.7 times more issues than human-authored poll requests.

00:31:22 ◼   ► Technical debt increases 30% to 41% after AI tool adoption.

00:31:27 ◼   ► Cognitive complexity increases 39% in agent-assisted repos.

00:31:31 ◼   ► Initial velocity gains disappear in the first few months.

00:31:34 ◼   ► Change failure rate is up 30%, and incidents per poll request is up 23.5%.

00:31:40 ◼   ► Then another article, Comprehension Debt, the ticking time bomb of LLM-generated code by Armin Shakar.

00:31:47 ◼   ► Comprehension debt refers to the gap between the code generated by LLMs and a developer's understanding of that code.

00:31:53 ◼   ► In this post, we will explore the implications of comprehension debt, its impact on software quality and maintainability, and strategies for mitigating its effects.

00:32:00 ◼   ► And then finally from John,

00:32:01 ◼   ► Your Brain on ChatGPT, Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Tasks by Natalia Cosmina, Cosmina, or something like that, and others.

00:32:11 ◼   ► Anyways, while LLMs offer immediate convenience, our findings highlight potential cognitive costs.

00:32:17 ◼   ► Over four months, LLM users consistently underperformed at neuro-linguistic and behavioral levels.

00:32:21 ◼   ► These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI's role in learning.

00:32:30 ◼   ► Okay, fellas, thank you for being patient.

00:32:32 ◼   ► What do you got?

00:32:33 ◼   ► So that was a lot, and I think it was structured in this way.

00:32:36 ◼   ► Steve Trouton-Smith, who is a frequent person cited on the show, is a very skilled, experienced developer on Apple platforms and just a very generally very good coder.

00:32:44 ◼   ► So this is not a novice.

00:32:46 ◼   ► He's using it.

00:32:47 ◼   ► He's having great success.

00:32:48 ◼   ► He's porting apps that he had tried to port himself, taken months and not done it, and then he's doing it in five minutes.

00:32:54 ◼   ► So it's not like somebody who is, I wasn't able to do this thing before because I'm not a skilled coder, but now that I have AI, I can.

00:32:59 ◼   ► This is a skilled, experienced coder who is fairly productive, does a lot of stuff, and is having great success.

00:33:05 ◼   ► Then he has ideas for apps and whips them up in two seconds.

00:33:08 ◼   ► He's impressed by it.

00:33:09 ◼   ► And then we have all the people who are scared about it, and then we have, but how good actually is it and these people studying it?

00:33:16 ◼   ► Oh, so AI, everyone seems to love it and they think it's great, but actually it's not helping productivity.

00:33:20 ◼   ► It's hurting it and all these studies finding that it's not giving a return on investment.

00:33:25 ◼   ► What's the truth?

00:33:26 ◼   ► How can we square this circle?

00:33:28 ◼   ► And I think everything that we just read is the truth.

00:33:33 ◼   ► The truth is that AI is massively overhyped.

00:33:36 ◼   ► There's a lot of money behind it.

00:33:38 ◼   ► Forcing your employees to use it, which many companies have been doing, which blows my mind.

00:33:42 ◼   ► By the way, when you have something that people like, you don't have to force them to do it.

00:33:45 ◼   ► Forcing people to use it.

00:33:47 ◼   ► It's like even if you don't want to use AI, if you're programmed for this company, you're forced to use it.

00:33:50 ◼   ► Because it's going to increase productivity and this, that, and the other thing.

00:33:52 ◼   ► It doesn't surprise me that it's not having productivity games.

00:33:56 ◼   ► And I'm not saying AI is a fad, but it reminds me of a lot of the sort of business fads that go around for practices that are good.

00:34:03 ◼   ► Things that are actually good, but that become sort of trendy or popular or like mandatory to do.

00:34:10 ◼   ► And then companies will just force like everybody to do them, even though they're good things.

00:34:15 ◼   ► Like, you know, pick Agile, which we always make fun of.

00:34:18 ◼   ► But all sorts of other sort of good programming practices that are actually good ideas and have longevity about, you know, ways to architect code or ways to organize people or whatever.

00:34:27 ◼   ► Anytime there is any idea that is actually good in the technology industry, it gets overused to the point where people hate it.

00:34:35 ◼   ► And it causes negative productivity because people feel like they have to use it.

00:34:39 ◼   ► And AI is like that times a million.

00:34:41 ◼   ► That's why you see AI in every single product in the world and every feature.

00:34:45 ◼   ► And, you know, companies have entire releases where the only directive from the company is whatever you're doing in this company.

00:34:51 ◼   ► I don't care.

00:34:52 ◼   ► Next release, you've got to have AI something.

00:34:53 ◼   ► Reportedly, Apple did that.

00:34:54 ◼   ► Look at any product that you see.

00:34:56 ◼   ► See that all the time.

00:34:57 ◼   ► So it doesn't shock me that people aren't seeing productivity games on it.

00:35:00 ◼   ► It's not just because, oh, they're using it wrong.

00:35:01 ◼   ► It's because it's not all things to all people.

00:35:04 ◼   ► And especially when we go into the essay writing and stuff like that, as you as you start to stray more into like the other things that LLMs can do, it becomes a lot spotty.

00:35:14 ◼   ► On the other hand, in the hands of a skilled developer doing specifically the few things that it's actually very good at, the value is tremendous.

00:35:21 ◼   ► Like Steve Trouten-Smith is just one person, so it's anecdated or whatever.

00:35:24 ◼   ► But there's a reason lots of coders are out there saying this is really doing things for me faster, better than I was before.

00:35:33 ◼   ► And as I said in the last episode, it really is a skill to learn how to use these things.

00:35:36 ◼   ► So even Steve Trouten-Smith would admit that he's climbing that learning curve of how to actually use this tool productively.

00:35:42 ◼   ► Again, my analogy last time was like Photoshop.

00:35:44 ◼   ► It can massively increase your power as a graphics designer, but it's a lot to learn if you've never used a computer before back in the day.

00:35:51 ◼   ► And this seems to be like that.

00:35:53 ◼   ► And then finally, wrapping all this up, as I tried to emphasize last week and continues to be true.

00:35:59 ◼   ► All right.

00:36:00 ◼   ► Well, so does that mean once we just learn how to use AI at the things that it's good at in a good way and we aren't all forced to use it and the hype, you know, the bubble bursts and the hype is gone and we figure out what is it actually good for and what is it not good for?

00:36:11 ◼   ► Everything will be fine.

00:36:12 ◼   ► It'll be smooth sailing.

00:36:13 ◼   ► No, because there's still absolutely the unresolved issue of how is it that these tools came into being and how are the sources of the power that they have drive being compensated?

00:36:23 ◼   ► I think maybe this is an unrealistically optimistic view.

00:36:26 ◼   ► I think it is plausible to have an LLM that is as useful as the thing Steve Trouten-Smith is using to, you know, be more productive and get more stuff done.

00:36:39 ◼   ► That is entirely trained in an ethical way.

00:36:43 ◼   ► In the same way that Adobe and Photoshop has that their models are trained entirely on licensed images.

00:36:48 ◼   ► Like, so they paid for the images.

00:36:50 ◼   ► They train them out on the images.

00:36:51 ◼   ► They put that feature in Photoshop.

00:36:52 ◼   ► I think that is a there's nothing wrong with that.

00:36:56 ◼   ► You know, a lot of people who wrote in with AI feedback were like, I'm against AI in all forms because it's just 100% evil.

00:37:01 ◼   ► I think even they would agree like, OK, but what about the Adobe thing?

00:37:04 ◼   ► That's fine, right?

00:37:05 ◼   ► They would say, no, you're you're draining the deserts of all their water or whatever.

00:37:08 ◼   ► But it's like you're not.

00:37:09 ◼   ► It's just it's not, you know, the energy usage of of the AI feature in Photoshop is not significant in the grand scheme of human energy usage.

00:37:16 ◼   ► And I think they're if they're telling the truth, which I believe they are about only training their models on data that they owned or legally licensed.

00:37:23 ◼   ► There you go.

00:37:25 ◼   ► I think it's perfectly fine.

00:37:26 ◼   ► And I think that is possible for coding agents.

00:37:30 ◼   ► Right now, it's not happening because, you know, even if you have like an MIT license, open source thing, does the MIT license say anything about AI training?

00:37:38 ◼   ► I bet it doesn't because it didn't exist when the MIT license was written.

00:37:41 ◼   ► So even even the most permissive open source license probably doesn't explicitly allow AI training.

00:37:47 ◼   ► So it's an open question.

00:37:48 ◼   ► But you can imagine a future in which I know Stack Overflow is doing this.

00:37:51 ◼   ► The Stack Overflow is essentially licensing their data to AI people.

00:37:54 ◼   ► And I think people are grumpy about that.

00:37:55 ◼   ► But legally, you know, probably good ethically, whatever.

00:37:58 ◼   ► I mean, Stack Overflow is dead with AI.

00:38:01 ◼   ► I know.

00:38:02 ◼   ► But like the only asset they have are all their questions that people train the models on.

00:38:05 ◼   ► So they're selling those to they're selling access to those.

00:38:07 ◼   ► Yeah, I'm saying like that, like let them do whatever they need to survive right now, because I like that's not a good place.

00:38:14 ◼   ► Like that's not a good business to be in right now because AI has is completely replacing it.

00:38:18 ◼   ► I know, but the AI the AI coding agents could not exist at their current level without training on things like Stack Overflow.

00:38:23 ◼   ► So, you know, get while the getting is good.

00:38:25 ◼   ► But also for open source projects, you can imagine a future in which licenses are made that say explicitly things about whether or not AI training is allowed and under what conditions.

00:38:34 ◼   ► And if that's the case, I can imagine a future Adobe style coding agent that is entirely trained on licensed code that is, you know, licensed from Stack Overflow, licensed from open source that explicitly allows us in a license.

00:38:48 ◼   ► And I think what we've shown with current technology is coding agents can be useful productivity boost if you don't think that they are completely replacement for everything and everybody has to use them.

00:38:59 ◼   ► That's, you know, like if you contain the hype and just look at what they're actually good at, that is possible.

00:39:04 ◼   ► Now, I grant that's not happening now.

00:39:06 ◼   ► Now, people are using these agents and they don't care where they came from, just like we don't care where our food came from.

00:39:10 ◼   ► We just eat it.

00:39:10 ◼   ► If it tastes good, we're happy with it.

00:39:11 ◼   ► And we don't care that what terrible things happen to get this food to our, you know, like it's human nature to say.

00:39:18 ◼   ► Steve Trouten-Smith to be excited by the fact that he made three apps in a day and did this porting thing that he could never do before, despite months of trying.

00:39:24 ◼   ► And he gets done in an afternoon.

00:39:26 ◼   ► As someone who's a technology enthusiast, you're going to be excited about that.

00:39:29 ◼   ► But people have to be, you know, it's important to still remember that just because you're excited about it and it did a cool thing doesn't mean every issue involved in this technology has been resolved because you have the good thing that you want.

00:39:40 ◼   ► But I'm still hopeful about it because I agree with the promise of it.

00:39:45 ◼   ► I agree that it can do great things in the right hands in limited circumstances.

00:39:48 ◼   ► And I think for coding specifically and not the entire universe of the rest of the things that people want AI to do, which is apparently mostly talking to it as a therapist, which is a whole other ball of wax.

00:39:57 ◼   ► But anyway, for coding specifically, my personal experience has shown me that this is a powerful, useful tool that I wish I could access in a much more ethical and also I would add cost effective way.

00:40:11 ◼   ► But even now, yeah, like it's that's the that's the challenge we face.

00:40:15 ◼   ► And I think like I would try to give this kind of balanced picture of AI last time, but people tend to latch on to the part that they either do or don't want to hear about it.

00:40:25 ◼   ► And the problem is it's a complicated topic.

00:40:26 ◼   ► Like I don't agree with the people who say it's 100 percent evil and we should stop it in all forms and never pursue it in any way.

00:40:33 ◼   ► I don't think that's the right thing to do, but I do agree with the people who say it still has problems and we should ignore those problems just because it's doing something useful for us.

00:40:40 ◼   ► The truth is what we just read, that it's it is great at certain things and there are a whole bunch of other problems that are unresolved.

00:40:47 ◼   ► And it's currently a bubble and it's being overhyped and it's there's a lot of bad that's going to come from it.

00:40:51 ◼   ► But, you know, it's a it's a complex topic.

00:40:54 ◼   ► I don't know.

00:40:55 ◼   ► What do you guys think?

00:40:55 ◼   ► I know we all got a lot of feedback about this.

00:40:57 ◼   ► I only saw the feedback that like ATP got and I got, but did you specifically get any feedback you want to address?

00:41:01 ◼   ► A little bit.

00:41:03 ◼   ► So earlier today I got my keyster to properly start working on the iPad layout and call sheet, which has been trashed since call sheet was around.

00:41:13 ◼   ► And it's been too long.

00:41:14 ◼   ► I should have already taken care of it, but I hadn't yet.

00:41:17 ◼   ► And my first cut at this was, all right, let me use the 26.3 beta and let me ask, you know, the Xcode agentic stuff via Claude.

00:41:26 ◼   ► And I don't recall what model I was on and I should have paid attention.

00:41:30 ◼   ► I didn't think about it, but let me ask it to turn the cast and crew list and call sheet, which right now on iPad is one column, which is awful.

00:41:38 ◼   ► It looks terrible.

00:41:39 ◼   ► It should at least be two columns.

00:41:41 ◼   ► Maybe.

00:41:41 ◼   ► And I have some ideas from maybe something even more advanced than that.

00:41:44 ◼   ► But my first cut was, hey, can you just make this two column?

00:41:48 ◼   ► And it took a couple of back and forths between me and Xcode, I guess, or Claude, or I don't know how to describe this, how to verbalize this.

00:41:56 ◼   ► But it took a couple of back and forths and it got it to the point that visually it was exactly what I wanted.

00:42:02 ◼   ► And that only took, I don't know, 10-ish minutes, something like that.

00:42:04 ◼   ► It really was not long at all.

00:42:06 ◼   ► But functionally, it was a mess.

00:42:07 ◼   ► And the particulars don't matter.

00:42:11 ◼   ► But suffice it to say, it totally screwed up navigation.

00:42:13 ◼   ► A lot of things were hosed.

00:42:15 ◼   ► And I've heard rumblings that these things don't do too well with SwiftUI.

00:42:20 ◼   ► And the theory is in part because SwiftUI is a moving target, in part because it's so new.

00:42:25 ◼   ► But one way or another, I can tell you that I was not exceedingly impressed by that experience.

00:42:32 ◼   ► And additionally, shoot, now I'm drawing a blank as to what it was.

00:42:35 ◼   ► But there's something else I had to do that was much less invasive.

00:42:39 ◼   ► Golly, it's really going to drive me nuts because I can't remember what specifically it was.

00:42:42 ◼   ► But something in call sheet.

00:42:43 ◼   ► And I wanted it to do something.

00:42:45 ◼   ► And, oh, I think I needed it to pump or to pipe a date from, you know, like high up in the object graph to deep within the object graph.

00:42:55 ◼   ► And the particulars are not that important.

00:42:58 ◼   ► But I asked it to do that for me.

00:42:59 ◼   ► And it did.

00:43:01 ◼   ► But it did it in a very clunky, very immature way.

00:43:04 ◼   ► And as I dug into what it was doing and what it had done to figure out, okay, is this really what I would have done?

00:43:10 ◼   ► Was I saving myself rote work or am I punting on something that I should really have a think about?

00:43:17 ◼   ► And it was quickly obvious to me that, no, I punted on something that I should have had a think about.

00:43:22 ◼   ► And I could go back and forth and say, no, don't do it that way.

00:43:25 ◼   ► Do it this way.

00:43:26 ◼   ► Or why did you choose that instead of this and blah, blah, blah.

00:43:28 ◼   ► And I could have done that.

00:43:30 ◼   ► And there have been instances in less complicated topics that that's worked out really, really well for me.

00:43:35 ◼   ► But in this particular case, it didn't give me the results I wanted.

00:43:40 ◼   ► And I ended up, for the piping through the object graph, I ended up doing that all by hand.

00:43:47 ◼   ► And for the iPad stuff, I had to put it away because I had other stuff I needed to do.

00:43:52 ◼   ► I think I'll probably take another stab at it probably tomorrow.

00:43:55 ◼   ► But I get the vibe that I'm going to end up going back and just rolling all of this by hand.

00:44:03 ◼   ► Because for this particular use case, I'm not trying to extrapolate to everything under the sun.

00:44:08 ◼   ► But for me, in CallSheet, it's been okay.

00:44:12 ◼   ► Like, simple stuff it can handle really well.

00:44:15 ◼   ► But anything even vaguely complicated, it's not great.

00:44:18 ◼   ► And it's also the learning curve of using the tools, like I said.

00:44:20 ◼   ► If you see the people who are more experienced at using these tools, it is actually a fairly complicated skill.

00:44:25 ◼   ► And that kind of leads to the, okay, but is the time investment in building up that skill?

00:44:31 ◼   ► And the time required to essentially execute it.

00:44:33 ◼   ► Like, Steve Troughton Smith posted.

00:44:33 ◼   ► We don't have a link to it, but maybe we can find it for, if you follow the threads, you'll probably see it.

00:44:37 ◼   ► He posted, what did I have to say to AI to make one of these apps that he made?

00:44:41 ◼   ► And it was 77 different prompts.

00:44:43 ◼   ► And they were fairly complicated.

00:44:44 ◼   ► And it's kind of like one of those things of like, like I said, if you're like mentoring a new employee, it's faster if you just do it yourself.

00:44:50 ◼   ► But like, yeah, but I'm mentoring a new employee.

00:44:51 ◼   ► Well, you're not mentoring an AI.

00:44:53 ◼   ► They're not learning from what you're talking about.

00:44:54 ◼   ► Like, in general, like, the model is the model until they get, you know, it's obviously you got your context window, but that's basically it.

00:45:00 ◼   ► So you have to think, like, do I want to spend the time mentoring a box of numbers that is just going to be replaced by a different box of numbers later?

00:45:07 ◼   ► Or do I just want to do it myself?

00:45:09 ◼   ► But still, like, there are targeted instances when used by a skilled practitioner where it can do amazing things that we're not possibly for, which is why people are excited for it.

00:45:17 ◼   ► And one thing I'll add to that, by the way, I still haven't let these things touch any of my apps code, not because my apps code is beautiful or anything, but just because I don't I don't know.

00:45:24 ◼   ► I just don't want to or I don't I don't think I need to at this point because I'm not doing anyway, whatever.

00:45:28 ◼   ► But I did find something to do because I'm trying to use of the last of my little hundred dollar subscription, which I did cancel, by the way.

00:45:34 ◼   ► Oh, good for you.

00:45:34 ◼   ► Yeah, I was trying to it's going to run out.

00:45:37 ◼   ► It's going to expire soon.

00:45:37 ◼   ► So I'm like, well, let me just use the last little bits of it.

00:45:39 ◼   ► Here is one of the things that I would recommend.

00:45:42 ◼   ► If you ever want to play with one of these things, you're like, well, I'm not going to, like, learn how to be a wrangler and do use a coding agent to write my codes for me.

00:45:48 ◼   ► It gets as complicated or whatever.

00:45:49 ◼   ► Like, but if you want to try something, one of my suggestions would be pointed at your code base.

00:45:54 ◼   ► You have to at least figure out how to give it stern instructions to never to not modify your code at all.

00:45:58 ◼   ► But it's mostly good about that.

00:46:00 ◼   ► But it helps to just maybe put in a Claude that MD or whatever your equivalent is and say, don't ever commit anything.

00:46:06 ◼   ► Don't ever modify any files.

00:46:07 ◼   ► You have read only whatever.

00:46:08 ◼   ► But just simply say this.

00:46:10 ◼   ► Find any bugs in this program.

00:46:11 ◼   ► And it will find a bunch of crap that are not bugs.

00:46:15 ◼   ► But it will find some actual bugs, too.

00:46:18 ◼   ► Even if it's just a typo and a variable name, it will find them.

00:46:21 ◼   ► Yeah, actually, I did that with Claude Code, not with the new beta.

00:46:24 ◼   ► I did that like a week ago with Claude Code.

00:46:26 ◼   ► And for the most part, it was nothing that was that super-duper important.

00:46:30 ◼   ► But it definitely did find some stuff.

00:46:32 ◼   ► And what I did, I think I might have said this already.

00:46:34 ◼   ► Apologies if I'm repeating myself.

00:46:35 ◼   ► But I just cloned the repo somewhere else.

00:46:38 ◼   ► And so this way, yeah, gave it its own little sandbox and let it run amok.

00:46:43 ◼   ► And I can choose to push that, you know, to main.

00:46:46 ◼   ► Or I can choose to just leave it on my computer.

00:46:47 ◼   ► No harm, no foul.

00:46:48 ◼   ► Yeah, not asking it to fix the bugs.

00:46:49 ◼   ► You're just saying find them.

00:46:50 ◼   ► And like I said, it's going to find stuff that's not bugs because it's not that smart, right?

00:46:55 ◼   ► But it will find legit bugs.

00:46:56 ◼   ► And the thing is, you can keep asking it.

00:46:58 ◼   ► Find any more bugs in this program.

00:46:59 ◼   ► Find any more bugs.

00:47:00 ◼   ► Find any performance issues.

00:47:02 ◼   ► Find any concurrency issues.

00:47:03 ◼   ► Find any memory safety issues.

00:47:04 ◼   ► You phrase it 75 different ways.

00:47:05 ◼   ► Just keep asking it questions and asking it to point you to things.

00:47:08 ◼   ► And all it's going to say is, I think this thing over here is wrong.

00:47:11 ◼   ► And it's going to be wrong a lot of the time.

00:47:13 ◼   ► It's going to be pointing to things that are not bugs, which is annoying.

00:47:16 ◼   ► Ask any open source maintainers getting bug reports from AI that are just not bug reports.

00:47:21 ◼   ► But for your own code, I as a programmer value very highly the ability to find even one bug out of like 10 things that it pointed out.

00:47:29 ◼   ► Because, hey, that's one bug you didn't know about before.

00:47:31 ◼   ► And when you fix it, it's gone.

00:47:33 ◼   ► And when it can't find any more legitimate bugs, does that mean there's no more bugs in your program?

00:47:37 ◼   ► Absolutely not.

00:47:38 ◼   ► There's always more bugs in your program.

00:47:40 ◼   ► Like, we're programmers.

00:47:40 ◼   ► We know there's always more bugs.

00:47:42 ◼   ► But this is one of the few tools I've ever seen, aside from like, you know, linters and memory safety checkers and stuff like that.

00:47:50 ◼   ► Like, all those things that are in Xcode can find like your Swift 6, you know, strict concurrency, finding places where there might be issues.

00:47:57 ◼   ► This is like that.

00:47:58 ◼   ► And I highly value tools like that.

00:47:59 ◼   ► So I hope these tools find a way to continue to exist in some reasonable form so that they can be used even for that targeted purpose.

00:48:05 ◼   ► Really quickly, I know we haven't given Marco a chance to talk, and I apologize.

00:48:08 ◼   ► But as an example of something that worked out super duper well for me, I think at this point I was using ChatGPT.

00:48:14 ◼   ► But for various uninteresting reasons, I was exposed to N8N, like the letter N, numeral 8, letter N, which is a workflow, like automation thing that you can self-host.

00:48:27 ◼   ► And I wanted to have a notification in the morning that would tell me what lamps thing the kids were doing.

00:48:35 ◼   ► So that's, shoot, library, art, music, shoot, I'm forgetting, sports, I guess?

00:48:42 ◼   ► No, PE and something else.

00:48:45 ◼   ► I forget.

00:48:45 ◼   ► And basically what special class they have each day.

00:48:49 ◼   ► And this is actually a really complicated problem to solve because it's not exposed anywhere by the school system or anything like that.

00:48:56 ◼   ► And so I built an N8N mostly by asking ChatGPT to do stuff, a whole thing that will, like, go and figure out what is the calendar for the school district.

00:49:06 ◼   ► Do they have school today?

00:49:07 ◼   ► Is it a half day?

00:49:08 ◼   ► Do they have a snow delay?

00:49:10 ◼   ► Or are they out because of snow?

00:49:12 ◼   ► And based on such and such a, like, reference date, count forward all the school days from that reference date to figure out, okay, this is day number three, which means Michaela has library and Declan has counseling.

00:49:22 ◼   ► And then send me a push notification at 645 while we're sitting down to breakfast so I can ask them, oh, are you excited for library and counseling and so on and so forth?

00:49:29 ◼   ► This is dumb.

00:49:30 ◼   ► This is very Casey.

00:49:31 ◼   ► I get that.

00:49:31 ◼   ► You don't need to make fun of me.

00:49:32 ◼   ► I am well aware.

00:49:33 ◼   ► Honestly, this sounds awesome.

00:49:34 ◼   ► Right?

00:49:35 ◼   ► I'm saying, like, this is a stupid thing.

00:49:38 ◼   ► It's a stupid thing because I could never frigging remember what their specials were each day.

00:49:42 ◼   ► And it's driving me crazy.

00:49:44 ◼   ► And so I built this thing, mostly with ChatGPT, that will help me do this and help me figure it out.

00:49:49 ◼   ► And so I'm just sitting there at breakfast, and then bloop, my phone says, hey, guess what?

00:49:53 ◼   ► Declan has counseling or whatever I just said, and Michaela has library.

00:49:56 ◼   ► It doesn't matter.

00:49:56 ◼   ► You get the idea.

00:49:57 ◼   ► And this is actually, again, very complicated because you're pulling in all these different pieces, and some of that is a testament to N8N, which is very cool and very, very weird, but very cool.

00:50:06 ◼   ► But a lot of this was me saying to ChatGPT, hey, I'm targeting N8N.

00:50:11 ◼   ► Can you write me a bunch of JavaScript that I'll plug into a node in N8N that will do this thing?

00:50:16 ◼   ► And I mostly told it, like, this is the step where I need you to parse out the school calendar and tell me whether or not today is a school day.

00:50:28 ◼   ► Okay?

00:50:28 ◼   ► And then the next node is, all right, given that today is a school day, count from this reference date and figure out how many school days there's been.

00:50:33 ◼   ► I forget exactly how it works.

00:50:34 ◼   ► We get the idea, right?

00:50:35 ◼   ► So I'm breaking down the major problem into micro problems and having ChatGPT figured out, and it took some back and forth, but I'll be damned if it didn't work, and it is freaking cool, and it solves a problem for me.

00:50:47 ◼   ► So in the one breath, I'll tell you, for CallSheet, I've had not great luck for the most part.

00:50:52 ◼   ► I had some good moments, but not great luck.

00:50:53 ◼   ► But for the N8N stuff, it was freaking amazing, you know?

00:50:56 ◼   ► And so I don't even know what I think about it right now.

00:51:00 ◼   ► Leaving aside the ethics, leaving aside the environmental impacts, just as like a tool, I definitely like it, but I'm not sure.

00:51:07 ◼   ► I can't figure out how much of this is overblown and how much isn't.

00:51:12 ◼   ► Sponsored this episode by DeleteMe.

00:51:14 ◼   ► DeleteMe makes it easy, quick, and safe to remove your personal data from hundreds of data broker sites online at a time when surveillance and data breaches are common enough to make everyone vulnerable.

00:51:23 ◼   ► DeleteMe does all the hard work of wiping you and your family's personal information from data broker websites.

00:51:30 ◼   ► So here's what these are.

00:51:31 ◼   ► These are these sites that, like, when you search in a search engine for somebody's name, that's all the sites that tell you, like, get their address, get their phone number, all these sites that will basically sell people's information to anybody who comes by looking for it.

00:51:43 ◼   ► There's hundreds of these sites out there, and they all have some kind of, like, opt-out or takedown procedure, but they're all different, and they make them kind of hard to find.

00:51:51 ◼   ► So you as an individual would never stand a chance at getting your stuff off of a meaningful number of these sites, and new ones, of course, are popping up all the time.

00:51:58 ◼   ► So even if you somehow did all hundreds of them out there, you'd need to constantly monitor and keep doing it for new ones.

00:52:04 ◼   ► So DeleteMe is a service that does that for you.

00:52:07 ◼   ► So they do all the hard work of figuring out all the data brokers that they can find out there, figuring out their opt-out and takedown procedures, and automating it for you.

00:52:17 ◼   ► So that way, you can just sign up for DeleteMe, you tell them what information you want to be removed, whether it's, like, your stuff, your family stuff, everything, and then they will go and automatically take it down from all the different data broker sites they know about, and as new sites come up, they'll take it down from them, too.

00:52:33 ◼   ► So you don't have to keep monitoring it.

00:52:35 ◼   ► They do it all for you.

00:52:36 ◼   ► They'll send you regular personalized privacy reports showing what they found and what they were able to take down.

00:52:41 ◼   ► So it works for you constantly, monitoring and removing the personal info you don't want on these data brokers.

00:52:47 ◼   ► So, take control of your data and keep your private life private by signing up for DeleteMe.

00:52:52 ◼   ► Now at a special discount for our listeners.

00:52:54 ◼   ► Get 20% off your DeleteMe plan when you go to joindeliteme.com slash ATP and use promo code ATP at checkout.

00:53:01 ◼   ► The only way to get 20% off is to go to joindeliteme.com slash ATP and enter code ATP at checkout.

00:53:08 ◼   ► That's joindeliteme.com slash ATP, code ATP.

00:53:13 ◼   ► Thanks to DeleteMe for sponsoring our show.

00:53:16 ◼   ► So, I've said a lot, and I apologize, Marco.

00:53:21 ◼   ► Please interrupt me.

00:53:22 ◼   ► Start talking so I don't keep talking for the next two hours.

00:53:25 ◼   ► You got it.

00:53:25 ◼   ► You can always count on me.

00:53:26 ◼   ► All right.

00:53:28 ◼   ► So, the way John led off earlier is correct.

00:53:31 ◼   ► This is all correct in some ways.

00:53:35 ◼   ► There was a great interview on Stratechery a few days back with Benedict Evans and Ben Thompson.

00:53:43 ◼   ► And one of the – they were talking about like, you know, basically will AI replace all of these like SaaS product companies?

00:53:51 ◼   ► And their argument and Benedict Evans' argument was no because, you know, what most apps really are, especially a lot of business apps, but like what most apps really are.

00:54:01 ◼   ► Think about like, you know, what is something like Salesforce, you know, it's – they're basically like databases with front ends.

00:54:07 ◼   ► And like that's what almost – that's a huge amount of software is like a database with a front end.

00:54:15 ◼   ► A huge amount of other software is basically a spreadsheet, but custom.

00:54:19 ◼   ► And when spreadsheets and databases came out and were made available with the personal computer revolution, they were very disruptive to a lot of jobs that came before them.

00:54:34 ◼   ► Ask accountants, you know, or ask – you know, rather ask calculators or computers.

00:54:39 ◼   ► What were they actually called computers?

00:54:40 ◼   ► The people who – spreadsheets made their job a lot less necessary.

00:54:45 ◼   ► But also we still have accounts and we still have, you know, people doing a lot of these jobs.

00:54:51 ◼   ► And we still have a lot of products that exist that are basically databases or customized spreadsheets despite everyone having access to database tools and spreadsheet tools.

00:55:01 ◼   ► Their argument was basically like, you know, AI is going to be a component of lots of software, but it's not going to replace the need for software.

00:55:09 ◼   ► And I think that's probably correct.

00:55:12 ◼   ► It is a little bit different than those in like the breadth of the types of problems it covers.

00:55:16 ◼   ► And the nature with which it solves them, you know, being kind of non-deterministic is a little bit unusual in the computing world.

00:55:24 ◼   ► Let's just say unreliable.

00:55:26 ◼   ► And certainly, you know, when you're looking at like, you know, businesses using AI in business roles, one of the biggest challenges is this stuff is constantly changing.

00:55:35 ◼   ► And so you can build a business process around calling, you know, the ChatGPT API or something.

00:55:41 ◼   ► But next month, whatever you built will break because something will change.

00:55:47 ◼   ► GPT 5.5 or 6.0 comes out and all of a sudden the integration you built, it works differently.

00:55:52 ◼   ► And now you have to rebuild it.

00:55:53 ◼   ► And there's going to be a lot of that tumult for a while, for a long time.

00:55:58 ◼   ► I think there will always be value in the market for customized software the same way there always has been value for it, even though people have spreadsheets and databases and they can build their own.

00:56:08 ◼   ► However, one major difference with AI is we are making the construction of these things even easier than it was before.

00:56:16 ◼   ► And I think to software developers, this is probably on the level of like the jump when we went from assembly code to compiled code and higher level languages.

00:56:29 ◼   ► That was a major jump in programmer productivity.

00:56:32 ◼   ► You can draw a lot of parallels to that.

00:56:34 ◼   ► It isn't a perfect analogy, but you can draw a lot of parallels to that kind of jump.

00:56:38 ◼   ► With AI, where I do think that the era of kind of hand customizing really nice code that you're making yourself is going to be, you know, commercially in the past.

00:56:52 ◼   ► Now, the same way, like, you know, if you are a custom furniture maker, there's still a market for custom furniture and you can you can be a really good woodworker and you can make really nice custom wood furniture for people.

00:57:07 ◼   ► But most people's furniture is, you know, large scale made by machines and flat packed into boxes at Ikea or whatever that made it a lot more affordable for a lot of people, which is good.

00:57:21 ◼   ► But if you were a furniture maker, it's like, well, hmm, the market just got a lot smaller once that once, you know, power tools and automation and factories started taking over that business.

00:57:31 ◼   ► You know, the AI revolution here with code generation in particular is that kind of moment for custom software.

00:57:38 ◼   ► Now, the good news is most software, most code that most people write is shuffling stuff around in a database or a spreadsheet or something similar.

00:57:50 ◼   ► Most code that most people write is really boring and really simple.

00:57:55 ◼   ► And that is easy to automate, as we're finding.

00:57:59 ◼   ► That is a major shift, but not unreasonable.

00:58:03 ◼   ► And certainly the industry can survive that.

00:58:06 ◼   ► It does change things.

00:58:08 ◼   ► It's going to keep changing things.

00:58:10 ◼   ► But most code that most people write in most apps most of the time is really boring.

00:58:14 ◼   ► And that is now much easier to automate.

00:58:18 ◼   ► The more difficult code is for a little while, at least.

00:58:22 ◼   ► I mean, maybe forever, but probably not forever.

00:58:24 ◼   ► For a little while, more difficult code or more high-level problems will still be generally better done by humans.

00:58:32 ◼   ► But that's probably not going to last that long.

00:58:35 ◼   ► I mean, we've had the AI tools for basically zero time, and they're already really good.

00:58:39 ◼   ► You know, where are they going to be in five years?

00:58:42 ◼   ► We don't know if they're going to plateau like the other uses.

00:58:45 ◼   ► I got some pushback last episode saying that, like, the chat interface had plateaued, but it actually is continuing to advance.

00:58:50 ◼   ► Not plateaued as in flatline, but the rate of advancement is much slower for those things than it is, right?

00:58:55 ◼   ► And so, like, I mean, what I'm saying about the coding thing is I haven't seen the rate of advancement slowing yet.

00:59:00 ◼   ► Surely it will at some point, unless we have some other advance, because these things don't scale forever, right?

00:59:06 ◼   ► But, you know, the line is still going up.

00:59:09 ◼   ► And as for your analogy with, like, assembly versus C and stuff, a couple things to say on that.

00:59:15 ◼   ► One, when we went from, like, assembly to, like, higher level languages, that resulted in tremendously more programmers and tremendously more code because it became more accessible.

00:59:28 ◼   ► Doing stuff in assembly is work that is only possible by a much smaller subset of people because it's so fiendishly difficult and unforgiving.

00:59:38 ◼   ► There's a reason we made higher level languages, and when we did, things became easier.

00:59:42 ◼   ► Manual memory management is more difficult than not having to do manual memory management.

00:59:47 ◼   ► In the days when we had to do manual memory management in C and you were allocating and freeing everything, there were fewer programmers.

00:59:54 ◼   ► Let me jump in.

00:59:55 ◼   ► Let me jump in right there.

00:59:56 ◼   ► You're exactly right.

00:59:57 ◼   ► But just in case you're not familiar, you know, I forget that not everyone that listens to this program is a developer.

01:00:02 ◼   ► So, assembly language is, and John will correct my piss-poor analogy here, or my definition here, but assembly language is basically a hop, skip, and a jump from ones and zeros.

01:00:11 ◼   ► Like, it is as close as you can get to actually telling the CPU, the processor, I would like you to add these two numbers and put the result in this piece of memory right here.

01:00:22 ◼   ► Like, it's very low level, and you have to think about a bunch of junk that the three of us don't generally have to think about.

01:00:27 ◼   ► And so, that's what makes it, like, you have to be the right kind of weirdo to enjoy assembly development.

01:00:33 ◼   ► It's difficult to do complicated things because you're thinking about so much stuff.

01:00:36 ◼   ► And every time you can remove things that you have to worry about, what has happened is not that, like, I will now, every programmer was an assembly programmer at one point.

01:00:44 ◼   ► Like, that was, back in the day, there were no high-level languages.

01:00:47 ◼   ► There was just assembly or even before that machine code where you're literally writing ones and zeros, which I have done, and it's not fun.

01:00:52 ◼   ► I probably had a piece of paper at that point or a punch card or whatever.

01:00:56 ◼   ► Anyway, and it's like, oh, all these people are going to be out of business because now Pascal exists, and you can just write words.

01:01:02 ◼   ► And it's, you don't, you're not writing assembly instructions, and it's, like, portable across different architectures.

01:01:06 ◼   ► And so, your skills of knowing this particular CPU assembly code, that's worthless.

01:01:09 ◼   ► You're out of a job.

01:01:10 ◼   ► There's not going to be any more programmers.

01:01:12 ◼   ► Just every time there has been an advancement in the art of programming, it has produced, it has resulted in us having more programs and more programmers.

01:01:20 ◼   ► And I don't think that will be any different with this new tool.

01:01:23 ◼   ► We will have more programs and more programmers.

01:01:26 ◼   ► It's just that what we call a programmer will be, you know, it's like, well, they're not a programmer.

01:01:30 ◼   ► They're just prompting an AI.

01:01:31 ◼   ► I was like, well, they're not a programmer.

01:01:32 ◼   ► They're just using C.

01:01:33 ◼   ► It does all the stuff for you.

01:01:34 ◼   ► You know, you have to know where the, they don't even know what the registers are.

01:01:37 ◼   ► They don't even know what CPU they're running on.

01:01:39 ◼   ► They don't even know who's allocating memory.

01:01:40 ◼   ► Oh, a garbage collector is going to come clean up for you.

01:01:42 ◼   ► That must be nice.

01:01:43 ◼   ► Like, that has always happened in every higher level language.

01:01:46 ◼   ► Now, obviously, this is different because those are all deterministic advances.

01:01:48 ◼   ► This is non-deterministic.

01:01:49 ◼   ► So I'm just saying this is an additional tool in our toolbox in the same way that having

01:01:53 ◼   ► like, I don't know, linters are a tool.

01:01:55 ◼   ► Linters are not a thing that is 100% accurate and they cite things that aren't actually problems

01:02:00 ◼   ► or whatever.

01:02:00 ◼   ► But you just use them as a tool to try to make programming more accessible and let us do a

01:02:05 ◼   ► better job.

01:02:05 ◼   ► As for your furniture analogy, which I think is a good one, but I would say that we have

01:02:10 ◼   ► long since been in the age of manufactured furniture before LLMs.

01:02:14 ◼   ► Like, forget about LLMs.

01:02:16 ◼   ► We went from the handcrafted furniture age into the mass-produced junkie furniture age, probably

01:02:22 ◼   ► in the 80s.

01:02:23 ◼   ► You're talking about in code, just to be clear.

01:02:25 ◼   ► If you've ever worked for a big company and seen the code behind your favorite product, it

01:02:30 ◼   ► looks a hell of a lot more like something stapled together in a crappy factory than a

01:02:35 ◼   ► does like a handcrafted, beautifully turned, solid wood thing made by artists and craftsmen.

01:02:40 ◼   ► Like, the only place that exists today, forget about AI.

01:02:44 ◼   ► The only place that exists today before AI was in small teams in small startups.

01:02:49 ◼   ► And even then, only if you're lucky.

01:02:51 ◼   ► Any big program does not look like that.

01:02:54 ◼   ► Like, maybe it started off as a beautiful hand-turned, you know, hand-assembled carved

01:02:59 ◼   ► piece of furniture made by artisans, but it doesn't stay that way.

01:03:03 ◼   ► And success makes it look a lot more like, it's like, imagine an Ikea sofa, not Ikea,

01:03:09 ◼   ► because I don't want to slam them.

01:03:10 ◼   ► Imagine like some mass-produced sofa held together with staples, but it's the size of Manhattan.

01:03:15 ◼   ► That's like the program, that's what's running Salesforce right now.

01:03:20 ◼   ► And into that fray, you throw one more tool, which is this, you know, LLMs that can help

01:03:26 ◼   ► you out with your coding process or whatever.

01:03:27 ◼   ► And honestly, I don't think it changes the nature of the product.

01:03:30 ◼   ► It is still a stapled together sofa the size of Manhattan.

01:03:33 ◼   ► It's just that people aren't manually putting the staples in anymore.

01:03:36 ◼   ► And who knows, maybe if this stuff continues to advance, we'll be able to shape that down

01:03:39 ◼   ► into something that looks closer to that hand-assembled piece of furniture.

01:03:43 ◼   ► But characterizing the pre-LLM age as we were all hand-making these beautiful turned pieces

01:03:48 ◼   ► of things on our lathes is not fair.

01:03:51 ◼   ► Even my dinky little programs are not that.

01:03:54 ◼   ► And they have like 30,000 lines of code.

01:03:56 ◼   ► Like, so that's why we programmers out here are excited by this.

01:04:00 ◼   ► Because again, if I can find a tool that will find me one additional bug that it couldn't

01:04:03 ◼   ► find before, that tool has value to me.

01:04:05 ◼   ► I think what we will see here is going to be very similar to those other advances.

01:04:11 ◼   ► You know, when those other advances came around, when we had higher level languages, well, you

01:04:15 ◼   ► know, every time you had programmers, I've been one of them, you've had programmers who have

01:04:20 ◼   ► said, I'm going to keep doing it the old way, either because I like it better, or because

01:04:26 ◼   ► I think I can do a better job.

01:04:28 ◼   ► Or there's some, you know, my code will perform better, or, you know, whatever it is.

01:04:31 ◼   ► And over time, all of those advantages for all of those steps, memory management, compiled

01:04:38 ◼   ► optimizations, or, you know, assembly optimizations, over time, all of those advantages have been

01:04:43 ◼   ► erased by other benefits or by just the sheer performance of the higher level tools.

01:04:47 ◼   ► Except perhaps that you liking it better, because that's up to you.

01:04:50 ◼   ► Yes, exactly.

01:04:51 ◼   ► And so that's why I think handwriting all of your code without having AI generate any of it

01:04:57 ◼   ► for you is going to be like having a woodworking hobby.

01:05:01 ◼   ► Like, you can do it.

01:05:03 ◼   ► And if you love coding, if you love the craft of it, if you love the process of it, you can

01:05:08 ◼   ► still do that.

01:05:09 ◼   ► But if you want to do this as a job, it's going to become increasingly difficult for you to be

01:05:17 ◼   ► competitive and relevant in that world if you're not using these higher level tools.

01:05:22 ◼   ► That's been true with all those other advances, and that's going to be true here, too.

01:05:26 ◼   ► The promise, though, here is that this is a bigger jump than any of those have been in

01:05:33 ◼   ► certain ways.

01:05:34 ◼   ► Not in every way, but in certain ways, it's a bigger jump.

01:05:37 ◼   ► And so the real promise of AI is not that it's going to replace all of us.

01:05:42 ◼   ► Although I do have concerns about, you know, junior level programmers coming up in the industry

01:05:47 ◼   ► right now.

01:05:47 ◼   ► I think that could be challenging for them, but we'll see, you know, it'll it'll shake

01:05:50 ◼   ► out.

01:05:50 ◼   ► Well, we have some questions about that in a future episode, so we will talk about that.

01:05:53 ◼   ► I know people have sent in a lot of stuff on that.

01:05:54 ◼   ► Yeah, and we'll see how that shakes out.

01:05:56 ◼   ► I mean, odds are they'll they will be just using these tools like everyone else.

01:05:59 ◼   ► But anyway, what the value here is, is not that we can replace all of the code that we

01:06:05 ◼   ► can and should write with AI.

01:06:07 ◼   ► It's that we can replace a lot of it and we can have it do a lot of things for us that don't

01:06:14 ◼   ► require perfect behavior all the time that don't require super high skill levels or super

01:06:20 ◼   ► high complexity levels yet.

01:06:22 ◼   ► We can have it do a ton of tasks that right now either burden us, things like writing

01:06:29 ◼   ► tests.

01:06:30 ◼   ► Maybe this will actually get me to write tests.

01:06:32 ◼   ► Marco's going to bring out the test.

01:06:33 ◼   ► Some people like writing tests, you know.

01:06:35 ◼   ► Some people like writing code.

01:06:37 ◼   ► I love writing code.

01:06:38 ◼   ► But, you know, anyway, so there's there's a lot of stuff that right now is a burden to

01:06:41 ◼   ► us that AI will be able to take off of our hands.

01:06:44 ◼   ► So that's a huge advance right there.

01:06:46 ◼   ► And then secondly, from that, which I was saying earlier about how, you know, as programmer

01:06:51 ◼   ► productivity has increased, we've had the need for more programmers.

01:06:54 ◼   ► The reason why that's happened is as making software has gotten cheaper, more people have

01:07:01 ◼   ► employed it to solve problems that that weren't worth doing when it was really expensive.

01:07:05 ◼   ► But when it got dramatically cheaper, some of those problems became worth doing with software.

01:07:10 ◼   ► Casey just did it with his N8N thing.

01:07:12 ◼   ► It was not worth doing when he if he had to write all that code himself, but now that

01:07:16 ◼   ► he has a tool that can help him suddenly becomes worth doing.

01:07:18 ◼   ► 100%.

01:07:19 ◼   ► And now you have, you know, a status board and a show bot and the reason or a tier list maker

01:07:24 ◼   ► rather.

01:07:24 ◼   ► And the reason why is because we like now things that were never worth making before might be

01:07:32 ◼   ► worth making.

01:07:33 ◼   ► Many of them will be worth making.

01:07:35 ◼   ► So first of all, that's great for basically every business and every hobby is like because

01:07:40 ◼   ► we can just make stuff now.

01:07:42 ◼   ► And as programmers, I don't know a single programmer who loves this craft, who doesn't have a bunch

01:07:49 ◼   ► of ideas they're never going to get to.

01:07:51 ◼   ► Now you can get to a bunch of them and that can be entire products.

01:07:55 ◼   ► It can be little things that just help you or it can be like, you know, the kind of side

01:08:00 ◼   ► apps that help you do your other stuff.

01:08:03 ◼   ► I know our friend underscore David Smith wrote about it recently on Mastodon, how he's made side

01:08:07 ◼   ► apps to help his business like admin control things or like internal tools to test out

01:08:13 ◼   ► ideas, just mini apps to try out ideas for things.

01:08:16 ◼   ► Yeah.

01:08:16 ◼   ► Or just things, things to kind of help automate something that like no one else but you is

01:08:20 ◼   ► going to need this, but you need it.

01:08:21 ◼   ► You know, I have tools with that.

01:08:23 ◼   ► I have like, you know, the shell scripts I've written to do like I have so much, so many

01:08:27 ◼   ► little tools like that.

01:08:28 ◼   ► And what AI can do is also make a lot of those things that just weren't worth making before

01:08:34 ◼   ► all of a sudden worth making.

01:08:36 ◼   ► And then finally, the major thing it can do, in addition to those other two major things,

01:08:40 ◼   ► another major thing it can do is it can keep working when you're not.

01:08:44 ◼   ► So you can have it like what John was saying, go find bugs in my app.

01:08:50 ◼   ► You can have it do that every night while you're asleep or doing anything else.

01:08:55 ◼   ► You can have it, you can assign it a task and be like, hey, go explore, you know, these

01:09:00 ◼   ► different, you know, different options for this problem I'm having or this idea I have.

01:09:04 ◼   ► Go explore that.

01:09:05 ◼   ► And then you can come back the next morning and see what it found and go through it.

01:09:08 ◼   ► There's already people who are doing this with great success playing with things like

01:09:12 ◼   ► cloud code.

01:09:12 ◼   ► Let me just give it some grunt work because it doesn't care that you're giving a grunt work.

01:09:19 ◼   ► You can give it the same grunt work every single night.

01:09:21 ◼   ► When it brings you six solutions in the morning, you can say all six of those suck.

01:09:28 ◼   ► Give me six more.

01:09:29 ◼   ► And it's not going to get upset or quit because it's not a person.

01:09:33 ◼   ► Think about what this opens up.

01:09:35 ◼   ► You can you can have an assistant doing a bunch of grunt work that costs you almost nothing.

01:09:41 ◼   ► And you don't have to worry about, like, are they miserable doing this because they're

01:09:47 ◼   ► not a person.

01:09:47 ◼   ► That's a huge opportunity for people like developers who are going to be able to harness this.

01:09:53 ◼   ► Think of all the different grunt work you can do.

01:09:56 ◼   ► Like and the biggest challenge to using these tools is not, you know, installing the software.

01:10:03 ◼   ► It's not paying for the 20 or 100 bucks a month.

01:10:06 ◼   ► The biggest challenge for existing professionals in the field like us to use these tools is

01:10:12 ◼   ► that we won't think of what to ask them to do.

01:10:13 ◼   ► That's the hardest part.

01:10:15 ◼   ► We need to we need to think of things to try to assign them.

01:10:18 ◼   ► Young people won't have this problem as young people come into the industry.

01:10:23 ◼   ► They're going to have grown up with these tools.

01:10:25 ◼   ► They won't have the preconceived notions of where the walls are that we have.

01:10:29 ◼   ► They'll just try a bunch of stuff and a lot of it will work.

01:10:32 ◼   ► That's the challenge for people, people like us, like, you know, us elder programmers

01:10:36 ◼   ► is that like we need to jump into this world and start getting ideas for what to have them

01:10:42 ◼   ► do.

01:10:42 ◼   ► By the way, if you want something that can generate a bunch of ideas for you for free,

01:10:45 ◼   ► LLMs are really good for that.

01:10:47 ◼   ► So you can literally go to ChatGPT or Gemini or Claude and you can say, give me 20 ideas for

01:10:54 ◼   ► what I can have Claude code or codex or whatever, what I can have it do in my existing code base

01:10:59 ◼   ► or a new code base that I am not thinking of.

01:11:01 ◼   ► You can literally have it tell you because you know what AI is really good for brainstorming

01:11:06 ◼   ► because AI is a BS generator and that's what brainstorming is, is generating BS.

01:11:11 ◼   ► So you can have it tell you how to use it and you can then at that point say, eh, these

01:11:18 ◼   ► don't sound very good.

01:11:19 ◼   ► Give me 50 more and it will happily give you 50 more and you can keep doing that because

01:11:26 ◼   ► it's not human.

01:11:26 ◼   ► Like that's the kind of enabling thoughts that we have to start training ourselves to

01:11:32 ◼   ► have if we're not going to get marginalized over time.

01:11:34 ◼   ► And again, like it's hard for like, you know, one of the posts that we mentioned earlier that

01:11:38 ◼   ► kind of hit me hard was the morning of our craft that like, yeah, it does kind of feel

01:11:43 ◼   ► like our craft is kind of dead and I do feel like I have built up all these skills that

01:11:51 ◼   ► are rapidly becoming marginalized in large ways.

01:11:56 ◼   ► I don't feel that at all.

01:11:57 ◼   ► I don't know.

01:11:57 ◼   ► We're not talking about that post, but I understand the sentiment where it's coming from.

01:12:01 ◼   ► But like if you've ever seen anyone use these tools to great effect, you need every single

01:12:06 ◼   ► one of the skills that I've gained as a professional programmer for 25 years.

01:12:09 ◼   ► Like those were all absolutely because if you think they're not give give one of these

01:12:14 ◼   ► tools to someone who's not a programmer and have them build an app and they can do it.

01:12:17 ◼   ► But it's going to be bad because they don't know like the skills, like the skills of the

01:12:22 ◼   ► programmer are not knowing what words to type.

01:12:24 ◼   ► Like you learn as you as you advance in your career and you learn multiple languages, you

01:12:28 ◼   ► eventually learn that, yeah, it is valuable to be an expert in a particular language or know

01:12:31 ◼   ► the details of any particular API.

01:12:33 ◼   ► But the real skills of a programmer is not at the level of languages or APIs or whatever.

01:12:38 ◼   ► The real skills of a programmer are at much higher level of how to organize programs, how

01:12:41 ◼   ► to how to address complexity, how to solve problems.

01:12:43 ◼   ► And yeah, the LMS can help a little bit with that.

01:12:46 ◼   ► But humans are just so far, so much farther ahead of them at this point.

01:12:49 ◼   ► So I don't I understand what they're saying.

01:12:52 ◼   ► Oh, I mourn my crap.

01:12:53 ◼   ► I want to write all my code myself.

01:12:54 ◼   ► Like, I think I feel like the skills that I'm most proud of are not the skills that involve

01:13:00 ◼   ► me typing particular API calls.

01:13:02 ◼   ► I wanted to call back very quickly to something that Marco said a minute ago, which I think

01:13:06 ◼   ► is important and I don't think I was explicit about.

01:13:08 ◼   ► And I would like to plus one what Marco said.

01:13:10 ◼   ► That N8N thing that I was talking about, I could have done that.

01:13:15 ◼   ► There is no freaking chance I would have wasted the time to do it.

01:13:19 ◼   ► Even though it would have made me happy, even though it would have been useful, even though

01:13:24 ◼   ► I'm more than capable of achieving it, it would have taken me hours.

01:13:28 ◼   ► And I just I don't think I would ever look at that task and be like, yes, this is worth

01:13:32 ◼   ► hours of my time rather than doing actual work or things that are actually productive.

01:13:36 ◼   ► You know, put the quote, scare quotes wherever you want.

01:13:39 ◼   ► There's no chance I would have done that.

01:13:40 ◼   ► And what what I think is unquestionably incredibly powerful in freeing, for lack of a better word,

01:13:47 ◼   ► about LLMs and the ability they have to write code specifically is that these sorts of things,

01:13:53 ◼   ► those things like underscores doing, like you had said, in this N8N thing and the tier list

01:13:58 ◼   ► thing.

01:13:59 ◼   ► You absolutely could have written that tier list, John, as you have said previously.

01:14:02 ◼   ► But would you have bothered?

01:14:03 ◼   ► Eh, probably not.

01:14:04 ◼   ► I mean, evidences was that I didn't, even though I exactly, even though I dislike the other site,

01:14:08 ◼   ► I just kept using the crappy one that was there just because it didn't take any time.

01:14:12 ◼   ► Right.

01:14:12 ◼   ► And so I think and you guys have each said this in your own way, but I just really want to

01:14:16 ◼   ► plus one it that that that that is what's so powerful to me about these LLMs and the agentic

01:14:23 ◼   ► coding or vibe coding, whatever you want to call it this week.

01:14:25 ◼   ► I think that's incredibly powerful and incredibly freeing.

01:14:28 ◼   ► And I don't think that's limited to professional software developers.

01:14:31 ◼   ► But I think that we are, to your point a moment ago, perhaps uniquely good at leveraging them

01:14:35 ◼   ► for these sorts of things.

01:14:36 ◼   ► I mean, I didn't we didn't address the entire world of things that LLMs can do, but we are

01:14:41 ◼   ► in a position to understand, to correctly assess how valuable are these tools for doing

01:14:47 ◼   ► programming tests.

01:14:47 ◼   ► Like, that's why I cited Steve Trouton-Smith.

01:14:49 ◼   ► Like, he's not a dummy.

01:14:49 ◼   ► He's a good programmer.

01:14:50 ◼   ► The more you know about program, like, I feel like you can trust his opinion.

01:14:54 ◼   ► This tool was useful for him.

01:14:56 ◼   ► I don't know how useful it is for other people in other industries or whatever.

01:14:59 ◼   ► I feel like there's lots of places that LLMs are super terrible and are not going to be

01:15:03 ◼   ► helpful at all.

01:15:04 ◼   ► But in this narrow, very narrow instance of using it in a particular way for particular things

01:15:08 ◼   ► for our profession, we can say that it does useful things.

01:15:12 ◼   ► Yeah.

01:15:12 ◼   ► And like, and, you know, Casey, you were just saying a minute ago, like, you know, whether

01:15:16 ◼   ► it's called, you know, Vibe Coding or Agenta Coding.

01:15:18 ◼   ► Well, in three months, it's just going to be called Coding.

01:15:21 ◼   ► Yeah, true.

01:15:22 ◼   ► True.

01:15:23 ◼   ► This is just what Coding is.

01:15:24 ◼   ► I don't know if that's going to happen in three months.

01:15:26 ◼   ► You're very optimistic, class pessimistic.

01:15:28 ◼   ► And I feel like I'm the Momento Mori guy.

01:15:31 ◼   ► You know, that's, if I'm remembering correctly, the person who would walk behind, like, the

01:15:35 ◼   ► emperor of Rome or whatever, as he goes on his parade and everyone celebrates him and

01:15:39 ◼   ► remind him that he's mortal, that he's going to die.

01:15:41 ◼   ► He's just behind them whispering that thing.

01:15:43 ◼   ► I feel like I'm the same way, as Marco said, here's the only problems.

01:15:46 ◼   ► The only problems we have with AI are, like, figuring out what to ask it and all these other

01:15:49 ◼   ► things.

01:15:49 ◼   ► Young people are better at it.

01:15:50 ◼   ► I didn't say that was the only problem with AI, to be clear.

01:15:53 ◼   ► Well, the problems that we were going to have with it.

01:15:54 ◼   ► But I just wanted to remind everybody.

01:15:56 ◼   ► I mean, I feel like we're still even in the Uber phase of this, where Uber had this amazing

01:16:00 ◼   ► service where you could get a car to come to you.

01:16:02 ◼   ► And they were just running at a loss for years and years until I could drive all the taxi

01:16:05 ◼   ► companies out of business.

01:16:06 ◼   ► But actually, the money we were paying for an Uber ride was not paying for enough to cover

01:16:10 ◼   ► the costs of the company and the drivers.

01:16:12 ◼   ► Remember all those years, right?

01:16:13 ◼   ► And start coming out of those years.

01:16:15 ◼   ► Like, now that we've killed the taxi companies, we can start turning up the price.

01:16:18 ◼   ► The prices we're currently paying for this are not covering the costs of the services we

01:16:24 ◼   ► are receiving.

01:16:24 ◼   ► Now, in theory, unlike the Uber drivers, which are kind of intractable, there's humans, there's

01:16:29 ◼   ► cars and while setting aside self-driving, in theory, the cost curve on inference and training

01:16:35 ◼   ► will go down or whatever.

01:16:36 ◼   ► We'll see how that goes.

01:16:37 ◼   ► So there's it's technology.

01:16:38 ◼   ► They're trying to drive the cost down, yada, yada.

01:16:39 ◼   ► But then, of course, there's the whole ethical, legal and moral problem, which is how do these

01:16:44 ◼   ► things come into being by, you know, consuming all this content without permission?

01:16:48 ◼   ► And where do the benefits of that go?

01:16:50 ◼   ► Oh, they go directly to open AI and not to all the people from whom they took the content.

01:16:55 ◼   ► You know, well, how is how are all the lawsuits from the New York Times and Disney and Paramount

01:17:00 ◼   ► and like, we don't know how that's going to go.

01:17:03 ◼   ► Like, it's very easy to get caught up into the amazing thing these tools can do for you

01:17:07 ◼   ► and forget about is this sustainable?

01:17:10 ◼   ► Is this affordable?

01:17:11 ◼   ► Is this ethical?

01:17:12 ◼   ► Will this create a world in which the things required to train these models are no longer

01:17:17 ◼   ► being produced?

01:17:18 ◼   ► And it's like a death spiral of disincentivization where everyone just wants to use the models,

01:17:23 ◼   ► but there's nothing for the models to train on except for the stuff that the models put

01:17:26 ◼   ► out and it just gets worse and worse.

01:17:27 ◼   ► We don't know.

01:17:28 ◼   ► That is still unsolved despite all these things.

01:17:30 ◼   ► But the whole point of this whole big thing, we keep going around in circles.

01:17:33 ◼   ► I know people hear this and they hear Marco talking to like Marco thinks there's no problems

01:17:36 ◼   ► and everyone should love AI and it's going to doom us all.

01:17:39 ◼   ► And they're like, it's all of this.

01:17:41 ◼   ► It's all the good things and it's all the bad things.

01:17:43 ◼   ► It's a complicated issue.

01:17:45 ◼   ► It's not simple.

01:17:46 ◼   ► I personally think it's not as simple as no one should ever use AI and it's not good

01:17:50 ◼   ► for anything.

01:17:50 ◼   ► I don't think it's that simple.

01:17:51 ◼   ► And it's also not as simple as this is the future.

01:17:55 ◼   ► We're all going to be doing this surrender and you'll be happy.

01:17:58 ◼   ► I don't think it's that either.

01:17:59 ◼   ► It's somewhere in between those extremes and everyone can decide where they think it is.

01:18:04 ◼   ► But I think we should be all be working to try to figure out like that's why I'm I'm

01:18:07 ◼   ► racking my brain thinking, is it possible like these tools are useful?

01:18:10 ◼   ► I want to use them.

01:18:11 ◼   ► But is it possible to get a tool like this where where it doesn't destroy the world of

01:18:16 ◼   ► open source software?

01:18:17 ◼   ► For example, like, is there a way to like, can we actually, you know, licensing the Stack

01:18:21 ◼   ► Overflow stuff legally, probably fine.

01:18:24 ◼   ► Ethically, maybe the people who contributed who didn't never foresaw this.

01:18:28 ◼   ► And so it's not great.

01:18:28 ◼   ► But like going forward, open source, can we change the licenses so that people can decide

01:18:32 ◼   ► if they want to allow their open source stuff to be used for training?

01:18:37 ◼   ► If they did that, it would make for a model that I think more people would feel comfortable

01:18:41 ◼   ► reusing or on the macro side of thing is, well, nobody cares about how ethical things

01:18:47 ◼   ► are.

01:18:47 ◼   ► We all buy shoes made by children in third world countries and nobody thinks about it.

01:18:51 ◼   ► And I think some people do think about it.

01:18:53 ◼   ► And I don't want to reproduce that in another industry.

01:18:55 ◼   ► So it's let's say it's a complicated issue.

01:18:58 ◼   ► And in summary, AI is a land of contrast.

01:19:02 ◼   ► Well, and again, like, and a lot of those are very valid concerns and questions.

01:19:06 ◼   ► My position is simply that, you know, A, there's huge utility to be had.

01:19:10 ◼   ► So we can't just write it off.

01:19:11 ◼   ► And B, you're not putting the genie back in the bottle.

01:19:13 ◼   ► So, you know, one thing that I was concerned about was, you know, as I was, you know, thinking

01:19:18 ◼   ► about using these tools myself, I thought, well, I'm not going to let it near overcast code

01:19:23 ◼   ► because among other things, overcast code contains pretty sophisticated audio processing

01:19:28 ◼   ► algorithms that I've written that took me a year to develop.

01:19:31 ◼   ► And I put a lot of work into all the, you know, the voice boost and smart speed algorithms

01:19:36 ◼   ► and everything to make mine work really well and sound really good and be very sophisticated

01:19:41 ◼   ► technically.

01:19:41 ◼   ► And I thought, I don't want any of these apps to like to find that and send it to their servers.

01:19:49 ◼   ► And no matter what I say, what if they're going to train on that?

01:19:51 ◼   ► And then anybody else can make.

01:19:52 ◼   ► They absolutely are.

01:19:54 ◼   ► And so here's the thing, though.

01:19:55 ◼   ► So I thought of that and I had that concern.

01:19:58 ◼   ► I'm not going to let this near my code.

01:19:59 ◼   ► But then I realized, like, it's going to be, you can probably already today go to any of

01:20:06 ◼   ► these agents or go to any chatbot and say, generate me some Swift code to do, you know,

01:20:11 ◼   ► this kind of processing on audio and make it sound really good.

01:20:14 ◼   ► And what else should I be looking at?

01:20:16 ◼   ► What other algorithms or what other techniques should I be using to make, to achieve these

01:20:19 ◼   ► things to make them sound good?

01:20:19 ◼   ► And it will generate it for you.

01:20:21 ◼   ► I know this because I have done that with other types of algorithms.

01:20:24 ◼   ► I have asked ChatGPT or Gemini or Claude to generate me code to do, you know, some kind

01:20:29 ◼   ► of relatively, I even asked it, like, here's the kind of problem I'm trying to solve.

01:20:33 ◼   ► What algorithms should I be looking at?

01:20:35 ◼   ► And it'll give me a nice overview of six different algorithms to solve this problem and what the

01:20:40 ◼   ► tradeoffs of each ones are, what the performance is.

01:20:42 ◼   ► And then I'll say, do you want me to generate this for you in Swift?

01:20:44 ◼   ► I'll say, yeah, sure.

01:20:45 ◼   ► And then it generates it for me in Swift.

01:20:46 ◼   ► And I think there's no way this is going to work.

01:20:48 ◼   ► And I try and it works.

01:20:49 ◼   ► Already, the value of, like, proprietary custom techniques and tricks that you do is plummeting.

01:20:57 ◼   ► I should be really upset about that.

01:20:59 ◼   ► And on some level, that is kind of a shame.

01:21:01 ◼   ► But also, on another level, this other problem I was trying to solve, I just got the answer

01:21:06 ◼   ► and I got to move on.

01:21:07 ◼   ► That's incredibly powerful.

01:21:09 ◼   ► That there's all these different algorithms that, like, you know, problems that I'm trying

01:21:13 ◼   ► to solve, I could have figured out the algorithm to do it the sophisticated way.

01:21:17 ◼   ► It would have taken me a day or more.

01:21:20 ◼   ► And I might have gotten the algorithm wrong.

01:21:22 ◼   ► And I might have had weird bugs as a result.

01:21:24 ◼   ► Or I might do something in a relatively straightforward, naive approach that might have a really efficient

01:21:30 ◼   ► algorithm that I'm not going to ever think of or know about, but the AI knows about it.

01:21:34 ◼   ► So the AI code is not only giving me the potential to make faster performing code by giving,

01:21:40 ◼   ► more sophisticated algorithms that cover more things with fewer bugs, but also it's allowing

01:21:45 ◼   ► me to make things faster because I'm not having to implement all of these tricky lines of code

01:21:52 ◼   ► myself.

01:21:52 ◼   ► I can actually have it generate the really complex, hairy parts of these kinds of things, and then

01:21:59 ◼   ► I can move on and I can work on higher value tasks.

01:22:02 ◼   ► And so it's taking over some of those really tricky things.

01:22:05 ◼   ► It's taking over some of the low-end stuff and some of the, like, shuffling around data code.

01:22:10 ◼   ► It's a huge win for programmers.

01:22:13 ◼   ► We will be, like, the people who stand to win the most from using these tools are, as John was

01:22:21 ◼   ► saying, experienced, good programmers.

01:22:24 ◼   ► We stand to win the most here.

01:22:26 ◼   ► It's a huge value creator for us.

01:22:28 ◼   ► So ideally, you know, once you are done morning, and I think that's a process that we're all going

01:22:37 ◼   ► through, once you're done morning, jump in.

01:22:40 ◼   ► I'm not morning.

01:22:40 ◼   ► I'm morning.

01:22:41 ◼   ► But I'm also going to jump in.

01:22:43 ◼   ► I just simply don't believe the argument in that article.

01:22:46 ◼   ► Like, I don't think there's really that much to mourn.

01:22:48 ◼   ► But, you know, anyway.

01:22:49 ◼   ► I mean, it's time for my momentum or anything you were mentioning, like, considering whether

01:22:54 ◼   ► you should allow this thing to look at your overcast code and deciding eventually you should

01:22:57 ◼   ► just do so.

01:22:58 ◼   ► It's nice that you got to make that decision and it didn't just take the code without your

01:23:00 ◼   ► knowledge or consent, which is the situation for every other piece of code that's in that

01:23:05 ◼   ► model, probably.

01:23:06 ◼   ► You are correct.

01:23:07 ◼   ► And also, it's irrelevant.

01:23:08 ◼   ► I think it's entirely relevant.

01:23:10 ◼   ► But when people only hear Marco's opinion on the show, apparently, and everything I say

01:23:15 ◼   ► doesn't count.

01:23:15 ◼   ► But anyway.

01:23:15 ◼   ► Oh, yeah.

01:23:16 ◼   ► John, let me assure you, that could not be further from the truth.

01:23:20 ◼   ► Have you ever read the email?

01:23:21 ◼   ► I swear to you, if you listen to the feedback from last episode, everyone was yelling at me

01:23:25 ◼   ► about things Marco said.

01:23:26 ◼   ► We are three different people.

01:23:27 ◼   ► I have a different position on this stuff, but it is a complicated issue.

01:23:30 ◼   ► Like, I'm saying there's no one simple answer.

01:23:32 ◼   ► Like, I'm not saying that I don't discount that person's morning articles.

01:23:35 ◼   ► That's how a lot of people are really feeling.

01:23:37 ◼   ► Like, the main point I want to get across with all this, with all this back and forth

01:23:40 ◼   ► on these things, is that it actually is complicated and not simply black and white.

01:23:45 ◼   ► And every time I see someone extreme in either position, it just makes me think that they're

01:23:49 ◼   ► not interested and they don't know enough about the topic because it actually is fairly

01:23:54 ◼   ► complicated and, you know, it's something that we're all going to have to grapple with.

01:23:59 ◼   ► And this is just in the narrow realm of programming, which is a tiny corner of the world of things

01:24:05 ◼   ► that people are using chat GPT for.

01:24:07 ◼   ► So, yeah, we've got a challenge ahead of us.

01:24:09 ◼   ► Hopefully, it's not like nuclear weapons, but, you know, maybe it's like the version of that

01:24:12 ◼   ► that does not destroy the entire planet.

01:24:14 ◼   ► Now, one final note on the AI stuff, and I know this has been going long and I'm sorry,

01:24:19 ◼   ► but you also wanted to call out that AI hardware use and the fact that it's slurping up, like

01:24:26 ◼   ► these LLM vendors are slurping up all the hardware in the world, has had some real world implications

01:24:31 ◼   ► so far in a lot of different ways.

01:24:33 ◼   ► But perhaps most recently, the Steam Machine and Steam Frame are delayed because of AI-driven

01:24:38 ◼   ► RAM and storage shortages and price increases.

01:24:41 ◼   ► From Valve, when we announced these products in November, we planned on being able to share

01:24:45 ◼   ► specific pricing and launch dates by now.

01:24:47 ◼   ► But the memory and storage shortages you've likely heard about across the industry have

01:24:50 ◼   ► rapidly increased since then.

01:24:51 ◼   ► Limited availability and growing prices of these critical components mean that we must revisit

01:24:55 ◼   ► our exact shipping schedule and pricing, especially around Steam Machine and Steam Frame.

01:24:59 ◼   ► This is part of the reason people hate AI, because they're like, the thing I wanted isn't going

01:25:03 ◼   ► to get here because AI is stealing all the chips to tell people to have a little Eliza bot tell

01:25:07 ◼   ► people about their feelings, which is apparently what the vast majority of the millions and

01:25:11 ◼   ► millions of people use AI every day are doing with it.

01:25:13 ◼   ► And lots of people think that is both not a good idea and bad for people and a waste of

01:25:20 ◼   ► electricity and time and resources.

01:25:21 ◼   ► And now I can't get my Steam Machine.

01:25:23 ◼   ► So again, anti-AI sentiment, like I get where it comes from.

01:25:26 ◼   ► Like it's not, it's not, you know, it's not frivolous.

01:25:29 ◼   ► They're not against AI because they're just mean and someone is paying them to be against

01:25:32 ◼   ► AI.

01:25:33 ◼   ► People, there's a lot of things, you know, not in AI's favor, especially since it's like,

01:25:38 ◼   ► you know, huge amounts of VC and all the benefits going to the small number of companies

01:25:42 ◼   ► and they're doing things in an unsustainable way.

01:25:44 ◼   ► And when this bubble pops, it's going to destroy all our retirement accounts and all that terrible

01:25:48 ◼   ► stuff.

01:25:48 ◼   ► And yet none of that, and yet none of that negates the fact that it actually can do some useful

01:25:53 ◼   ► things in a tiny subset of stuff.

01:25:55 ◼   ► And I think we as a people should find a way to extract the good while avoiding as much

01:26:00 ◼   ► as the bad as we can.

01:26:01 ◼   ► We are sponsored this week by Quince.

01:26:05 ◼   ► A well-built wardrobe is about pieces that work together and hold up over time.

01:26:08 ◼   ► And that's what Quince does best.

01:26:10 ◼   ► Premium materials, thoughtful design, and everyday staples that feel easy to wear and easy to

01:26:14 ◼   ► rely on, even as the weather shifts.

01:26:16 ◼   ► Quince has the everyday essentials that you'll love and with quality that lasts.

01:26:21 ◼   ► There are organic cotton sweaters, polos for every occasion.

01:26:25 ◼   ► You know I love a good polo.

01:26:26 ◼   ► Lighter jackets that keep you warm in the changing season.

01:26:28 ◼   ► The list goes on.

01:26:30 ◼   ► Quince works directly with top factories and cuts out the middle people.

01:26:33 ◼   ► So you're not paying for brand markup, just quality clothing.

01:26:36 ◼   ► Everything is built to hold up to daily wear and still look good season after season.

01:26:40 ◼   ► Plus, and this is my favorite part, they only partner with factories that meet rigorous

01:26:44 ◼   ► standards for craftsmanship and ethical production.

01:26:46 ◼   ► So recently, Quince sent me a couple of pieces.

01:26:49 ◼   ► They sent, actually for Michaela, they sent an organic cotton long sleeve fit and flare pocket

01:26:54 ◼   ► dress, which has not only a couple of pockets, but it also has these adorable little like

01:26:58 ◼   ► two-tone hearts all over it.

01:27:00 ◼   ► Michaela genuinely loves this dress.

01:27:01 ◼   ► And then for me, they sent comfort stretch traveler five pocket pants slim cut.

01:27:06 ◼   ► And I love these things by genuine sheer coincidence.

01:27:10 ◼   ► I was wearing them earlier today because we'll probably talk about it on the show, but I needed

01:27:14 ◼   ► to look a little fancier for something.

01:27:16 ◼   ► And I wanted to have something that you could dress up.

01:27:19 ◼   ► And these pants very much look dressy, but feel casual, if that makes sense, which is a

01:27:25 ◼   ► compliment, let me assure you.

01:27:27 ◼   ► They're a little bit stretchy and they fit really nicely.

01:27:31 ◼   ► And I feel like I look good in them.

01:27:32 ◼   ► And that's probably the best thing of all, right?

01:27:35 ◼   ► Because if you feel like you look good, then other people will probably think that too.

01:27:38 ◼   ► So what are you going to do?

01:27:40 ◼   ► You're going to refresh your wardrobe with Quince.

01:27:42 ◼   ► Go to Quince.com slash ATP for free shipping on your order and 365 day returns.

01:27:47 ◼   ► Quince is now available in Canada as well.

01:27:49 ◼   ► That's Q-U-I-N-C-E dot com slash ATP.

01:27:52 ◼   ► Free shipping and 365 day returns.

01:27:55 ◼   ► Quince.com slash ATP.

01:27:57 ◼   ► Thank you to Quince for sponsoring the show.

01:27:59 ◼   ► We've known for a while that Ferrari is going to be making an EV.

01:28:08 ◼   ► I don't think we knew that the interior was going to be designed by Love From, which you know.

01:28:13 ◼   ► We did know that.

01:28:14 ◼   ► And we knew it in one of the dumbest ways, which I think I would comment on this show.

01:28:18 ◼   ► Like months ago, someone was doing like some Johnny Ive interview and they mentioned in

01:28:22 ◼   ► an offhanded manner that Johnny Ive was working on a Ferrari steering wheel.

01:28:25 ◼   ► And I think it was like a New Yorker story.

01:28:28 ◼   ► And I think I remember commenting the show of like, like the fact that that's in this

01:28:32 ◼   ► article, what are they talking about?

01:28:34 ◼   ► Because the Ferrari steering wheels of the time sucked.

01:28:37 ◼   ► So I'm like, that surely can't be a Johnny Ive thing.

01:28:39 ◼   ► But did he just out Johnny Ive and the fact they're working with Ferrari or have they already

01:28:43 ◼   ► worked with Ferrari?

01:28:44 ◼   ► And I just forgot about it until this came back.

01:28:45 ◼   ► I'm like, oh, this is the thing that guy spoiled like a year and a half ago.

01:28:49 ◼   ► Good job, dude.

01:28:50 ◼   ► So this is the Ferrari Luce, I would assume.

01:28:53 ◼   ► You got it.

01:28:54 ◼   ► Okay, good.

01:28:54 ◼   ► I'm one for a million.

01:28:57 ◼   ► But anyway, this is their first EV and they, I guess a couple of months back or something

01:29:03 ◼   ► like that, they announced all, a lot of the tech behind it.

01:29:06 ◼   ► Like they showed, they showed like the whole chassis with nothing, just the, like the chassis,

01:29:10 ◼   ► the motors, the battery, but no other part of the car.

01:29:13 ◼   ► So now in the last week, we've seen a lot of the cabin, not a full cabin, mind you, but

01:29:19 ◼   ► a lot of the like interfaces.

01:29:20 ◼   ► They would never show you the full cabin.

01:29:21 ◼   ► Why would they do that?

01:29:22 ◼   ► Right.

01:29:22 ◼   ► But there was, we've seen the interfaces, the steering wheel, the center display.

01:29:26 ◼   ► Next month, we're going to see a headlight.

01:29:27 ◼   ► Right.

01:29:28 ◼   ► Wait, are you, you might be serious.

01:29:30 ◼   ► Are you serious?

01:29:30 ◼   ► No, I'm not serious.

01:29:31 ◼   ► Okay.

01:29:31 ◼   ► Because the way they've been rolling this out, I didn't know.

01:29:33 ◼   ► I mean, you can read that, read the little blurb from the Verge.

01:29:35 ◼   ► It summarizes it well.

01:29:36 ◼   ► All right.

01:29:37 ◼   ► So from reading from the Verge, Ferrari released the first interior images of the company's

01:29:41 ◼   ► first all electric supercar called the Ferrari Luce, which is light in Italian.

01:29:45 ◼   ► This is the second time the Italian automaker has teased the Luce, which was formerly the

01:29:50 ◼   ► Electrica.

01:29:51 ◼   ► I hope I got that right.

01:29:52 ◼   ► Without showing us the actual car or even a silhouette, but the interior images should

01:29:56 ◼   ► suffice given the bold face name of the designer, Johnny Ive.

01:29:59 ◼   ► Ferrari decided to outsource the work of designing Luce's interior to Ive and his partner, Mark

01:30:03 ◼   ► Newson, who together run the design shop Love From.

01:30:06 ◼   ► Ferrari and Love From have been quietly collaborating for five years, and Luce is the first time

01:30:11 ◼   ► we're seeing the results.

01:30:12 ◼   ► The tech underpinnings were revealed last year in Italy.

01:30:15 ◼   ► The interior will, excuse me, the exterior will debut in May of 2026.

01:30:19 ◼   ► Well, there's a timeline for you.

01:30:21 ◼   ► We'll eventually see the car, but this is the weirdest car rollout I've ever seen.

01:30:24 ◼   ► It's still going to show us the chassis, then pieces of the interior, and then in 2026,

01:30:29 ◼   ► we'll show you the car.

01:30:29 ◼   ► Right.

01:30:30 ◼   ► So I would like to just state right up front, I watched a fascinating video.

01:30:36 ◼   ► by Jordan Golson from PRDNDL, which is, you know, park reversed, a neutral drive, low.

01:30:42 ◼   ► Anyways, I'd not heard of that publication before, but this dude was really good and did

01:30:46 ◼   ► like a 20-minute video walking through all the different pieces, which they did in the, what

01:30:50 ◼   ► is it, the Transamerica Tower in San Francisco?

01:30:52 ◼   ► Somewhere in San Francisco.

01:30:53 ◼   ► Yeah.

01:30:53 ◼   ► This was basically like the reason you're seeing this.

01:30:55 ◼   ► There's a web page of Ferrari, but also Love From did basically press.

01:30:57 ◼   ► Like they invited press to come into this room where they show all the things.

01:31:00 ◼   ► It's very much like an Apple press event where everything's set up on little tables and stuff.

01:31:04 ◼   ► And you can, you know, and they have to talk to the creators and, you know, it's, it's in

01:31:07 ◼   ► a very Apple style press event.

01:31:09 ◼   ► Right.

01:31:10 ◼   ► And so, uh, Jordan has a video, which I definitely recommend also a blog post if you prefer that.

01:31:15 ◼   ► Uh, Ferrari's website actually is really, really good.

01:31:19 ◼   ► Uh, and I really enjoyed looking through that.

01:31:21 ◼   ► It had a bunch of photos and whatnot.

01:31:22 ◼   ► Um, I'm going to come out right out and say it.

01:31:26 ◼   ► I'm so sorry, John.

01:31:27 ◼   ► I really like it.

01:31:29 ◼   ► I think it looks really good.

01:31:30 ◼   ► I don't know why you're apologizing to me, but like, like the second I saw this, I, I knew

01:31:35 ◼   ► exactly what I thought of it, but I'm like, well, hold off.

01:31:38 ◼   ► Maybe you just have it.

01:31:39 ◼   ► And then I watched all the videos and all the things or whatever.

01:31:41 ◼   ► And just my opinion got stronger and stronger.

01:31:42 ◼   ► And it's, it's founded in, uh, remember we talked about CarPlay Ultra, whatever it's

01:31:48 ◼   ► called.

01:31:48 ◼   ► Uh, and I think all, I think it was just CarPlay Ultra.

01:31:51 ◼   ► We talked about it a lot and trying to figure out what's the deal with this, where Apple takes

01:31:54 ◼   ► over all the screens and your car and you have to collaborate with the car vendor and

01:31:57 ◼   ► they have a way to customize it or whatever.

01:31:58 ◼   ► I'm like, how is this going to work?

01:32:00 ◼   ► And do car vendors want their, their car?

01:32:01 ◼   ► Like, how are they going to make it look like Aston Martin or like whatever company and stuff.

01:32:05 ◼   ► And the pitch for directly from Apple's mouth of what is the deal with CarPlay Ultra is we

01:32:12 ◼   ► have this product.

01:32:12 ◼   ► And if you use this product, the whole idea is we will make something that is a blend of

01:32:18 ◼   ► your brand and Apple's brand.

01:32:21 ◼   ► So 50, 50, I guess what maybe Apple would think are 51, 49 or whatever, but like, that's

01:32:26 ◼   ► the pitch.

01:32:27 ◼   ► The inside of your car with CarPlay Ultra will be half you and half Apple.

01:32:32 ◼   ► It'll be our child together.

01:32:34 ◼   ► We'll mix our genetic material, right?

01:32:37 ◼   ► This Ferrari interior made by Love From is half Ferrari and half Love From.

01:32:44 ◼   ► And I think that's wrong.

01:32:47 ◼   ► I think the Ferrari interior should be a hundred percent Ferrari and that a designer should say,

01:32:54 ◼   ► we will find a way to make the most Ferrari interior you have ever seen.

01:32:59 ◼   ► But when I look at this, I say, this is not a 100% Ferrari interior.

01:33:04 ◼   ► This is a 50% Love From interior and a 50% Ferrari interior.

01:33:08 ◼   ► And I, that's not the approach I think that you should take with this interior.

01:33:12 ◼   ► And worse, Love From is very different from Ferrari in terms of what are the adjectives

01:33:20 ◼   ► that you would use to describe that.

01:33:21 ◼   ► Like, it's not harmonious.

01:33:23 ◼   ► This does not look like a Ferrari interior.

01:33:26 ◼   ► This looks like an Apple car interior that Ferrari drove by and splashed a bunch of paint

01:33:33 ◼   ► over to make it look kind of, and some, and some logos to make it look like Ferrari.

01:33:37 ◼   ► And I find that fairly heartbreaking, but I will say though, the chassis and everything you

01:33:42 ◼   ► can see, this is a four seat car.

01:33:43 ◼   ► So let's be honest, it's probably just going to be some stupid Ferrari SUV that I won't care

01:33:46 ◼   ► about, but yeah, probably.

01:33:48 ◼   ► But you know, like it's, I, it's, it's heartbreaking to me because there's a lot of beautiful design

01:33:54 ◼   ► in this and a lot of beautiful work that is like, you look at it, if you're, if you're

01:33:58 ◼   ► just familiar with Johnny Ive, you know, like they have so many Apple designers there.

01:34:01 ◼   ► You look at this and you're like, yep, nope, they did.

01:34:03 ◼   ► That's the thing.

01:34:04 ◼   ► That's what Johnny Ive does.

01:34:05 ◼   ► There's rounded rectangles, there's beautifully machined parts, there's aluminum, there's

01:34:08 ◼   ► glass, there's clever ideas, there's attention to detail.

01:34:11 ◼   ► There's, there's all that stuff.

01:34:12 ◼   ► It's too damn bad.

01:34:14 ◼   ► It doesn't look like a Ferrari interior.

01:34:15 ◼   ► I get that.

01:34:17 ◼   ► I really do.

01:34:18 ◼   ► And I think part of the reason I don't mind that is because I don't have the unabashed

01:34:22 ◼   ► love for Ferrari that you do.

01:34:24 ◼   ► Not to say that I dislike Ferraris.

01:34:25 ◼   ► I mean, they're.

01:34:26 ◼   ► Oh, and to be clear, recent Ferrari interiors have been terrible, but they've been terrible

01:34:30 ◼   ► in a Ferrari way.

01:34:31 ◼   ► Well, I mean, it's, to me, it's kind of like if, if this were a, a Lamborghini, which I think

01:34:38 ◼   ► trades on the fact that it's weird and loud and obnoxious.

01:34:41 ◼   ► No, that, that, that's the perfectly perfect example I was going to do.

01:34:44 ◼   ► Lamborghini.

01:34:45 ◼   ► What do you, give me some adjectives to describe a Lamborghini, interior or exterior?

01:34:48 ◼   ► Loud, obnoxious, angular, you know.

01:34:51 ◼   ► Angry, dangerous.

01:34:53 ◼   ► Yep.

01:34:53 ◼   ► Yep.

01:34:54 ◼   ► Evil.

01:34:54 ◼   ► Eh, maybe.

01:34:56 ◼   ► That's, that's, that's Lamborghini's styling and brand.

01:34:59 ◼   ► Inside of Lamborghini's look like the inside of a freaking transformer from the Michael Bay

01:35:02 ◼   ► transformer movies, the outside looks like them too, right?

01:35:05 ◼   ► That for you, like it or hate it.

01:35:07 ◼   ► That's the, the, the, the, the, the Countach for crying out loud.

01:35:10 ◼   ► That's the brand.

01:35:11 ◼   ► Can you imagine?

01:35:12 ◼   ► And the Ferrari is not that exact same thing, but like this interior has no danger, has no

01:35:18 ◼   ► menace, has no, no sexiness to it.

01:35:22 ◼   ► No, no sensual Italian sexy.

01:35:25 ◼   ► It's like, it's a Ferrari.

01:35:26 ◼   ► The inside, this looks like a, inside of an electric Fiat with Ferrari logos on it.

01:35:30 ◼   ► I'm sorry.

01:35:30 ◼   ► I think, I think the reality is between you and me.

01:35:33 ◼   ► I don't, I think you're not giving it enough credit, but I don't necessarily disagree with

01:35:38 ◼   ► your thesis.

01:35:38 ◼   ► It, it definitely does not have that.

01:35:41 ◼   ► What is it?

01:35:42 ◼   ► Joie de vivre.

01:35:43 ◼   ► I don't, I don't speak French.

01:35:43 ◼   ► It's a love child of Love From and Ferrari, but yeah, Love From's genes dominated.

01:35:48 ◼   ► It does.

01:35:49 ◼   ► But I got to tell you, I really think they did a good job with it.

01:35:52 ◼   ► Let's leave aside the fact that it's a Ferrari.

01:35:54 ◼   ► If, if you can divorce yourself of that piece of knowledge, I think it's really well done.

01:35:58 ◼   ► So the way this works is there's a, a screen behind the steering wheel, which they're calling

01:36:05 ◼   ► the binnacle.

01:36:05 ◼   ► I presume that's a common term.

01:36:07 ◼   ► I've not heard it in this context before.

01:36:09 ◼   ► I just, I just, I just blame, blame, uh, Love From for coming up.

01:36:12 ◼   ► Oh, sure.

01:36:13 ◼   ► Um, but there's a binnacle, a binnacle is not a Love From term.

01:36:16 ◼   ► It is in the car industry.

01:36:17 ◼   ► So the binnacle has three gauges, which, uh, are their screen.

01:36:22 ◼   ► The gauges are screens, but they might that, I don't know if this one does or not, but there's

01:36:26 ◼   ► other gauges that have physical real life hands.

01:36:28 ◼   ► The center dial has a physical hand.

01:36:31 ◼   ► There you go.

01:36:31 ◼   ► Dial thing.

01:36:32 ◼   ► Yeah.

01:36:33 ◼   ► Yeah.

01:36:33 ◼   ► And then around the three gauges is a screen, but it's all really well integrated.

01:36:37 ◼   ► But the point is there's a mixture, at least some modicum of, of physicality to it.

01:36:43 ◼   ► And then the steering wheel has the flattened bottom, which is John's favorite thing in the

01:36:49 ◼   ► world.

01:36:49 ◼   ► Um, there are no stalks there.

01:36:51 ◼   ► The turn signals are, it's a, it's a T steering wheel or a three, three spoke wheel.

01:36:55 ◼   ► The turn signals are inside the spokes that, that go laterally.

01:37:00 ◼   ► Then below those spokes, there's a couple of control, uh, pods, if you will.

01:37:04 ◼   ► And, uh, then on the center, they have basically an iPad, which I don't love that.

01:37:11 ◼   ► It's very rounded corners.

01:37:12 ◼   ► Yes.

01:37:13 ◼   ► Very rounded corners.

01:37:14 ◼   ► I do love that.

01:37:15 ◼   ► They have a little bar in front of it that you can use as a wrist rest as you're operating

01:37:20 ◼   ► the screen.

01:37:20 ◼   ► I don't know how I feel about the fact that you can tilt the screen toward or away from

01:37:26 ◼   ► the driver.

01:37:27 ◼   ► Like I get the idea here, but I don't know if I like it.

01:37:29 ◼   ► Um, but in the corner of the screen is a dial, a rounded dial that is separate from the rest

01:37:36 ◼   ► of the screen inset within the screen.

01:37:38 ◼   ► And it has physical hands on it.

01:37:40 ◼   ► And by default it can show, I think it's like a stopwatch and there's a couple of buttons

01:37:43 ◼   ► on the outside of the screen, physical buttons on the outside of the screen to like start and

01:37:47 ◼   ► lap, you know, hit the, hit the start, stop and lap buttons.

01:37:49 ◼   ► If it's a stopwatch or it can be a clock.

01:37:51 ◼   ► And because what's behind these physical hands is another screen, the HVAC controls are in the

01:37:56 ◼   ► center screen, but there are a fair amount of physical controls, uh, the center consoles where

01:38:03 ◼   ► the gear shift is, which is very interestingly done.

01:38:05 ◼   ► And I think the windows might be there or something like that.

01:38:08 ◼   ► That's where you insert the key.

01:38:09 ◼   ► If you so choose, I really think leaving aside the fact that I don't really think you're wrong,

01:38:14 ◼   ► John, that this doesn't exactly scream Ferrari to me.

01:38:16 ◼   ► I do think this is really well thought out and looks really, really cool.

01:38:20 ◼   ► Yeah.

01:38:20 ◼   ► Well, I mean, it, it looks like, I mean, I don't care about Ferrari the way John does, but

01:38:25 ◼   ► like this looks very much like a designer who idolized Dieter Rams made a car interior.

01:38:35 ◼   ► Yeah, very much so.

01:38:37 ◼   ► You know, Dieter Rams was a really, is a really famous designer for good reason.

01:38:42 ◼   ► He's an amazing designer.

01:38:43 ◼   ► Like all the, all those classic mid-century brawn products, like that's that level of design.

01:38:49 ◼   ► Like it's, it's a really great style.

01:38:51 ◼   ► It has, it has evolved over time to represent like, and, and Johnny Ive, uh, you know, Johnny

01:38:57 ◼   ► Ive is a very good designer, but when, but you can see his influences, like with the Apple

01:39:01 ◼   ► watch, he borrowed a lot from his Patek Philippe Nautilus.

01:39:05 ◼   ► Um, with, with many Apple products, he borrows a lot from Braun, um, and from Dieter Rams, like,

01:39:12 ◼   ► cause it's, it is a really good style.

01:39:14 ◼   ► This looks like that.

01:39:17 ◼   ► If the logo on the steering wheel was not the Ferrari logo.

01:39:20 ◼   ► Just put an Apple logo there.

01:39:22 ◼   ► It could be, well, it's, there's too many buttons for it to be Apple, but it could be any other

01:39:27 ◼   ► car brand.

01:39:28 ◼   ► Like, you know, it, it may be some modern brand that nobody has that much of an opinion about

01:39:32 ◼   ► like Polestar.

01:39:32 ◼   ► Like if that could be a Polestar logo and everybody would be like, okay.

01:39:35 ◼   ► No matter, no matter what logo you put there, I would have been able to tell you, do you

01:39:38 ◼   ► think, uh, people who made this are, you know, like Apple's design and be like, yeah, these

01:39:43 ◼   ► people obviously love everything Apple is doing.

01:39:45 ◼   ► Would it surprise you to learn that it was actually Apple people?

01:39:48 ◼   ► No, it would not surprise me.

01:39:49 ◼   ► Well, but I mean, honestly, I think the more Apple like design is like what Tesla and Rivian

01:39:54 ◼   ► do, where they try to minimize physical controls and have everything on a small number of screens.

01:39:58 ◼   ► I don't, I don't think so.

01:39:59 ◼   ► I think a photograph of this interior makes me, screams Apple to me more than any Tesla interior.

01:40:04 ◼   ► I mean, you can certainly see the resemblance, but you can also see like, okay, there's a

01:40:08 ◼   ► lot of physical controls.

01:40:09 ◼   ► They're done.

01:40:09 ◼   ► They look beautiful, but they also, the, the energy, the style, the, like the, like that,

01:40:18 ◼   ► that kind of like, you know, sex appeal of a high end sports car.

01:40:22 ◼   ► That's not this.

01:40:23 ◼   ► This is nice German design done well by Johnny Ive.

01:40:29 ◼   ► Like that's, that's what that looks like.

01:40:30 ◼   ► It does not look like Italian sports car.

01:40:34 ◼   ► So I, I agree, John, like I, even though I don't care about Ferrari and Lamborghini.

01:40:38 ◼   ► And even though there are so many references to old Ferraris, like that's the thing about

01:40:42 ◼   ► this.

01:40:42 ◼   ► Like, it's like, how can you say it doesn't look like a Ferrari?

01:40:44 ◼   ► Like that steering wheel is directly ripped out of an old Ferrari and those air vents and

01:40:47 ◼   ► this and that.

01:40:47 ◼   ► Ferrari's current styling, forget about the interiors, Ferrari's current styling trend

01:40:54 ◼   ► is a similar example in that they reference their old stuff like crazy.

01:40:58 ◼   ► And you're just like, how can you say this doesn't look like a Ferrari?

01:41:00 ◼   ► Let me show you all the elements that reference old classic Ferraris.

01:41:03 ◼   ► It's like, yeah, but like the whole thing cohesively has to say to me Ferrari.

01:41:08 ◼   ► And I think the exterior is still kind of say Ferrari, even though I think they're ugly.

01:41:11 ◼   ► They're going through an ugly phase.

01:41:12 ◼   ► It's difficult for all of us, but they, they don't look, the outside of their cars would

01:41:18 ◼   ► never be mistaken for a Fiat.

01:41:20 ◼   ► Let's put it that way.

01:41:20 ◼   ► Like you can't just take out the Ferrari badge on the outside of like the new Testarossa and

01:41:25 ◼   ► put on a, you know, like a badge from BMW.

01:41:28 ◼   ► Like no one would believe it.

01:41:29 ◼   ► Like it's, it's clearly it's, I'm not sure if people identify as Ferrari, but there's no

01:41:33 ◼   ► way in hell they'd think it was a Volkswagen or something like it's, it's something.

01:41:37 ◼   ► Whereas this interior is just, it's a, it's a Johnny I've special, like, and getting to

01:41:40 ◼   ► the specifics of this, by the way, like setting aside, Oh, you're upset.

01:41:43 ◼   ► Doesn't look like Ferrari or whatever.

01:41:44 ◼   ► Just in the specifics, there's a couple of areas where they did well and where areas where

01:41:50 ◼   ► they didn't Casey already mentioned, uh, the physical controls Ferrari itself did capacitive

01:41:55 ◼   ► controls on its steering wheels for way too long.

01:41:56 ◼   ► Everybody hated it.

01:41:57 ◼   ► It takes them so long to change it.

01:41:59 ◼   ► They've like, this is this year we're getting rid of physical controls in the steering wheel.

01:42:01 ◼   ► And they, they got rid of the physical start button, but they left other capacitive controls.

01:42:05 ◼   ► It's like, no, for right, please just everything physical.

01:42:07 ◼   ► So finally in 2026, they'll roll out a car, which reverts to their previous thing, which

01:42:13 ◼   ► is like, let's give me actual buttons on the steering wheel.

01:42:15 ◼   ► So that's good.

01:42:16 ◼   ► I applaud them for that.

01:42:17 ◼   ► Um, the clever thing with the dials and the screens, lots of other people have done similar

01:42:20 ◼   ► things.

01:42:21 ◼   ► This is obviously the much more expensive, much nicer looking Johnny I version of that idea,

01:42:25 ◼   ► but it's not a bad idea.

01:42:27 ◼   ► I don't, you know, I don't fault them for that at all.

01:42:29 ◼   ► It's interesting and it's fun.

01:42:30 ◼   ► Um, Casey, you talked about the iPad in the middle and I'm with you that I think this

01:42:35 ◼   ► shows a decided lack of imagination.

01:42:39 ◼   ► Like, oh yeah, everything is Johnny.

01:42:41 ◼   ► I was just so obsessed with freaking rounded rectangles and simplicity, but like slapping

01:42:45 ◼   ► an iPad on the dashboard is surrendering.

01:42:47 ◼   ► And it's like, no, I'm not surrendering.

01:42:48 ◼   ► Look at all the physical controls.

01:42:49 ◼   ► Look at the handle.

01:42:49 ◼   ► Look at the wrist rest.

01:42:50 ◼   ► It's like, okay.

01:42:51 ◼   ► But you put an iPad on the dashboard and you, now you like, it's not like he's not the first

01:42:54 ◼   ► person to put physical controls on the bottom of an iPad.

01:42:56 ◼   ► Like Ford has done it for crying out loud.

01:42:58 ◼   ► Like lots of manufacturers said, oh, it's an iPad, but there's a physical dial on it.

01:43:02 ◼   ► It's an iPad, but there's physical buttons along the bottom.

01:43:04 ◼   ► Is that, is that the peak?

01:43:06 ◼   ► Is that, you know, that's the best I would do we have to come up with?

01:43:09 ◼   ► I know it's a difficult problem, but like the, you know, the ergonomics of a cockpit, I refuse

01:43:15 ◼   ► to believe that an iPad with some physical buttons on it that you can tilt is the correct solution

01:43:20 ◼   ► to that problem.

01:43:21 ◼   ► Like I do feel like the cockpit, like, look, you're stationary, you're in a seat, you're

01:43:26 ◼   ► not going anywhere.

01:43:27 ◼   ► It should be possible to make a layout of controls that accommodates a lot of people that you can

01:43:33 ◼   ► get at too easily.

01:43:33 ◼   ► That is not as weird as this.

01:43:35 ◼   ► And the tilting thing that you mentioned, Casey, one of the problems with tilting is now

01:43:39 ◼   ► it becomes slightly, if you, if, unless you never move it, unless it never moves on its

01:43:43 ◼   ► own, which I would wonder about now, you have slightly more difficult time, no look reaching

01:43:48 ◼   ► for like the fan up speed or whatever, because you have to know what position you last tilted

01:43:52 ◼   ► the thing in.

01:43:52 ◼   ► Yeah.

01:43:53 ◼   ► And those tilt things, if the, if the Johnny Ive, uh, iMac, uh, monitor arm thing, uh, is

01:43:58 ◼   ► any indication they might get loose over time.

01:44:00 ◼   ► So it's very difficult to pull that off and then go down to the center console, their idea

01:44:05 ◼   ► for the center console.

01:44:06 ◼   ► You'll never guess.

01:44:07 ◼   ► Let's make it out of a bunch of rounded rectangles.

01:44:09 ◼   ► Oh my God.

01:44:10 ◼   ► It's, it's rounded.

01:44:12 ◼   ► Like the, the binnacle, that being a rounded rectangle is probably the biggest sin because it's

01:44:16 ◼   ► like for crying out loud, like a Ferrari would never have a rectilinear rounded rectangle poking

01:44:23 ◼   ► up at like, it does not look sleek or cool or aggressive in any way.

01:44:29 ◼   ► You're selling me on this.

01:44:30 ◼   ► Especially since they don't do anything with those corners.

01:44:32 ◼   ► They're just there.

01:44:32 ◼   ► Cause so it can be a rounded rectangle.

01:44:34 ◼   ► Like make like even BMW, which does a bunch of rectangles across its dashboard.

01:44:38 ◼   ► At least they have sharp corners and look Germanic.

01:44:40 ◼   ► This is not like Italian, but then the center console, a bunch of rounded rectangles stacked

01:44:46 ◼   ► on top of each other.

01:44:47 ◼   ► And Johnny has best idea for the center console thing is he doesn't like it when there's a

01:44:50 ◼   ► big door that you have to open up.

01:44:52 ◼   ► So I'm going to make two other doors.

01:44:53 ◼   ► Guess what, Johnny?

01:44:53 ◼   ► A million people have tackled this problem before.

01:44:55 ◼   ► You can have them open from the side entirely in both directions.

01:44:59 ◼   ► There's lots of different ways to do this, but the way he chose was that you can open half

01:45:03 ◼   ► of each of the round, the sides of the rounded rectangle.

01:45:05 ◼   ► But then that like the two, when both sides are open, there's just the spine down the

01:45:10 ◼   ► middle, which makes it so you have to like shove your hand down.

01:45:12 ◼   ► It's like, you can never get it fully open.

01:45:15 ◼   ► There's always, it's always the middle of it is always blocked by that thing.

01:45:18 ◼   ► The shifter arrangements like, oh, we made it of glass and it's a little thing.

01:45:21 ◼   ► I feel like Ferrari already had a kind of a good idea with its gated shifter, like play

01:45:25 ◼   ► on the gated shifter.

01:45:25 ◼   ► But that doesn't really matter that much, I guess.

01:45:27 ◼   ► But like, yeah, the center console certainly doesn't look like a Ferrari and ergonomically

01:45:31 ◼   ► it's nothing special.

01:45:32 ◼   ► The cup holders poking out there all the time.

01:45:33 ◼   ► Yeah, they're beautiful aluminum cup holders, but like they're always in your face.

01:45:36 ◼   ► There's a reason a lot of cars have them tuck away, especially sports cars.

01:45:39 ◼   ► I just like how much space you get for how much space it takes up the way the rear controls

01:45:44 ◼   ► look.

01:45:44 ◼   ► I'm just not impressed from a car design perspective.

01:45:46 ◼   ► The only thing that's impressive about it is that it, it looks like a Johnny, a beautiful

01:45:51 ◼   ► Apple type interior and everything is extremely expensive, made of beautiful materials.

01:45:56 ◼   ► I'm sure it feels great to you like all of that, right?

01:45:58 ◼   ► But this is going to be like a, you know, $300,000 car or something.

01:46:01 ◼   ► So I'm sure keep going, but yeah, well, maybe 500.

01:46:04 ◼   ► But anyway, I'm sure, I'm sure like, yeah, the materials better feel good.

01:46:08 ◼   ► And it's fun to do this as an exercise.

01:46:09 ◼   ► And the people who made it should be proud of the beautiful thing that they've made.

01:46:12 ◼   ► But I don't even think it's that great of a car interior in terms of ergonomics.

01:46:16 ◼   ► I feel like the people who design like the interior of like the Toyota Sienna or the Honda

01:46:20 ◼   ► Odyssey, like those are the best designers in the industry because they find a way to

01:46:24 ◼   ► have like insane amounts of storage space and adjustability and cubbies for everything and

01:46:29 ◼   ► things that make kids happy in the back seats and everything.

01:46:31 ◼   ► That is a harder design problem, better solved than this.

01:46:35 ◼   ► Granted, that's just a bunch of cheap plastics and things that are going to be durable to

01:46:38 ◼   ► spills and stuff.

01:46:39 ◼   ► And it doesn't look as beautiful as this.

01:46:41 ◼   ► But I feel like being a designer is not just like making something that satisfies the itch

01:46:46 ◼   ► that you have in your brain, which is essentially when you're Johnny out of this stage in his

01:46:50 ◼   ► career can afford to do that.

01:46:51 ◼   ► So good for him.

01:46:52 ◼   ► But like there's more to design than making a an item that you go to sleep with a smile

01:46:58 ◼   ► on your face, knowing that you made it that exists in the world.

01:47:01 ◼   ► It has to solve problems for users.

01:47:02 ◼   ► And I don't think this is a bad interior, but I don't think it solves problems for users

01:47:07 ◼   ► in a better way.

01:47:08 ◼   ► Like it's not the best in every area than any other control.

01:47:11 ◼   ► I can think of any other interior.

01:47:13 ◼   ► I can think of other cars that are better of it in lots of areas, setting aside the fashion,

01:47:17 ◼   ► setting aside the Ferrari and it's just judging it as an ergonomic car interior.

01:47:20 ◼   ► It does a lot of things right, but it does a lot of things in a very uninteresting way.

01:47:24 ◼   ► And one of the things I thought about this is if you're if you're interested in this type

01:47:27 ◼   ► of thing, like I love how everything is beautiful and machined and it's like little jewels

01:47:31 ◼   ► and stuff like that.

01:47:31 ◼   ► I have another multimillion dollar car you might be interested in.

01:47:35 ◼   ► I'll put a link in the show notes to Gordon Murray talking about his selection of knobs

01:47:39 ◼   ► and switch gear for his ridiculously expensive cars that he makes.

01:47:43 ◼   ► Plus some more videos you can see on his channel about I will have timestamp linked to him talking

01:47:48 ◼   ► about his knobs and his shift lever and stuff.

01:47:51 ◼   ► It's totally up Johnny Ives alley.

01:47:53 ◼   ► It's like, oh, we spent this amount of money on this thing because it just had to be just

01:47:56 ◼   ► right.

01:47:56 ◼   ► And we kept going back and forth until the switches felt just so or whatever.

01:48:00 ◼   ► If you have a multi-billion dollar car, by all means do that.

01:48:03 ◼   ► But here's the thing about Gordon Murray's car.

01:48:04 ◼   ► Go look at the videos on the channel and the tours of it.

01:48:07 ◼   ► It looks like a Gordon Murray automotive car.

01:48:10 ◼   ► It looks like there's no mistaking it for any other brand.

01:48:13 ◼   ► It has him and his like his design ethos is through the entire thing.

01:48:18 ◼   ► And compare the interior of that car to the interior of this car.

01:48:21 ◼   ► Like one of them screams Gordon Murray automotive.

01:48:25 ◼   ► And maybe you don't know what that brand is or like, like it is very cohesive.

01:48:29 ◼   ► And this one, it seems, I mean, it's kind of interesting that they showed us the interior

01:48:33 ◼   ► not inside a car because it really does look like an interior floating around in space waiting

01:48:38 ◼   ► to be put into the Apple car.

01:48:40 ◼   ► You're not wrong.

01:48:41 ◼   ► But it isn't even really Apple.

01:48:42 ◼   ► Like, again, like, yeah, I think Dieter Ram's obviously huge influence.

01:48:45 ◼   ► I think also one thing I didn't think of earlier was Leica cameras.

01:48:48 ◼   ► Like this looks a lot like Leica style.

01:48:51 ◼   ► Although I will say with these, the influences, which you're totally right about, like everyone

01:48:55 ◼   ► who works there, like love like the cameras, love Dieter Ram's.

01:48:57 ◼   ► But here's the thing.

01:48:59 ◼   ► At this point in time and in all their careers, Johnny Ive, Mark Newsom, all those people from

01:49:05 ◼   ► Apple, their footprint on history in terms of design is bigger than all the, all of their

01:49:11 ◼   ► influences, which is very often the case.

01:49:13 ◼   ► Those influences are important to them, but the imprint they have put on the world with

01:49:18 ◼   ► like the iPod, the iMac, the iPhone, iOS, like that footprint in terms of design influence

01:49:26 ◼   ► is bigger than all of their past influences.

01:49:29 ◼   ► So at this point, yes, we see all their influences, but at this point they are not referencing

01:49:34 ◼   ► themselves, but they are essentially setting a template for all the designers in the future

01:49:38 ◼   ► who are going to be like, we're going to look at them and say, oh, that reminds me of Johnny

01:49:41 ◼   ► Ive's design.

01:49:41 ◼   ► It's like, well, you know, Johnny Ive was really inspired by Dieter Ram's and Leica

01:49:44 ◼   ► cameras like who like the it's, it's almost absurd to even talk about their influence at

01:49:50 ◼   ► this point because their particular personal tastes and little brain itches are the only

01:49:56 ◼   ► things that matter because they have so much money and power.

01:49:58 ◼   ► It boggles my mind that Ferrari even let them do this to their cars because they said, we're

01:50:02 ◼   ► going to, we're going to, we're going to contract the best designers in the world.

01:50:05 ◼   ► And then it's almost like they were intimidated and didn't feel like they could push back and

01:50:09 ◼   ► say, you're going to put a rounded rectangle behind this steering wheel and then an iPad

01:50:15 ◼   ► on the dashboard.

01:50:16 ◼   ► And it looks like a Fiat.

01:50:18 ◼   ► Like, shh, it's Johnny Ive.

01:50:20 ◼   ► We paid him how many billions of dollars for this?

01:50:22 ◼   ► Look at the books he gave us.

01:50:23 ◼   ► He gave us a stack of books.

01:50:25 ◼   ► They're beautifully bound.

01:50:26 ◼   ► We have to do what he says.

01:50:27 ◼   ► I'm like, oh, Ferrari, you don't.

01:50:28 ◼   ► Yeah.

01:50:29 ◼   ► I will say though, one of the things that I thought was really interesting, which I don't

01:50:33 ◼   ► recall having seen before is the rear control panel.

01:50:36 ◼   ► So on the back of the center console, there's a screen that kind of mimics the control panel,

01:50:40 ◼   ► which they call the iPad.

01:50:42 ◼   ► Or sorry, the iPad that they can control.

01:50:44 ◼   ► We call the iPad.

01:50:44 ◼   ► Yeah, exactly.

01:50:45 ◼   ► The control panel.

01:50:46 ◼   ► But it has on there like your current speed, how much power the car is generating, a graph

01:50:51 ◼   ► of like recent power and recent speed, which I thought was kind of neat.

01:50:54 ◼   ► You can see that because they're using Celsius, there's decimals like animals.

01:51:00 ◼   ► Oh, okay.

01:51:01 ◼   ► But anyways, no, this does look incredible, but you've convinced me that this should not

01:51:05 ◼   ► be in a Ferrari and I take back everything I've said.

01:51:07 ◼   ► The other thing, by the way, the toggle switches, which is a great idea, but they didn't go full

01:51:11 ◼   ► physical because they're basically like, it's like those dials that turn forever, you know,

01:51:14 ◼   ► as opposed to a dial that has stops on it.

01:51:16 ◼   ► These are toggle switches where you can press up and press down, but it's actually like a

01:51:19 ◼   ► D-pad where you're going up, up, down, down.

01:51:21 ◼   ► There's no, there's no physical persistence of the switches, which is not great because

01:51:24 ◼   ► for example, every time I do this now, I think of this because of all the rants I've

01:51:27 ◼   ► had in this program.

01:51:28 ◼   ► When I'm driving around the winter, a couple of my cars have seat heaters in them and I

01:51:33 ◼   ► can turn the seat heaters on, turn them on high, low or off for me and the passenger.

01:51:39 ◼   ► And I can also tell whether they're currently high, low or off for me and the passenger without

01:51:45 ◼   ► looking at them at all, because they are physical stateful switches that, that rock.

01:51:50 ◼   ► I don't have to, I know where they are, they never move and I can tell what position they're

01:51:54 ◼   ► in by feeling them through my winter gloves, feeling what position they're already in without

01:51:58 ◼   ► ever looking, putting my eyes anywhere.

01:52:00 ◼   ► You cannot do that with any of these controls.

01:52:02 ◼   ► So that's one thing because they're just, you know, D-pad type things.

01:52:04 ◼   ► And the second thing is they're machined aluminum things poking out of the dashboard of a car.

01:52:11 ◼   ► I know they have the little things that prevent you from accidentally hitting them, like the

01:52:16 ◼   ► little, little loops of metal, kind of like roll bars to prevent that someone's knee,

01:52:20 ◼   ► from accidentally hitting the, like the seat temperature up thing or whatever in the back

01:52:23 ◼   ► seats and stuff like that.

01:52:24 ◼   ► Uh, but you could still accidentally hit them with your knee.

01:52:27 ◼   ► First of all, there's kind of a reason why they don't have those toggles, which in the

01:52:29 ◼   ► second of all, God forbid in any kind of accident or something, now you have these little metal

01:52:35 ◼   ► baggers in the dashboard.

01:52:36 ◼   ► Now you should be belted in airbags and yada yada and a little crash tested.

01:52:40 ◼   ► And I understand, but like, there's kind of a reason why car interiors aren't festooned with

01:52:45 ◼   ► stiff pieces of metal sticking out of the dashboard, Johnny.