PodSearch

Cortex

133: The Ethics of AI Art

 

00:00:00   So this isn't an actual episode out of time, but it's real close.

00:00:04   This episode should be released just before the iPhone event.

00:00:09   So by the time most people have listened to this,

00:00:11   but we're doing it now a little earlier than we would normally

00:00:15   because I'm going to be traveling and I got a lot going on in September.

00:00:20   So we're trying to get this done just before.

00:00:23   - I don't understand how this makes it an episode out of time at all.

00:00:26   - It doesn't. - Yeah, okay.

00:00:27   - It's close. - Is it?

00:00:28   I don't know. I don't agree. I feel like this is more like,

00:00:31   "Oh, there's just a bunch of stuff that's going to happen?"

00:00:34   But that's not an episode out of time.

00:00:35   That's not even an episode out of time adjacent.

00:00:37   You're always talking about things that are going to happen in the future.

00:00:40   That doesn't make that an episode out of time.

00:00:42   I always feel like episodes out of time are made more out of time-y

00:00:46   based on how dated they are due to something we would normally talk about.

00:00:50   That's kind of how I imagine it.

00:00:52   So, like, if we were releasing this episode a week later,

00:00:55   it would be an episode out of time.

00:00:56   'cause so much would have changed at that point.

00:00:59   Things we'd normally talk about, like new iPhones,

00:01:01   new Apple Watches, whatever, that I think it would qualify.

00:01:04   - Wait, I feel like I'm in crazy town here.

00:01:07   Like, don't episodes out of time,

00:01:08   we normally try to pick something that isn't time sensitive?

00:01:11   That's why it can be an episode out of time?

00:01:13   - No. - It's like, oh,

00:01:14   we're gonna record this thing,

00:01:15   and then we'll release it two months from now.

00:01:17   - But then we would never reference

00:01:19   that it was an episode of time.

00:01:20   We reference that they're episodes out of time

00:01:22   because of the fact that it's obvious

00:01:24   that time would have passed because there's something

00:01:27   that we would have otherwise spoken about,

00:01:29   or are worried that something could have happened

00:01:31   that would date the episode in a way that was weird.

00:01:34   So we say this is an episode out of time.

00:01:36   - I don't know. - That's why we do it.

00:01:38   - Oh, okay, I feel like I've fallen through

00:01:39   some kind of time vortex,

00:01:40   and I have no idea how this works anymore.

00:01:42   - 'Cause like if we did a book, right,

00:01:44   and just did a whole episode about a book,

00:01:46   we could record and release that whenever,

00:01:48   and that wouldn't be an episode out of time,

00:01:49   because it's just about the book.

00:01:51   But if we're doing an otherwise normal episode,

00:01:54   where we just talk about topics,

00:01:56   things that we wanna discuss.

00:01:57   If there's like a long time between recording and releasing,

00:02:01   then it becomes an episode at a time.

00:02:03   Because if we leave it for like four weeks, five weeks,

00:02:05   which we do sometimes, anything could happen

00:02:08   that would otherwise date the show.

00:02:11   And typically this happens around,

00:02:13   I think the last couple, if my memory serves,

00:02:16   has been because there's gonna be some kind of Apple event

00:02:18   or whatever that we would otherwise talk about.

00:02:20   But we can't talk about it because it hasn't happened yet.

00:02:23   - Okay, I feel like the only thing that is making this,

00:02:27   even remotely, an episode out of time

00:02:29   or episode out of time adjacent,

00:02:32   is this conversation right now.

00:02:34   - September is a huge time for a bunch of reasons.

00:02:38   It's busy time, work-wise, so we got new iPhones coming.

00:02:43   iOS comes out, all that kind of stuff, so busy time.

00:02:46   But around these parts, it's especially busy

00:02:50   because we turn our attention to raising money

00:02:52   for St Jude Children's Research Hospital. Because September is Childhood Cancer Awareness

00:02:57   Month. For the fourth consecutive year, the Real AFM community is rallying together to

00:03:03   support the life-saving mission of St Jude Children's Research Hospital.

00:03:08   Finding Cures, Saving Children This year, St Jude is celebrating its 60th

00:03:13   anniversary. Since opening its doors in 1962, St Jude Children's Research Hospital has

00:03:19   grown in size and capabilities for one special reason. They believe that children all over

00:03:24   the world deserve the same chance at survival. Treatments developed at St Jude Children's

00:03:29   Research Hospital have helped increase the overall childhood cancer survival rate from

00:03:34   20% to more than 80% in the 60 years that it's been around. While these tremendous strides

00:03:40   have been made, 1 in 5 children diagnosed in the US will not survive, and globally the

00:03:45   the numbers are shockingly reversed,

00:03:47   with four in five children in some developing countries

00:03:50   not surviving of childhood cancer.

00:03:53   Limited access to high quality affordable medicines

00:03:55   and the financial burden of research and care

00:03:58   are hallmarks of the childhood cancer challenge

00:04:00   that many developing countries face.

00:04:02   And this is how it is across the globe,

00:04:04   and this is the kind of stuff that St. Jude is changing.

00:04:07   St. Jude is a hospital in Memphis.

00:04:10   It has a very important link to us here at Relay FM.

00:04:13   You know, if you're not aware, you wonder why do they raise money for St. Jude Children's

00:04:17   Research Office every year.

00:04:19   My co-founder, Stephen Hackett, lives in Memphis, was lucky to live in Memphis as his eldest

00:04:25   son was diagnosed with a brain tumor when he was very, very young.

00:04:31   And St. Jude took him as a patient, saved his life.

00:04:36   and he is now an incredible young man who I love dearly and wouldn't have been around

00:04:47   without St. Jude's help.

00:04:49   So that is our personal connection to this place.

00:04:52   But the more time we have spent and learnt about St. Jude, the more time I have had to

00:04:57   interact with doctors and patients and families, it just shows how special a place this is.

00:05:04   It is a very unique institution in America because of the way that they work in that

00:05:08   families do not pay for the treatment and St. Jude also provides food and housing for

00:05:14   patient families so they can focus on the health of their child.

00:05:19   But St. Jude is also a research institution and they share what they learn with the world.

00:05:25   For example, recently in 2021 they did some medical trials that saw a 20 point improvement

00:05:31   in survival rates for high risk neuroblastoma, which is the second most common solid tumour

00:05:36   in children. They produced this antibody at the campus in St Jude and have then since

00:05:42   shared this knowledge with the world. So we turn our attention to raising money for St

00:05:48   Jude for a bunch of reasons, primarily because they want to save the lives of children everywhere.

00:05:57   Childhood Cancer Awareness Month is an opportunity for us to come together in the fight against

00:06:01   childhood cancer because together we can make a big impact.

00:06:05   This year you are able to support the Relay FM for St Jude campaign in multiple ways.

00:06:10   So as always we invite you to make a donation.

00:06:13   Donors who make an individual gift of $60 or more will receive a digital bundle including

00:06:17   a wallpaper and a Mac OS screensaver pack.

00:06:21   If you donate $100 or more you get this plus some special stickers.

00:06:29   If your employer offers a matching gift program, check with your employer if they do matching

00:06:35   gifts.

00:06:36   If they do, there's a form at stjude.org/relay that can be credited to our campaign.

00:06:40   If you work for a big company, it's very likely that they do charity matching.

00:06:44   Ask.

00:06:45   And new this year, if you want to get more hands on, you can now start your own fundraising

00:06:50   campaign to help us reach our goals while also earning exclusive Relay FM merchandise.

00:06:55   If you sign up to fundraise and you raise $1 or more, you'll get an exclusive St. Jude

00:06:59   limited edition of a relay of fame challenge coin. If you raise $250 or more you will get

00:07:05   a very weird and unique desk mat which features the cartoon heads of me and Steven. It's very

00:07:10   strange but it's great. If you may have seen last year that I covered my desk in stickers

00:07:16   of Steven's face, we are now allowing you to do that on a temporary basis because you

00:07:21   will, unlike me, you will be able to remove the desk mat and put it back whenever you

00:07:26   want.

00:07:27   like something out of a nightmare but it's also kind of amazing yeah.

00:07:31   So you've seen my office right you've seen their desk. This year I'm doing the big one.

00:07:36   Oh you're doing the big one okay so that one's gonna get completely covered?

00:07:39   More Steven stickers. So basically we do this is like for every couple of hundred dollars

00:07:43   raised we each put a sticker on some piece of furniture in our offices so my recording

00:07:48   desk is completely covered and now it's gonna be spreading to another but if you want to

00:07:52   do it sign up to be a fundraiser raise $250 or more and you'll get a desk mat so you

00:07:57   you can take it off when you don't want it.

00:07:58   Now the fundraising thing is really cool.

00:08:02   What it does is it enables you to be able to extend

00:08:06   this message yourself to your friends, your family,

00:08:08   your coworkers, and help us raise more money.

00:08:10   This is money that our campaign would never see otherwise,

00:08:13   'cause maybe these people don't listen to our shows.

00:08:15   And also if you don't have the money to donate yourself,

00:08:20   this is also a way to get involved.

00:08:22   So you're able to say go to family, friends, and say,

00:08:25   here's this thing, it's an amazing institution,

00:08:28   would you like to donate some money?

00:08:30   Go to stjude.org/relay, you can donate

00:08:33   and find out more about fundraising.

00:08:34   So that's what's going on there.

00:08:36   I wanna talk about the podcastathon.

00:08:38   So this will be the fourth podcastathon,

00:08:42   which is an eight hour event that we do.

00:08:45   It's a variety show, features many Relay FM hosts,

00:08:48   special guests, me and Steven host it together.

00:08:51   This year it's gonna be on September 16th,

00:08:53   from 12 to 8 p.m. US Eastern Time.

00:08:56   Now, the plan is we'll be back in person.

00:09:00   - Wow, okay.

00:09:01   - So the first one we did in person,

00:09:04   then we had to do two remote.

00:09:05   So me here in Mega Studio, Steven at St. Jude.

00:09:08   We really wanted to do it last year,

00:09:13   but we just couldn't make it work.

00:09:14   It was very unfortunate with travel restrictions,

00:09:17   'cause the travel restrictions ended in October,

00:09:20   so we kind of just missed out on it.

00:09:23   - Yeah, it was a heartbreakingly close call last year.

00:09:26   I remember that, that was awful.

00:09:27   - It was rough, it was rough.

00:09:29   But this year, the plan is that I'll be back in Memphis

00:09:33   and we're gonna do it live together.

00:09:35   I'm terrified because of my last big trip to America.

00:09:40   So I have like real kind of COVID anxiety right now.

00:09:44   We have tried to plan for all of the potential contingencies

00:09:49   but really we're planning for one thing,

00:09:52   which is doing it in person.

00:09:54   And if it doesn't work, we'll just fall back.

00:09:57   But it's kind of like, there's no point,

00:09:59   I think, going deep on all the contingency plans

00:10:02   because there's so many variables for like,

00:10:06   what if I get COVID or what if Steven gets COVID

00:10:10   or what if somebody in our family's had-

00:10:11   - Layover in Austin, you're gonna record from there?

00:10:14   You never know.

00:10:15   - We are doing a layover.

00:10:16   Chicago the way there, Dallas the way back,

00:10:18   another 90 minute layover on the way back.

00:10:20   So I think we'll be spending our third night in Dallas.

00:10:23   - Okay.

00:10:24   - Right, I see no other potential here.

00:10:27   We will be in Dallas for the third time this year.

00:10:29   I'm saying it right now, I'll report back on this later,

00:10:32   but I am convinced that we will be in Dallas again.

00:10:34   (laughing)

00:10:35   It's just, I see no other way.

00:10:37   But it's kind of one of those things where

00:10:39   we had a conversation as a group

00:10:40   because we work with a wonderful team of people

00:10:43   at St. Jude who help us plan these events.

00:10:46   And we're like, okay, well, we'll just see what happens.

00:10:49   'Cause I think it's too complicated

00:10:52   to try and specifically plan out every contingency.

00:10:55   It's just like high level, this is what we'll do,

00:10:58   but we're all just like headstrong

00:10:59   on we're gonna be together.

00:11:01   We're gonna bring in a bunch of our favorite elements

00:11:03   of the past three events, especially the last two,

00:11:07   because I've done things and Steven's done things,

00:11:10   but we've not been able to do those things together.

00:11:12   So like last year, he had this big challenge wheel

00:11:15   that he would spin and it would land on certain things

00:11:18   to have to do them. I've never got to spin the wheel. And we're also, we're constructing

00:11:23   a balloon room.

00:11:24   Ah, balloon room! Yeah, I was gonna ask, like, there better be a balloon room.

00:11:27   So I'm actually really excited. It's gonna be levelled up in a fun way. So I will still

00:11:34   have my balloon room, but Steven will be able to experience the balloon room, which I think

00:11:38   he's actually very scared of. I think he doesn't like the idea of the being surrounded by balloons.

00:11:43   It always would freak him out when I would dive in them. So we'll see what that would

00:11:46   be like.

00:11:47   gonna be like a greatest hits at the last four years which I'm very excited about.

00:11:53   It actually also might be the day that iPhones come out. It's own little thing.

00:11:57   Yeah it is it is absolutely insane how much work the two of you put into this podcast

00:12:03   athon every year and yeah getting there in person to do it together this year like I'm

00:12:09   so glad that that's gonna work out again it was just it was so heartbreaking last time

00:12:13   that you had to do it remotely, so I'm very happy about that.

00:12:17   I'll be curious to see Steven's apparent fear of balloons explored,

00:12:21   and that can always be a nice motivator for people to donate.

00:12:24   This man is afraid of balloons!

00:12:26   Donate, and we had a balloon, right?

00:12:28   Like, that's his classic.

00:12:29   - Oh, it's like how I always have to eat those disgusting jelly beans.

00:12:32   And like, give more money and Myke eats the, like, jelly beans

00:12:36   that are flavored like dishwater and rotten eggs.

00:12:39   It's like, great, I'll just—I'll keep eating this.

00:12:41   - If there's one fundraising crowd pleaser,

00:12:43   It's making someone uncomfortable in some way that goes along with it.

00:12:47   Right?

00:12:47   People, people just love it for the, for the fundraisers.

00:12:50   I hate the thought of it, but at the same time, it donates money for childhood

00:12:54   cancer, so I'm like, all right, I'll do it then I suppose, you know, like,

00:12:58   I'll suffer this.

00:13:00   Yeah.

00:13:00   People like people suffering for a cause.

00:13:03   So that's, that's, that's what that is.

00:13:04   Yeah.

00:13:04   They've got so much happening September.

00:13:07   We're doing a bunch of extra streams.

00:13:08   For the first time this year, we're going to be doing a campaign closing

00:13:11   stream at the end of September.

00:13:13   where we announce the final total and all that kind of stuff.

00:13:16   We've got a bunch of milestone streams that we're unlocking

00:13:19   with different Relay FM hosts and we're playing video games and that kind of stuff.

00:13:22   All of this is over at sttude.org/relay

00:13:25   so please go and check it out.

00:13:27   Go to sttude.org/relay, you can learn more,

00:13:30   you can donate, you can sign up to fundraise

00:13:32   and let's cure childhood cancer together.

00:13:34   This episode of Cortex is brought to you by Squarespace.

00:13:39   They are the all-in-one platform for building your brand

00:13:41   growing your business, growing your presence online.

00:13:46   With Squarespace you will be able to stand out with a beautiful website, engage with

00:13:49   your audience with beautiful design and sell anything.

00:13:52   Your products, services and even the content you create, Squarespace has got you covered.

00:13:58   With Squarespace you will be able to get blogging.

00:14:00   They have powerful blogging tools so you can share your stories, your photos, your videos,

00:14:04   your updates.

00:14:05   You will be able to categorise, share and schedule your posts to make your content work

00:14:09   for you.

00:14:10   You can then get stuck in with SEO tools, they have a suite of integrated features and

00:14:14   useful guides that will help you maximise prominence among your search results.

00:14:18   Then maybe as your blogging presence grows, you might want to start a store of your own,

00:14:23   whether you sell physical or digital products, Squarespace has the tools you need to start

00:14:27   selling online.

00:14:28   But this is just one potential path, Squarespace will let you build just about any type of

00:14:32   website you want, whether it is to publish something online for people to see, or maybe

00:14:36   it's for something to share with people that are important to you.

00:14:39   When I got married we set up a website with Squarespace, they have beautiful templates

00:14:43   actually just for wedding websites and they come with all the structure that you're going

00:14:47   to need so it makes it really easy for you to be able to make sure your guests have everything

00:14:51   that they're going to need.

00:14:52   Squarespace really is amazing, I've used it for so many projects now, no matter what it

00:14:56   is they have all of the tools so if you want to build a website, trust me Squarespace is

00:15:01   the first place to go and you can go see it for yourself.

00:15:04   If you go to squarespace.com/quartex you can sign up for a free trial, there's no credit

00:15:07   card required and you can go in and set up your entire website. You can choose from a

00:15:12   beautiful template, you can customise it to how you want, you can even start putting your

00:15:15   content in and everything, all in their trial. You don't need to give them a credit card.

00:15:20   But then when you're ready to launch it to the world, you sign up and use the offer code

00:15:23   Cortex. This will save you 10% off your first purchase of a website or domain. They have

00:15:28   domain options too. Ask Squarespace.com/Cortex and when you decide to sign up, use the offer

00:15:33   You will get 10% off your first purchase

00:15:35   and show your support for this show.

00:15:37   Squarespace really is an all-in-one package

00:15:40   for you to get what you want online.

00:15:41   Our thanks to Squarespace for the support of this show

00:15:44   and all of Real AFM.

00:15:45   Do you use Git now?

00:15:48   - Oh, no. (laughs)

00:15:50   I made a terrible mistake.

00:15:52   - You've ruined our subreddit.

00:15:54   You've ruined it. - I didn't ruin it.

00:15:55   I just, I didn't know what I was asking, so.

00:15:58   (laughing)

00:16:00   - Okay, so listeners who don't go to the subreddit,

00:16:03   previously on Cortex, Gray was discussing

00:16:07   his syncing problems with Obsidian,

00:16:10   and Myke was bullying him in an entirely correct way.

00:16:14   And so we had this big conversation about like,

00:16:16   how to have a workflow in my writing program

00:16:20   that syncs between me and my assistant

00:16:23   that can be used for making changes on PDFs

00:16:27   and a whole bunch of stuff.

00:16:27   Anyway, the fundamental question is,

00:16:30   "Hey, I would like a text document that synchronizes between me and my assistant,

00:16:36   and there are reasons why it is difficult to do in my writing app of choice, Obsidian."

00:16:42   And so on the subreddit, of course, very helpful people were proposing suggestions,

00:16:50   and one of them was Git. And Git is one of these things of like, "I've heard about Git. I've heard

00:16:57   programmer friends talk about this thing for you know you like you push and you pull on the gits

00:17:03   there's like a tree and the tree merges there's a hub is there a hub i don't know i think there's

00:17:13   forks and pulls and and requests requests yeah there's like stuff i don't know it's just one of

00:17:19   these things like you're aware you can't have programmer friends without being vaguely aware

00:17:23   of like, oh, Git is a thing that programmers use,

00:17:27   and it's, oh, programmers don't be angry,

00:17:30   it's like a way to collaborate for programmers

00:17:33   on source code, I guess.

00:17:35   - Well, it does, as far as I'm aware.

00:17:36   - Yeah, as far as I'm aware, too, but Myke,

00:17:38   on the internet, you never know, right?

00:17:40   - Oh, come on. (laughing)

00:17:42   Who could be mad about that?

00:17:43   I also, like Federico, I friend Federico

00:17:46   who writes wonderful iOS reviews every year,

00:17:49   I know that he started using Git

00:17:52   as a version control system and a way to share

00:17:57   his reviews with proofreaders and stuff like that.

00:18:01   I know that he started doing that.

00:18:02   - Oh, okay, interesting.

00:18:03   - Multiple years ago.

00:18:04   It's like a backup system and like a checking in system,

00:18:08   like similarly to how programmers use it.

00:18:10   Like, this part's done, do you wanna look at it?

00:18:14   And it's got version control and history

00:18:16   and all that kind of stuff.

00:18:18   So it ends up being a, I think, pretty decent tool

00:18:22   that people use for collaboration of things

00:18:24   other than coding, just because of the way that it works

00:18:28   from a fundamental level.

00:18:30   - Yeah, it was one of those moments

00:18:31   where someone suggested, they said,

00:18:32   "Hey, have you looked into Git?"

00:18:34   And I said, "Oh, I would like to know more."

00:18:36   And boy, did I get more.

00:18:38   And I went a bit quick.

00:18:40   People were very helpful,

00:18:42   but it also became very clear, very fast, like,

00:18:46   Okay, I think this is one of those tools where if you are a programmer who is already familiar

00:18:52   with this, it is a great tool that is totally useful in a lot of circumstances.

00:18:58   But if you're not already familiar with this, this is like, "Oh God, here's another layer

00:19:04   of thing that can go wrong."

00:19:07   And if I'm just using it in this one scenario, it just it struck me as dangerous overkill

00:19:13   for the current situation.

00:19:14   - Yeah, it seems like a big learning curve, I think.

00:19:17   Like if you're coming in cold,

00:19:19   we use GitHub for the management

00:19:22   of our software development projects,

00:19:25   every layer found, like we run our own systems,

00:19:27   we maintain our own publishing systems

00:19:29   and ad platform system and that kind of stuff.

00:19:32   So I have to use GitHub every now and then, right?

00:19:35   Which uses Git, but it's its own tool, right?

00:19:37   It's Git fundamentally,

00:19:39   but then it has its own stuff on top.

00:19:41   And like, I can see how good it is,

00:19:43   but oh boy does it confuse me when I'm in there.

00:19:45   - Oh, so you're actually hands on pushing and pulling

00:19:49   and forking and fitting or whatever?

00:19:51   - I don't do the pushing and the pulling

00:19:51   and the forking and the, I use issues,

00:19:54   which is just like a bug tracking request system.

00:19:57   I do no forking or hubbing or anything like that,

00:20:01   but I see that it happens.

00:20:03   I see that the push is merged and you know,

00:20:07   all that kind of stuff.

00:20:08   - The push is merged and you know,

00:20:12   all that kind of stuff. - You know.

00:20:13   And then the other plots happen.

00:20:15   Everything we said, and that's the one you take.

00:20:18   (laughing)

00:20:20   - Yeah, so I want to thank everybody

00:20:22   who was trying to convince me to use Git.

00:20:24   I think it's overkill.

00:20:25   It's also one of those situations where,

00:20:28   we talked about this with Dropbox for me,

00:20:30   and it was like you were discussing

00:20:32   like Google Notes as well.

00:20:33   There's an additional layer, which is always when

00:20:36   it has to be a tool not just for you,

00:20:38   but a tool when you're working with multiple people.

00:20:40   And it was one of these things of like,

00:20:43   I might be able to use Git, but selling this as a,

00:20:47   hey, here's what we're gonna do between the two of us

00:20:50   with me and my assistant.

00:20:51   I was like, this is completely out of the question.

00:20:54   There is no way this is gonna be the right tool

00:20:56   for the two of us.

00:20:57   - I saw somebody ask on Twitter,

00:20:59   would iCloud file sharing not work for this?

00:21:02   - Okay, so here's the current state of affairs.

00:21:05   Someone else mentioned that on the Reddit.

00:21:08   And it's one of these interesting moments

00:21:10   because I realized that I had gotten something into my head

00:21:15   that wasn't true.

00:21:17   Sometimes when you've used systems for a long time,

00:21:19   you can have this idea of, oh, I know the way things work

00:21:23   because I know how a tool that I've used for years works,

00:21:27   and you don't realize, oh, it's actually changed since then,

00:21:30   and so you have artificial restrictions in your brain.

00:21:32   So in iCloud Drive, there are these,

00:21:36   I don't even know what to call them.

00:21:38   Apple has a specific name for them,

00:21:40   but there are like these folders that your applications use,

00:21:44   which is sort of different from,

00:21:46   oh, you just have a folder.

00:21:48   And often they're putting like,

00:21:49   oh, these are files that are not really meant

00:21:52   for the user to access.

00:21:54   It's like a bunch of stuff that the app

00:21:55   just keeps for itself.

00:21:56   - Well, some apps keep all of your projects in there, though.

00:21:59   - Yeah, yeah, yeah.

00:22:00   But like sometimes the apps are just using it

00:22:01   for themselves, sometimes it's,

00:22:03   oh, hey, you can actually put files in this.

00:22:06   And so this exists at the root level of iCloud Drive.

00:22:09   And in my head, precisely because of my long experience

00:22:12   with Apple and the way iCloud Drive used to work,

00:22:16   I always filed those in my head as like,

00:22:20   oh, these are like library folders.

00:22:22   They're not really here for the user.

00:22:24   You just interact with them entirely through the app.

00:22:27   And I have a separate folder,

00:22:29   which is acting like basically Dropbox,

00:22:32   where it's like gray's iCloud stuff.

00:22:34   And this just has a bunch of folders in there.

00:22:36   And I had artificially in my head-

00:22:38   something about gray is like gray's iCloud stuff keep out top secret yeah exactly no boys allowed

00:22:44   yeah it was in kind of gross i think was the sign um but good old covenant hops but so anyway like i

00:22:55   just i had divided in my head this totally artificial line which i realized through the

00:23:00   conversation on reddit so again very helpful to everybody who was leaving comments i was like

00:23:04   wait a minute, I think I can just dig into this Obsidian iCloud folder where I'm syncing stuff.

00:23:10   And can I just share a single subfolder of this with my assistant? And it turns out, yes, I can.

00:23:19   So I can, I can go into there and just share a folder with her. Now, we've currently run into

00:23:26   one more technical problem, which I was literally checking in as we were putting the show notes

00:23:31   together this morning. I'm like, "What's the current status of us trying to figure this out?"

00:23:34   So the current limitation that we don't know if it's technical or not is,

00:23:40   "My assistant will be doing this work on a Windows machine."

00:23:43   Why the sharp intake of breath there, Myke?

00:23:48   I would not trust that at all.

00:23:51   Okay, so here's the fun we've been having, because I literally haven't had my hands touch a Windows

00:23:58   computer in greater than a decade.

00:24:00   Like the last time I used a Windows computer, I was working as a teacher.

00:24:04   And so we were going through a bunch of steps, but I figured out, oh, Apple

00:24:08   makes a program that they call iCloud for Windows, which, and now this is,

00:24:16   this is where we get into it.

00:24:17   I'm trying to help my assistant with like technical support on her Windows

00:24:21   machine upon which I know nothing.

00:24:24   So it feels like I'm a blind person, like, reaching into the void and being like,

00:24:29   "Oh, maybe this box goes over there? I don't have any idea."

00:24:32   She sent me screenshots and like, it was so funny to get a screenshot of just like,

00:24:36   "Oh, here's what I'm looking at." And like, I know what nothing on the screen is.

00:24:40   I don't even know if I'm looking at a file browser. Like, what is this? I have no idea.

00:24:43   So she is able to access the files on her Windows computer.

00:24:49   But what we haven't been able to figure out is there doesn't seem any way for her to

00:24:53   edit them in place. Like she has to make a copy, she can edit the copy, then delete

00:25:00   the original and rename the copy the same as the original.

00:25:04   Don't like that at all.

00:25:06   I know, right? So I'm not exactly sure what the current state of this project is, but it does turn out, yes, in theory, you can share subfolders of any iCloud

00:25:22   folder, even those weird app ones which I just didn't think was the case, with

00:25:26   another user, even if they're on Windows. So my question is, now look I'm not

00:25:32   trying to make anyone change platform right? How impossible would it be for

00:25:37   your assistant to do this work on an iPad? So we have also been exploring iPad

00:25:41   as an option. I think that that's the way to go. I would tell you right now, do not

00:25:46   trust the Windows tool. No, you don't trust it? I just don't. Okay so iCloud

00:25:50   file sharing has been a problem on the Mac in the past, right? Like it has

00:25:57   suffered from unreliability which is a this is one of those things for me. Do

00:26:02   you remember for a really long time nobody trusted photosyncing? None of us

00:26:06   trusted photosyncing but over time we learned to trust it. I feel this way

00:26:11   about iCloud file sharing because even things like shared iCloud folders took

00:26:18   - It took over a year to be,

00:26:20   like it was supposed to be in a version of iOS

00:26:22   and just never showed up.

00:26:24   And I think it's like just recently started

00:26:26   to actually work.

00:26:27   So like, I'm not even convinced that this is something

00:26:31   that had enough long term user testing on.

00:26:35   - What if I use it for something mission critical though?

00:26:37   - Yeah, that's a good point.

00:26:39   This surely is not an issue that way, right?

00:26:42   So I wouldn't trust this.

00:26:46   (sighs)

00:26:48   with the Windows tool, because, I don't know, man.

00:26:53   (laughs)

00:26:55   So like, this was, I'm just looking it up now, right?

00:26:58   This was supposed to come out with Mac OS Catalina,

00:27:01   and it got delayed on the Mac for a long time,

00:27:03   like file sharing, or folder sharing, actually.

00:27:06   I don't know if you're doing folder sharing

00:27:08   or file sharing here.

00:27:09   - It's folder sharing, it has to be folder sharing.

00:27:11   - This is a relatively new thing.

00:27:13   I wouldn't feel comfortable.

00:27:16   My biggest issue is that if she is actually unable

00:27:19   to just edit in place,

00:27:21   I just can't, I just can only imagine

00:27:27   that's going to cause some kind of problem.

00:27:29   - Yeah, yeah.

00:27:30   - Like some kind of syncing issue with Obsidian.

00:27:32   Like Obsidian's gonna be looking for a file

00:27:34   that doesn't exist.

00:27:35   - I also do have that flagged in my head as like,

00:27:37   okay, this is workable, but this also feels like

00:27:39   the exact kind of thing that can confuse syncing

00:27:43   very easily. You're gonna get some kind of conflict somewhere.

00:27:46   Yeah, someone made a copy, edited the copy, deleted the original, renamed the original

00:27:50   to be the copy. Like that feels great if everyone's computer is online at the same time while

00:27:57   it's happening, but that feels like it could be real bad if someone's computer is offline

00:28:02   and wakes up with a different version of the file. And then it's like, "Wait, what happens

00:28:06   here?" I have in the past noticed that iCloud Drive

00:28:11   syncs not as quickly as Dropbox.

00:28:14   - Yeah, iCloud Drive, I feel very much the same way as you,

00:28:17   where when it first came out, I just didn't trust it at all.

00:28:20   For me as an individual user,

00:28:22   I think it's basically rock solid now,

00:28:25   like I haven't had any problems with it.

00:28:27   - I'm good with it now.

00:28:28   I use it for not, I still use Dropbox,

00:28:29   but like I don't worry about using iCloud Drive

00:28:33   for my own stuff anymore.

00:28:35   Like if an app uses it or whatever,

00:28:37   like I use it as like, okay,

00:28:38   Like I'll have some stuff in here, that's totally fine.

00:28:41   - Yeah, exactly, yeah, I'm perfectly fine with it as well.

00:28:44   But it does still have this thing of iCloud Drive

00:28:47   is never in a rush to synchronize anything.

00:28:50   I don't know, it always feels like iCloud Drive

00:28:52   just spends a lot of time in a cafe, sipping tea,

00:28:57   looking out the window, like reading the newspaper.

00:28:59   And like it'll do its job, but maybe not right now.

00:29:02   You know, it'll do it when it's convenient.

00:29:04   - And that's fine if I can tell it sync now

00:29:09   like I can with Dropbox.

00:29:11   - Yes, exactly.

00:29:12   - But you have no control over iCloud Drive.

00:29:15   It would just get to it when it's ready.

00:29:16   Like if I think something's stuck with Dropbox,

00:29:18   I can pause and unpause it and it will start going.

00:29:21   Or I can look in the little Dropbox app

00:29:24   and it will show me what's syncing.

00:29:26   - Yes.

00:29:27   - iCloud Drive has none of this.

00:29:28   Sometimes there's a little spinner, right?

00:29:31   Or like a little progress bar, but that's all you get.

00:29:33   And like, uh, what's particularly delightful right now is of course, we're recording this.

00:29:38   I'm on the writing computer, which is used for podcasting as well.

00:29:41   And whenever we do that, I have a little shortcut that I run that turns off Dropbox

00:29:46   for four hours, and then it will turn Dropbox back on after we're done recording.

00:29:50   Cause you don't want Dropbox spinning and eating up all the

00:29:52   the internet while that's happening.

00:29:53   Some people don't have a problem with this.

00:29:55   My Dropbox is currently uploading this conversation.

00:29:58   Yes, I know Myke.

00:30:00   We're all very impressed with your super fast internet connection.

00:30:02   (laughing)

00:30:05   - Yes, it's (beep) fantastic for you.

00:30:08   - We had a conversation beforehand

00:30:10   where Gray was telling me about his internet woes,

00:30:13   which is why I am now cyberbullying him.

00:30:15   - Yes, so it is vital for all podcasters

00:30:18   to minimize everything that's using the internet

00:30:21   while you're doing something that's gonna be real time.

00:30:24   - For some podcasters.

00:30:24   - For some podcasters, for all podcasters.

00:30:26   But so in this conversation, of course,

00:30:28   while we're discussing iCloud Drive,

00:30:30   I opened up my iCloud Drive folder

00:30:31   just to look at the Obsidian thing.

00:30:33   And even right now on the bottom,

00:30:35   there's a little message which says,

00:30:38   "iCloud Drive is currently downloading 15 items,

00:30:41   "five gigabytes out of 12 gigabytes."

00:30:44   And it's like, what are you doing, iCloud Drive?

00:30:47   I have no idea what that could even possibly be.

00:30:51   And it's extra funny to me

00:30:53   because we're recording the show.

00:30:55   I have just come back from a two-day trip

00:30:59   where I didn't bring any devices other than my phone.

00:31:02   Like I didn't really take any pictures.

00:31:04   This is the first time I'm using any of the computers.

00:31:07   Also, all the computers have been on the entire time

00:31:12   I was gone and it's like,

00:31:13   hey, this is what gives iCloud Drive that feeling of like,

00:31:16   oh, it's got other stuff to do.

00:31:18   And oh, iCloud Drive, I guess was on vacation

00:31:21   over the past couple of days.

00:31:22   And right now while we're recording a podcast,

00:31:25   it's decided to download 10 items.

00:31:27   like 11.35 gigabytes worth of stuff

00:31:30   like right now it's gonna download.

00:31:33   I don't have any problem with it 'cause it does work.

00:31:35   It syncs, everything is there.

00:31:36   But there is a little bit of like,

00:31:38   what you doing buddy?

00:31:39   Where were you?

00:31:40   What 12 gigabytes of stuff do you have to sync

00:31:44   when I literally haven't touched any of my devices

00:31:47   other than my phone for the past two days?

00:31:49   It's so weird.

00:31:50   - Well, I mean, maybe it's downloading a bunch of files

00:31:53   from someone's Windows PC

00:31:55   that have been accidentally added in.

00:31:57   Yeah, I mean, but that's also where I'd love to see what it's actually downloading so I could go like,

00:32:01   "Oh, it's downloading, you know, system 32 folder."

00:32:03   Like, okay, well then at least I can debug that and I can have some idea of what's going on.

00:32:07   So I don't really know quite where we are in this technical project.

00:32:12   I do think the next step is to try to figure out if there is a way to do it on the iPad.

00:32:18   I wasn't quite sure how to debug that, but it does just occur to me.

00:32:22   I can try to do this as a test with my wife's iPad.

00:32:25   so I can actually just have a device in front of me

00:32:27   that isn't mine and see, okay,

00:32:30   can I make it work on my wife's iPad?

00:32:32   And then I have a prayer of doing tech support

00:32:35   over the phone of how to get this to work for somebody else.

00:32:37   - My recommendation is if you're gonna go down this route,

00:32:40   if it does work, you should keep it to Apple platforms.

00:32:44   - Yeah, that would be ideal, yes.

00:32:46   (laughs)

00:32:47   - One, if there's an issue, you can't debug it.

00:32:49   But two, it's just, if it's going to be supported

00:32:52   well anywhere. This is the best place.

00:32:55   Like I guarantee you no matter

00:32:57   how many people are working on it,

00:32:58   and I'm sure there are really smart

00:33:00   people working on it,

00:33:01   it's just not going to get the

00:33:02   resources for the Windows software

00:33:05   that you're going to get for iPad OS

00:33:07   for the sinking of documents, right?

00:33:10   Oh yeah, for sure. For sure.

00:33:11   And that's just that's just naturally

00:33:12   the way I would expect it to be.

00:33:13   Of course Apple has to make it

00:33:15   absolutely rock solid and like

00:33:16   making you work on Windows is great,

00:33:18   but it it just can't ever be

00:33:19   priority number one.

00:33:20   Yeah.

00:33:21   It would be insane if it was that way.

00:33:23   You just couldn't possibly buy it.

00:33:24   - I mean, especially 'cause the open in place thing

00:33:26   is just sort of freaking me out.

00:33:28   - Yes, no, there's also,

00:33:29   as a long time power user,

00:33:32   that's a real red flag for like,

00:33:35   "Ooh, I don't like this at all."

00:33:36   - 'Cause there was a time

00:33:38   when that was just what iCloud Drive did.

00:33:40   You could not open files in place with iCloud Drive.

00:33:45   And it was one of the things I didn't like about it.

00:33:47   Every time I opened it on another machine,

00:33:49   it would make a duplicate.

00:33:50   and it's just like, this isn't what I want.

00:33:52   Or like if an app wanted to use something

00:33:56   in another app's folder,

00:33:57   there was just no way to access those documents.

00:34:01   - Yeah. - Right?

00:34:02   You would have to copy them.

00:34:03   But now you don't need to do any of that

00:34:04   because those app folders are now just treated

00:34:07   as folders inside of the whole thing

00:34:09   that you're able to freely access.

00:34:11   So it works fine.

00:34:12   But if it's still doing that kind of stuff on Windows,

00:34:16   I'd be a bit like, oh, I don't know.

00:34:17   - Yeah. - I don't think

00:34:18   it's gonna work great.

00:34:19   I will say, if you can get this to work better, it will be better for you than Git.

00:34:24   I'm almost convinced of that.

00:34:25   GIT is not an option.

00:34:27   You know, I now know way more about Git than I ever did before.

00:34:30   And I still feel like, oh no, this is, this is, this is just too much.

00:34:33   But yes, I do want to thank everyone for technical support.

00:34:36   I want to anti-thank Myke for bringing this up at the end of the show, because

00:34:40   it really has gotten under my skin of, I would like to have a system that is much

00:34:45   better for doing hand annotations on the script.

00:34:48   I really do feel like that's a critical part of my workflow that is just, is not where it should be.

00:34:55   I don't know why I'm getting anti-thanked then.

00:34:57   You're getting anti-thanked because I had brushed this to the corner of my mind and I just was mostly able to not think about it.

00:35:03   I'm just here trying to make your work better, you know?

00:35:05   I know!

00:35:06   And I'm getting anti-thanked for it.

00:35:08   It's super annoying!

00:35:09   Like, I didn't have a problem before and now I have a problem, right?

00:35:12   So, the script process may have been slightly suboptimal but I was able to deal with it.

00:35:17   And now I feel like, "Oh great, I have this really annoying technical problem to try to

00:35:21   solve with a bunch of tools that might not want to work well together to make this happen."

00:35:24   So…

00:35:25   I actually think that what you just said then is just a pretty good summary of our show

00:35:30   and working life together in general.

00:35:33   Oh yeah?

00:35:34   I didn't have a problem, now I have a problem.

00:35:38   I feel like that sums everything up really nicely.

00:35:44   This episode is brought to you by FitBud.

00:35:47   Between balancing all of the things you have going on in your life, whether it's your

00:35:50   social life, your work life, your family life and everything in between, this can make it

00:35:55   hard to make fitness a priority.

00:35:59   Especially if it's not working for you, maybe you've been trying to do something

00:36:02   and you're not really feeling it, this is why you need Fitbud.

00:36:05   Because Fitbud has an algorithm that will learn about you, your goals and training ability

00:36:09   that you have, so it can craft a personalised exercise plan that is unique to you.

00:36:13   Their app makes it incredibly easy to learn how to perform every exercise.

00:36:18   When you have that plan that is made just for you, that's when it's going to stick.

00:36:22   You'll see the results you are looking for.

00:36:25   They use data to create and adjust the dynamic fitness plan so you can have instant access

00:36:29   to it from anywhere with your own personalised routine in their fantastic app which makes

00:36:34   making progress on your goals really fantastic, really simple and possible from wherever you

00:36:40   are.

00:36:41   Everybody's fitness path is different. Fitbod does all this work to make sure they can customise

00:36:46   things exactly to suit you because they know it's important. They make sure to learn from

00:36:50   your last workout so your next is going to be even better, whether you work out twice

00:36:54   a day or twice a week, it doesn't matter. Fitbod will also track your muscle recovery

00:36:58   to make sure things are balanced with a variety of exercises so you're not overworking or

00:37:02   underworking anything. I really love how easy their app is to use, I love how easy it integrates

00:37:07   of my Apple Watch, so when I'm in the workouts I can advance to the next ones really easily

00:37:12   and if I want to go and see what this new workout is all about then I'm going to be

00:37:15   doing I can go and watch their really awesome videos that they have which makes it super

00:37:19   simple for you to learn every exercise.

00:37:22   Personalised training of this quality can be expensive. Fitbod is just $12.99 a month

00:37:26   or $79.99 a year but you can get 25% of your membership by signing up right now at fitbod.me/cortex.

00:37:35   So go there today and get your customized fitness plan.

00:37:38   That is fitbod.me/cortex to get 25% off your membership.

00:37:43   That's F I T B O D.me/cortex for 25% off.

00:37:47   Our thanks to fitbod for the support of this show and relay FM.

00:37:51   Speaking of problems, I think Apple's gonna help me out.

00:37:54   How's Apple gonna help you out?

00:37:55   This might be the best thing Apple has done for me personally ever.

00:37:59   Oh yeah?

00:38:00   Okay.

00:38:01   announced that I think starting next year

00:38:04   They are no longer going to require

00:38:07   People to put email addresses in RSS feeds for podcasts. Oh

00:38:12   Wow, okay. I was like what on earth could this be? All right, so I

00:38:17   spoke about this when I was complaining when I had like a real just like

00:38:21   breakdown about email a while ago on the show

00:38:24   Mm-hmm, because I was just fed up with the amount of absolute I get my email inbox every day and

00:38:31   And I know where it's coming from.

00:38:33   It's like people getting my email address

00:38:35   or Relay FM email addresses from podcast RSS feeds.

00:38:39   And they just get added to mailing distribution lists.

00:38:42   And I get spam, endless spam about,

00:38:45   hey, do you want this cryptocurrency expert

00:38:47   on your podcast, right?

00:38:49   Like that's just what I get all day, every day.

00:38:52   And now Apple is saying,

00:38:55   no longer will we be requiring this

00:38:58   because they've now like, you know,

00:39:00   In the intervening time, Apple now has a whole system

00:39:04   for submitting podcasts to Apple Podcasts

00:39:07   called Podcast Connect, which is like,

00:39:09   basically it's based I think on what the App Store Connect

00:39:12   thing is, where you submit stuff and it gets reviewed

00:39:15   and then you can view statistics and all that kind of stuff

00:39:17   about your shows.

00:39:19   They've now made it that basically that's what they care

00:39:22   about is you have an account on Podcast Connect

00:39:25   and you can have multiple users and all that kind of stuff.

00:39:28   And so that's just kind of how a contact email address

00:39:31   for Apple is given.

00:39:33   And that no longer needs to be confirmed in the RSS feed.

00:39:37   I will be intrigued to see,

00:39:39   I haven't actually looked into this,

00:39:40   like how are they getting ownership claim of feeds?

00:39:44   'Cause I think that was part of it, right?

00:39:45   That like, this is something that a lot of platforms

00:39:49   have used and it's gonna be interesting

00:39:51   to see how this changes.

00:39:52   They're like, you submit an RSS feed

00:39:55   and then to a platform,

00:39:57   the platform looks for the email address in the feed,

00:40:00   then contacts that email address with a code,

00:40:03   which you then validate as like, I own this feed.

00:40:05   So I haven't actually looked into that.

00:40:08   I'm sure that they've,

00:40:08   but like just like a thought that popped

00:40:10   into my mind right now, but whatever,

00:40:12   I'm sure there's a system.

00:40:13   But the reason this is good for me is we will now be able

00:40:16   to remove email addresses, hopefully, from our RSS feeds,

00:40:20   which will stop some of the crap email that I get,

00:40:25   because I'm sure I'm already on a bunch of lists,

00:40:28   so I know I'm gonna get it forever,

00:40:31   but there will be a selection of new lists

00:40:33   that my email address won't be on,

00:40:35   which I'm excited about.

00:40:36   - Yeah, you can kind of stop the future spread of this

00:40:41   is the main thing.

00:40:42   - Like eventually, I might be able to weed out

00:40:45   a lot of the spam, right?

00:40:46   And it's just not new spam that's coming.

00:40:49   And there'll still be some, but it'll be less,

00:40:51   like in over time, I can kind of get rid of it,

00:40:53   get rid of it. So I'm excited about that.

00:40:55   Yeah, I'm happy to hear about that as well.

00:40:57   I was going to talk about this either on State of the Apps or in the theme

00:41:01   episode, but might as well mention here,

00:41:02   just so people can truly understand the scale of this issue of like, Oh,

00:41:06   when you have to have your email address in a publicly accessible

00:41:11   database like this with Apple,

00:41:13   I made a real push over the past few months to try to clear up my email and try

00:41:18   to get back on top of the system for a bunch of reasons.

00:41:21   And in the past three months, I made a note when I started so

00:41:25   that I could see where this went.

00:41:27   I have deleted over 100,000 emails.

00:41:32   And it's like a huge portion of that is obviously coming from the fact that like,

00:41:37   oh, my email address has been in this public area and is like accessible.

00:41:43   And you just get, like Myke said, so many weird offers and just yeah, like this

00:41:49   whole level of emails where they're directed at you, but they're not personal to you.

00:41:55   It's just monstrous to try to deal with, to actually find the useful emails under

00:42:01   this enormous dump of things.

00:42:03   And it's part of the reason why, like, I think when people hear like, Oh, I just

00:42:06   have largely ignored my email for huge amounts of time, people are like, how can

00:42:10   you do that?

00:42:10   It's like, well, it's just because the, like the signal to noise ratio is so awful.

00:42:16   The majority of that wasn't.

00:42:18   straight up spam, like my little like how many messages do you think have thing

00:42:22   doesn't track the spam messages, that's like the stuff that is deleted

00:42:26   it's all this weird semi-spam stuff that that comes through through this kind of thing so

00:42:32   I'm very happy to hear about that too so yes I would like to get my email out of those lists

00:42:36   if remotely possible. This is one of those things that talking about the software

00:42:40   we will be able to make that change for relay FM shows because we control our feeds

00:42:44   But then it's up to other podcast platforms and systems to also decide that they want to make that change.

00:42:50   Oh yeah, yeah, yeah.

00:42:52   But hopefully they will and it will make things easier for a lot of people.

00:42:57   While we're talking about big companies and podcasting initiatives,

00:43:01   some time ago we spoke about the idea of YouTube getting into podcasting.

00:43:07   It's like a thing that we were just wondering, like, what is this sleeping giant going to do?

00:43:13   And I don't know if they've actually done anything,

00:43:17   but there's been a bit of a hubbub in the podcasting community

00:43:20   about the fact that YouTube now has a dedicated page,

00:43:25   which is youtube.com/podcasts.

00:43:27   And if you're in America, currently,

00:43:31   if you're in the USA, it shows you a page

00:43:34   where they have curated a bunch of podcasts

00:43:38   and made some playlists for them.

00:43:40   I used a VPN and took a screenshot of this page at some point,

00:43:45   which I've put in the notes for you and I'll put in the notes for other people.

00:43:48   It's not really much of anything right now.

00:43:52   >> Okay.

00:43:52   >> It's like, here are a bunch of popular podcast careers,

00:43:56   and here are a bunch of playlists of different types, and there you go.

00:44:05   And so they're kind of curating podcasts,

00:44:09   which I'm not really sure why they feel the need to do this what seems manual curation,

00:44:14   kind of seems antithetical to what YouTube is, but it definitely feels like there is manual

00:44:20   curation going on here, because I have not found, unless you have a way to tell

00:44:26   YouTube, "Hey, we're a podcast," right?

00:44:29   - So, okay, so here's what I wonder about this.

00:44:36   So when we talked about it last time, I'm really big on like, "Oh, if I was in charge of YouTube, I think podcasts is a big growth area for them, actually."

00:44:45   I think there's a lot that they can do here. I genuinely really want YouTube to get into the podcasting area, both as a producer and also as a listener,

00:44:54   because I think they can really help solve the discovery problem, which is part of what podcasts are the worst at now.

00:45:01   - See, I don't know if I agree with you. - Okay, why don't you agree with me?

00:45:05   Because people are already doing it. YouTube don't need to have a hand in this at all.

00:45:10   What do you mean?

00:45:11   People are already putting video versions of their podcasts or like people are already treating like

00:45:16   here's a bunch of people in a room we have microphones we have cameras this is now a podcast

00:45:21   and it's on YouTube right?

00:45:22   Right yeah yeah.

00:45:23   This is already happening and the YouTube algorithm is serving this content to people.

00:45:29   YouTube does not need to have a podcasting initiative of any kind for this to happen on its own.

00:45:34   So like, I would prefer them to do nothing,

00:45:39   because my concern is, if they try and turn their hand to this,

00:45:43   it's going to upset a bunch of stuff.

00:45:45   What are you worried about getting upset?

00:45:47   It's another big platform saying they do podcasts when it's...

00:45:53   Okay, I'm getting a bit tired of the debate of like,

00:45:55   a podcast is only a podcast if it's in an RSS feed.

00:45:58   I've said this a bunch of times,

00:45:59   but I feel like now it's just been repeated too many times.

00:46:02   Right, it's just like, whatever.

00:46:03   But a YouTube video is not a podcast.

00:46:08   It just isn't.

00:46:09   Like if I can't get it in a podcast app,

00:46:13   it's not a podcast.

00:46:15   It's just here is a long YouTube video.

00:46:18   And my concern is like Spotify,

00:46:22   YouTube will be encouraging people

00:46:24   to just publish their stuff on their platform.

00:46:27   That's the part that bugs me, right?

00:46:29   The RSS thing, whatever,

00:46:31   but a podcast should be available

00:46:33   wherever you get your podcasts.

00:46:35   That's my kind of conceit on it.

00:46:37   The particular is about exactly how, I don't care about,

00:46:41   but I do believe that this is a type of content

00:46:44   which is best served when people can get it

00:46:46   wherever they want to get it.

00:46:47   Now that could mean that you do what we do.

00:46:50   We have a podcast we also publish on YouTube, right?

00:46:53   That I'm cool with.

00:46:54   And if YouTube stays out of this space,

00:46:58   this keeps happening, right?

00:46:59   people maybe make it their YouTube video,

00:47:02   and then they release the audio, right?

00:47:05   My concern is if YouTube keeps pushing in,

00:47:07   they're gonna go, "No, no, no, no, no, no.

00:47:09   "Don't do that part.

00:47:10   "Here's $20 million.

00:47:11   "Just keep it on our platform."

00:47:15   - Okay, so you're worried they're going to provide

00:47:17   financial incentive to be,

00:47:18   'cause I was trying to think,

00:47:19   well, why wouldn't someone just also publish it

00:47:21   as an RSS feed? - Right.

00:47:22   They're gonna do what Spotify's doing.

00:47:24   - Hmm.

00:47:25   - 'Cause my point is I think people say,

00:47:28   Myke you're saying people should stop making one argument but it's the same argument.

00:47:31   I don't think that it's the same thing.

00:47:34   Because we participate in what are considered closed platforms.

00:47:38   You can get all of our podcasts on Spotify but that's because we're not entering any

00:47:42   particular deal with them.

00:47:44   They're just another destination for us to publish our shows.

00:47:48   But if Spotify said to us like if you want to put your stuff here we're the only place

00:47:53   that you can publish it.

00:47:54   We'd be like no I don't want to do that.

00:47:56   now we're not participating in the Neopen ecosystem.

00:48:00   And so if YouTube just continue doing what they're doing right now, people can choose

00:48:04   to do whatever they want, right?

00:48:06   They can put their video version on YouTube, they can put their audio, maybe their video

00:48:10   on Spotify, they can put their audio wherever they want, right?

00:48:14   It can go into Apple podcasts, it can go into every single other third-party podcast app.

00:48:19   My concern is if they're like, "Hey, we have a podcasting initiative.

00:48:23   "Hey, Logan Paul, here's 50 million dollars.

00:48:27   "Now you just publish your podcast with us."

00:48:30   - Right. - You know what I mean?

00:48:31   And that's the thing I don't want to happen,

00:48:35   and I would worry would happen

00:48:37   if YouTube continues pushing into this arena.

00:48:40   - Right, okay, I guess I think that is fair

00:48:42   because I have run across a couple of channels

00:48:45   that even call what they're doing a podcast,

00:48:47   and then I discover there is literally no place

00:48:49   to watch it other than YouTube.

00:48:51   - And that's not a podcast to me.

00:48:52   Now it's just a YouTube show, which is great, like go for it.

00:48:57   If you're not on Apple Podcasts and Spotify,

00:49:00   like you're not doing it, right?

00:49:01   It's not a podcast anymore, but yeah.

00:49:04   - What's particularly weird about that is I found

00:49:06   a couple of examples that are just audio only too.

00:49:09   They're not even doing the, oh, we have cameras

00:49:12   and we have that fancy whatever Rode microphone

00:49:15   or the Shure SMB or whatever the heck it is.

00:49:18   They're not even doing that.

00:49:19   It is just, it's like what we do with Cortex,

00:49:21   is just audio only and then I've gone to search for adding it to overcast and it's like oh there's

00:49:27   no RSS feed at all that I always think is very interesting like what are you guys doing yeah

00:49:31   why are you just publishing the audio to youtube it seems very strange to me so okay I think that's

00:49:38   a that is a fair fear I think I think you're not wrong there I guess I just I feel like youtube is

00:49:45   less bad because it, I don't know, maybe it's wrong.

00:49:49   It just feels much better than something like a Spotify.

00:49:52   - Sounds like Stockholm Syndrome to me.

00:49:54   - What I was just trying to think here is like,

00:49:56   can I actually come up with an actual reason why?

00:49:59   And nothing is leaping into my head.

00:50:02   - YouTube just works for you, right?

00:50:04   So like to you, YouTube's fine 'cause you use it

00:50:08   and understand it and you benefit from it.

00:50:10   You understand how to be successful with it

00:50:13   in a way that I think a lot of people don't.

00:50:16   And so to me, YouTube is really nerve-wracking

00:50:20   because I've built an audience however I've built it,

00:50:24   whatever the way is.

00:50:25   I don't know how it's happened,

00:50:26   but it's happened over time, right?

00:50:28   I don't think I would have been able to do that

00:50:32   the same way with YouTube deciding what's shown to people.

00:50:37   We talk about this a lot, right?

00:50:41   The subscriber and the algorithm.

00:50:42   And I think that my content and the shows that I make have succeeded based on the fact that people made a choice

00:50:48   and then they kept tuning in because there was a new episode in their list

00:50:52   and I feel like if it was YouTube, I don't think that would happen.

00:50:57   And also, one of the things I love about podcasting is the barrier to entry is low.

00:51:02   You just need a microphone and zoom or Skype or whatever, right?

00:51:07   And you can go for it.

00:51:09   With YouTube, that barrier to entry is so much higher.

00:51:14   - Yeah.

00:51:15   - Because as soon as you put video into it,

00:51:18   to get something that looks professional,

00:51:20   so much more time, so much more effort,

00:51:22   than to get a podcast that sounds professional.

00:51:25   You really don't have to do a lot

00:51:26   to get a podcast that sounds good.

00:51:28   With YouTube, it's now a whole different kettle of fish.

00:51:31   'Cause I dug through this page,

00:51:33   as far as I could see, every podcast that was featured

00:51:37   has a video component.

00:51:38   It's not like what we do.

00:51:40   - So one of the reasons why I think this has struck me

00:51:43   as an interesting project that I wonder

00:51:44   if YouTube is getting into,

00:51:46   is that YouTube does have a feature,

00:51:49   which I don't think it's accessible to all channels.

00:51:53   Like I think we can't do it on Cortex.

00:51:55   I think I remember looking for it once.

00:51:56   It may just be in beta.

00:51:58   But they have a way, as a creator,

00:52:01   where you can, they don't say like, "Mark as podcast,"

00:52:05   but I feel like this is functioning what they're doing.

00:52:08   They're asking if the video that you have uploaded

00:52:12   is like an audio only experience,

00:52:15   or if it is something that is a listenable experience.

00:52:18   I forget exactly what their language is.

00:52:20   But what they're asking for there,

00:52:22   and I've seen it on the YouTube app

00:52:24   for stuff that is basically like listening to a thing,

00:52:27   YouTube will change the interface

00:52:29   to act like audio controls.

00:52:31   - It's like their YouTube music.

00:52:32   They show the YouTube music controls,

00:52:34   which is like big play, pause, and skip buttons.

00:52:36   - Yes, yes.

00:52:37   And so that's the thing that has seemed to me interesting.

00:52:40   And like, I wonder looking at this page.

00:52:42   - I don't think they wanna do that.

00:52:44   - You don't think that's what they're up to?

00:52:45   - No, I think they want people to make video podcasts.

00:52:49   I don't think they want audio podcasts.

00:52:51   - Oh, but see, Myke, I want them to want audio podcasts.

00:52:53   - But I don't think that they do.

00:52:55   I really don't.

00:52:56   I think that YouTube know video,

00:52:59   and so they just want video.

00:53:02   And just poking around this page.

00:53:05   - It's a very strange page.

00:53:06   - It seemed incredibly clear to me,

00:53:08   because look, I'm not trying to be that guy.

00:53:10   There were shows that had significantly smaller audiences

00:53:14   than even we do on YouTube that were included in this.

00:53:18   And they were included in it, I believe,

00:53:20   because they make video shows and we don't.

00:53:23   - Right, okay.

00:53:25   - So I don't want podcasting to become a video medium.

00:53:30   As a consumer, I watch lots of podcasts on YouTube.

00:53:35   But I don't want that to be the only way that this content is created, because I think it would stop a lot of great shows from being made.

00:53:44   When I heard this announcement, I was looking around to try to see this podcast page and what it was, and I couldn't find it.

00:53:50   And so now, obviously, that's why it's only in the US. Can't see it over here.

00:53:54   And so I just have this screenshot of yours to go by.

00:53:56   And it is a strange selection of things.

00:54:00   things. The one that particularly strikes me is the one at the very top, which is the

00:54:04   nightly news broadcasts from like clearly a TV channel. And that's a bit like, "Is that a podcast?"

00:54:11   No, I think that's a news program that you could listen to. So maybe you're right in the long run.

00:54:18   Maybe this isn't like the thing that I would be thinking of if I was in charge of YouTube podcasts.

00:54:24   This is more like a watch time initiative?

00:54:27   How can we min/max content that is really long but also still has a visual component

00:54:33   and highlight those as things for people to watch? I don't know.

00:54:37   Hmm. Yeah, I have to say that the more I look at this screenshot, the less I like...

00:54:42   I like it.

00:54:45   So I'm not convinced that this would be a good thing.

00:54:49   I like that we're able to take our podcasts to whatever platform we want and make them available to whoever wants to get them.

00:54:57   And my worry is another big platform coming in and locking that content down.

00:55:04   Because I genuinely think that's bad for everybody. It's bad for the creators, even though they make a bunch of money. Not all of them do.

00:55:11   Yeah, yeah, yeah. It's bad for the creators collectively and it can be very good for individual creators.

00:55:16   - Like especially, yeah, individual creators, whatever,

00:55:19   they've just made their money, like make your money, right?

00:55:22   And I don't begrudge creators that do that individually,

00:55:26   but what I don't like is when there are like companies

00:55:30   that decide they're gonna do this exclusive arrangement,

00:55:33   and now there are a bunch of podcasters who would,

00:55:35   their next move is now hampered.

00:55:38   Like if you're a person who works for a podcast company,

00:55:41   that podcast company decides they're gonna put

00:55:43   all their content on Spotify,

00:55:46   you're now kind of stuck with the amount of people that can be listened to on Spotify,

00:55:51   which is growing all the time, I will say.

00:55:54   Recently, you know, I've been tracking kind of our show, it's like 15% of our audience

00:55:59   now listen to the show on Spotify.

00:56:01   I'm genuinely shocked by the number, I cannot believe it's that big.

00:56:04   I would have guessed like 3% I feel like would have been my first pass at that.

00:56:09   No, it's approaching and in some cases passes what we do on YouTube now.

00:56:15   And so I don't have a problem with that as a creator,

00:56:18   because we choose to be there.

00:56:19   But if we were only on Spotify,

00:56:22   well now we've potentially taken 85% of our audience away.

00:56:26   - Right, right.

00:56:27   - Right, and so these are the things that concern me

00:56:30   for people that work in the industry.

00:56:33   Well now if you work for a large publisher,

00:56:35   they've decided to go all in on one platform.

00:56:37   If you want to now go and start something

00:56:38   on your own one day, you've restricted the amount of people

00:56:41   that could possibly know about you

00:56:43   for you to now go make that move, you know?

00:56:45   So little things like that bother me.

00:56:47   And so I don't want another big company to do this.

00:56:51   And the problem I have is YouTube is,

00:56:57   I think, the company that could have

00:56:59   the biggest negative impact

00:57:01   because of their size and scale.

00:57:03   - And you think that negative impact

00:57:04   would be directly through exclusives?

00:57:06   - Yeah.

00:57:07   They have a lot of money and could write a lot of checks

00:57:12   If they wanted to, they don't even need to pay you cash to just change your rev share.

00:57:16   Yeah, that is true.

00:57:17   Say like, we can make you a ton of money over time if you just publish here.

00:57:21   So stuff like that, it's just like, I don't, I don't like the thought of that.

00:57:25   Really.

00:57:25   Well, you've made me less happy about this project than I was previously.

00:57:32   It's like, Myke, but I would like to discover new podcasts and I think

00:57:39   YouTube could be really good at that.

00:57:40   So do I.

00:57:41   That's my only interest.

00:57:42   And they're like, "Oh."

00:57:43   - But it also is like counter to what you,

00:57:46   like YouTube doesn't want to recommend a channel to you.

00:57:49   - Yeah.

00:57:50   - It wants to recommend one video.

00:57:51   So is it really gonna help?

00:57:53   - That is also a good point.

00:57:55   Even the very top here is like popular episodes.

00:57:58   And again, we can get into this like platonic question

00:58:02   of what is a podcast,

00:58:03   but having individual popular episodes feels like

00:58:08   that does not fit the platonic ideal of what podcasts are?

00:58:13   - It's like if I, you know, YouTube will recommend

00:58:15   an episode of a podcast to me now

00:58:17   because there's some kind of drama surrounding it.

00:58:19   And then I might go, "Ooh, lovely!"

00:58:22   And I watch 20 minutes of that podcast.

00:58:25   And that's the end of anything useful for anybody.

00:58:28   Like, I have not decided I want to subscribe

00:58:31   to that podcast.

00:58:32   People got a big breakout hit,

00:58:33   now they think they're podcast successful,

00:58:35   but it's not changed.

00:58:36   Yeah, and a lot of podcasts, like the very experience depends so much on the fact that

00:58:44   you have already listened to other episodes or like you know the hosts.

00:58:48   That also feels like a fundamental thing of like, "What are podcasts?" as opposed to videos

00:58:53   is like, "Oh, I make a video and it is about a topic and the topic is the primary focus

00:58:59   of the video."

00:59:00   Like a video is about a thing.

00:59:02   There are podcasts that are like that, but there are far fewer of them than how YouTube channels work.

00:59:07   But this is why I'm just trying to think about general characteristics.

00:59:11   And like, "Oh, what makes a general characteristic of a podcast?"

00:59:15   And one of those general characteristics is, I am interested in the host's take on this thing

00:59:23   because I have heard the host talk about lots of other things.

00:59:27   other things. So like I am curious about the opinion of this person who I have some sense

00:59:34   of who they are through long experience, which is at the absolute opposite end of the spectrum

00:59:40   of this is a video about a thing and you don't need any previous experience with the thing

00:59:44   because you're watching a video about the thing. You don't need to know anything about

00:59:48   the person who's talking about it. It's just this is topic.

00:59:51   typical YouTube parlance, I guess like podcasts are kind of more like vlogs in

00:59:56   that way like channels you have to like just start and understand that you're

01:00:02   not gonna get a lot of context for a while until you've maybe gone back and

01:00:05   watch more videos to get more context about this person and why are they doing

01:00:10   these things right now. It's kind of like podcasting right like you could jump

01:00:14   into an episode you could pick it up and start listening and you'll get there

01:00:17   eventually but it might help the fact if you go back and listen to more of it

01:00:21   maybe all of it maybe just a big chunk of it it helps right but like that's one

01:00:26   of the great things about podcasts I think is like once you find a new one

01:00:29   oh there's so much stuff to go and listen to if you enjoy it but it makes

01:00:33   it harder to just jump in then like your typical YouTube video about runway

01:00:38   numbers or whatever mmm like if you if there was like a six part podcast series

01:00:42   about runway numbers that included follow-up from the previous one it's

01:00:45   not as easy to jump into.

01:00:47   - Right, yeah, yeah.

01:00:48   Hmm.

01:00:49   - So I don't know if the medium ultimately

01:00:52   is really that good for YouTube to really push on,

01:00:56   as this is where podcasts should be,

01:00:58   and then recommend them to people of like,

01:01:00   hey, here's a popular episode.

01:01:02   You're just like, so?

01:01:03   Like I never see,

01:01:07   'cause I get recommended all kinds of stuff, right?

01:01:09   Like everybody.

01:01:10   And I'm sure I get recommended,

01:01:13   "Hey, here's a two-hour podcast from these people you don't know."

01:01:16   Like, why? Why would I listen to that?

01:01:19   But sometimes it's like,

01:01:21   "Here's a creator you're already following, and this is their podcast channel,

01:01:25   and they're now talking about this thing. Are you interested?"

01:01:28   And some people go, "Yeah."

01:01:30   Which is honestly like the main reason the Cortex YouTube channel exists, right?

01:01:36   Because sometimes we talk about things in your videos,

01:01:38   and then people get suggested them, and they're like,

01:01:40   "Oh, I'd like to know more about that."

01:01:42   and then they go and listen to the Cortex episode

01:01:44   about whatever video.

01:01:45   But I don't know if that video just works cold to someone.

01:01:52   - Yeah, it's interesting.

01:01:55   You're also just pointing out something that,

01:01:56   for my own podcast listening on YouTube,

01:02:00   that is also true.

01:02:01   I've actually realized, oh, my behavior actually expresses

01:02:05   interest in the guest and total lack of interest

01:02:09   in the hosts.

01:02:10   So YouTube will recommend me stuff like, "Oh, you're interested in Person X.

01:02:14   They've appeared on this show where they're talking to them."

01:02:18   And I guess I never really thought about it, but my experience is almost

01:02:21   entirely like, "I don't care at all about the people who are talking to the person."

01:02:25   In fact, most of the time it's like, they're an obstacle to hearing like the

01:02:29   person that I want to hear discuss something and yeah, I've never, ever

01:02:34   subscribed to any of those kinds of things.

01:02:38   just like, "Oh, YouTube is recommending videos where a guest is making the rounds on podcast

01:02:43   shows or whatever."

01:02:44   So it's like, "Oh, I want to hear the guest talk about the things."

01:02:46   But that absolutely never helps those channels, at least from my perspective, to subscribe,

01:02:50   because I never have.

01:02:52   And I never really thought about that.

01:02:56   But I guess from YouTube's perspective, that fits into what they would be thinking of with

01:03:00   this very top line here of like, "Popular episodes."

01:03:05   is going to be very, again, more on the video side of it, more like the content side, not the

01:03:12   long experience with the hosts side. So then does it push podcasts to be like,

01:03:18   "you'll never guess what"? That's where I'm kind of going with this,

01:03:21   is because even when you mentioned like, "oh, it's more like vlogs on YouTube",

01:03:27   I feel like the golden age of vlogging has totally passed, and it's partly because vlogs

01:03:34   slowly mutated into this same kind of format of, oh, you can't really count on people just

01:03:41   to show up each week to see what a person's up to, even if that person is quite interesting.

01:03:47   Vlogs kind of mutated more in the direction of stunts and yeah, and it's like I'm actually

01:03:56   trying to think.

01:03:57   And just excess.

01:03:58   Yeah, excess.

01:03:59   Excess is a really good way to put it.

01:04:02   Excess of everything.

01:04:04   of everything. An excess of energy from the person who the vlog is about, an excess in

01:04:13   like just horrific lifestyles, like an excess in what they're up to, you know, "we're

01:04:20   skydiving out of an airplane onto another airplane today!" It's like yeah, that's

01:04:23   what everybody does for their weekend. So yeah, I guess, because I think YouTube just

01:04:28   totally caused that with people realizing, again, because the subscribers are not really

01:04:32   you cannot hold on to a regular audience, even if those people are interested in your vlog, like they just won't see that it exists.

01:04:39   And it pushed all of them into more MrBeast end of the spectrum.

01:04:44   And so yeah, I guess...

01:04:47   I guess inevitable what I should have realized at the very start of this is that YouTube, if it gets into podcasts, will push

01:04:54   podcasts into being more like the most viral of YouTube videos,

01:04:59   which is not really at all what I want to get out of a podcast, and so now I'm sad.

01:05:03   This episode of Cortex is brought to you by Wealthfront.

01:05:08   There's so much going on with the economy right now.

01:05:11   Are we in a recession?

01:05:12   Is it a bear market?

01:05:13   What's going on with inflation?

01:05:14   Well I can't tell you the future, but I can tell you about a great place to earn

01:05:18   more in your savings, and that's Wealthfront.

01:05:20   Wealthfront is an app that helps you save and invest your money, and right now you can

01:05:24   earn 2% APY with the Wealthfront cash account.

01:05:27   That's 20 times more interest than if you left your money in a savers account at the

01:05:32   average bank according to fdic.gov. Getting a cash account is easy, it takes just a few

01:05:37   minutes to sign up and then you'll immediately start earning 2% interest on your savings.

01:05:42   And you know how much, I'm a big fan, of great design, it's something I look for

01:05:47   on the apps I use. Wealth

01:05:56   extra $50 bonus with a deposit of $500 or more. There are already nearly half a million

01:06:02   people using Wealthfront to save more, earn more and build long term wealth. So why wait?

01:06:08   Earn 2% on your cash today. Visit wealthfront.com/cortex to get started. That's wealthfront.com/cortex.

01:06:17   This high interest good news has been a paid endorsement from Wealthfront. Our thanks to

01:06:21   Wealthfront for their support of this show and Relay FM.

01:06:25   How much of this AI art stuff going around have you seen?

01:06:29   - Hmm, interesting.

01:06:31   - Yeah?

01:06:31   I mean, okay, so you know what I'm talking about, right?

01:06:33   - Yes, yeah, I know what Dali is.

01:06:36   There's a bunch of them.

01:06:37   Dali, I think, started this, right, into popular culture.

01:06:41   - Yeah.

01:06:42   - And then a bunch of other things came about.

01:06:45   I don't know whether they're connected or whatever.

01:06:47   I know that AI art exists.

01:06:50   I know that the way you generate it

01:06:52   is by just giving a string of text

01:06:56   to one of these AI programs

01:06:59   and it will spit something out for you.

01:07:00   Like I've seen YouTube videos about it.

01:07:02   I have friends that have participated

01:07:05   and soon some searches.

01:07:07   I have honestly don't care for me, right?

01:07:13   Like I know that if people are like, this is fun,

01:07:16   I'm gonna see what comes out when I put this text in.

01:07:19   For me, I have zero interest in it right now.

01:07:24   And I have negative interest in it for some stuff

01:07:29   that probably we're gonna touch on

01:07:32   if we're gonna talk about this.

01:07:33   - Ooh, negative interest.

01:07:34   - Yeah.

01:07:35   - Wait, negative thoughts or negative interest

01:07:38   as in it's below boring?

01:07:40   It's like--

01:07:41   - I think it's interesting, right?

01:07:43   I don't think this is boring.

01:07:44   I think that this is a really fascinating technology thing.

01:07:49   I don't think it's a good thing.

01:07:54   I think it's a bad thing.

01:07:56   - Okay, and let me explain for the listeners

01:07:57   if they're not familiar with this.

01:07:58   'Cause isn't one of these,

01:07:59   I've been kind of asking people about this,

01:08:01   and it's very interesting to see the varied response

01:08:06   where some people are very like,

01:08:08   "OMG, I can't believe what's happening about this."

01:08:10   And other people are like,

01:08:11   "I've never heard of this before."

01:08:12   - Right, right, right.

01:08:13   - The first thing where it really came across my radar is,

01:08:17   I don't know if you've seen, there's a Twitter thread here,

01:08:21   which is the first example that caught my attention

01:08:23   of AI-generated art.

01:08:25   And it's from this program called Dolly.

01:08:28   And yes, the idea with Dolly is,

01:08:31   you give the computer a sentence,

01:08:34   and it tries to generate,

01:08:37   Dolly in particular is like generating artwork,

01:08:39   but it's generating an image to match the sentence.

01:08:42   Dali is apparently the best of these.

01:08:44   Like there's a bunch around, Dali seems to be the one

01:08:48   that is the most advanced, and it's also closed.

01:08:52   Like not everyone can use it, but people get invitations

01:08:55   and then like a limited amount of like questions

01:08:58   that they can give it, I think.

01:08:59   - Yeah, so a lot of this stuff

01:09:01   is incredibly computationally intensive.

01:09:04   So it's like, it is very expensive to run.

01:09:06   You cannot run it very well on just normal computers.

01:09:09   You need like these Amazon style clusters of machines

01:09:12   to actually do this stuff.

01:09:14   But so this particular thread caught my attention because

01:09:17   this thread is incredible.

01:09:19   Someone came up with an idea of, "Hey, I'm gonna try to ask Dolly to make images of Kermit

01:09:26   the Frog in various style movies."

01:09:29   You need a prompt, this is the sentence, so the prompt is, "A still of Kermit the Frog

01:09:34   in Blade Runner 2049."

01:09:37   And so then you get an image which is supposed to match that.

01:09:41   And I have to say, it looks exactly like you would imagine a still of Kermit the Frog in

01:09:46   Blade Runner would look like.

01:09:48   It's it's photorealistic, it captures the style of that futurism.

01:09:54   And what makes this thread really great is they continue onward.

01:09:58   And so it goes, a still of Kermit the Frog in the matrix.

01:10:02   And so here's what Kermit would look like if he was in the matrix.

01:10:06   The one that really caught my attention was a still of Kermit the Frog in Spirited Away.

01:10:12   - It just looks like it's in the movie.

01:10:16   It's unbelievable.

01:10:17   - The Spirited Away one was the most shocking because Miyazaki has a real distinctive art

01:10:24   style and it just absolutely nails it of "oh this is how Kermit the Frog would look like

01:10:32   if he was in a Miyazaki movie."

01:10:34   - Right, but here's the problem, right?

01:10:36   - Mm-hmm. - How?

01:10:37   - Well, let's get to that later.

01:10:39   - Okay, okay. - We'll get to the how later.

01:10:40   - Okay. (laughs)

01:10:41   - But it, like, don't think about it too hard.

01:10:44   I just think the thing to establish here

01:10:48   is that it is possible.

01:10:50   And if you're listening to this show

01:10:52   and you have never heard of this stuff before,

01:10:54   like, go look at this link in the description to see.

01:10:58   It looks much better

01:11:00   than however you're imagining it would be.

01:11:02   - Yeah, it's legit.

01:11:03   Like, it's honestly, you could show me this

01:11:06   and tell me somebody made a puppet and put it,

01:11:09   and I would believe you.

01:11:10   Like, I would believe that it is real in many of these,

01:11:14   or that somebody sat down and drew this.

01:11:17   Like, I would believe you.

01:11:18   - Yeah, you wouldn't look at it and think,

01:11:20   oh, that was obviously a computer-generated image.

01:11:22   - Absolutely not.

01:11:23   For like, 75% of these look perfect.

01:11:27   - And in particular, the kind of weird photo-realistic ones

01:11:30   are interesting, like there's a still of Kermit the Frog

01:11:33   in total recall, which doesn't look like an animation. It really looks like Oh, someone

01:11:38   made a model and then photograph or the Twin Peaks one.

01:11:45   Way down there. Yeah. The other thing is about this that I think is interesting to realize

01:11:51   immediately is since this is a computer and you can just type in sentences, you can generate

01:11:57   hundreds of these very quickly in the time like it would take an actual artist to make

01:12:02   this stuff. You're looking at like, oh, this is thousands of human hours compressed down

01:12:07   to a person's interesting afternoon. So which is the one that you wanted me to find? I don't

01:12:12   know. I can't even Twin Peaks. Twin Peaks. Okay, I've got Bojack Horseman, Borderlands.

01:12:18   It's above that one. Eraser. The eraser head one is also really good. Twin Peaks is above

01:12:24   the Twin Peaks ones. Right. Okay, there we go. Yeah, it just looks like a vinyl puppet.

01:12:28   Yeah, it looks like a vinyl puppet.

01:12:30   These things are really quite good.

01:12:32   This is the first one that caught my attention.

01:12:34   And so I think you're totally right in the timeline of how is this stuff going?

01:12:41   Dolly is the first one that started getting people's attention.

01:12:45   And I think it's very easy to think like Dolly is the best one.

01:12:49   But I've been following this more closely.

01:12:51   And what is really quite amazing is how fast this area is improving.

01:12:57   That Twitter thread is from several months ago, and I'm going to send you one which is

01:13:01   a much more recent one.

01:13:04   This is a Twitter thread that's comparing some of the different programs that are coming

01:13:08   along.

01:13:10   And so Dolly, like you said, is a closed project, you can't get public access to it.

01:13:16   But there's a group that's making another version of this, which is called stable diffusion.

01:13:21   And this is like a publicly available, open source version of the same thing.

01:13:28   And the amount of improvement that this one is making is absolutely freaking breathtaking

01:13:34   of how good stable diffusion is.

01:13:38   And the thing with stable diffusion is unlike Dolly, which is sort of explicitly trying

01:13:43   to be like art, stable diffusion is just trying to make images.

01:13:48   So it's not necessarily trying to be like, "Oh, this is what the Mona Lisa would look

01:13:52   like if Picasso painted it."

01:13:54   I don't understand the distinction.

01:13:56   So I think the distinction can be summed up in one of the examples here, which is if you

01:14:03   scroll a little bit of the way down, the prompt is "behind the scenes shooting of the moon

01:14:09   landing Hollywood studio 1969."

01:14:14   So the prompt there is create an image of the moon landing being faked in 1969.

01:14:22   And I think if you look at those images, Stable Diffusions 1 looks a lot more like just a

01:14:27   natural photograph that someone would have taken.

01:14:31   Whereas the Dolly 2 one almost has a little bit of a feeling of like a Renaissance painting,

01:14:36   right, with the lighting and the way people are staged.

01:14:38   Right, but this is terrible.

01:14:40   Right?

01:14:41   So this is, but do you see what I mean?

01:14:43   Like if someone showed you that stable diffusion image, it's like, "Oh, that just looks

01:14:47   like a photograph that somebody took while the moon landing was being faked?"

01:14:51   - Can't we with Dali though ask it to be photorealistic?

01:14:55   Maybe that wasn't the prompt?

01:14:56   I don't know.

01:14:57   - Yeah, so I think you have to be more explicit with the Dali stuff about trying to make it

01:15:01   look real.

01:15:02   - Right, 'cause like right at the top of this thread, there's one which is the prompt

01:15:05   is, "Film still portrait of an old man, wrinkles, dignified look, gray silver hair."

01:15:11   Like the Dali 2 one is like, I mean stable diffusion is very good too, but like the,

01:15:18   you know what actually looking at them, I think I see what you mean from this one.

01:15:21   The Dali 2 one looks like…

01:15:24   It's like an amazing Pixar still.

01:15:26   If Pixar did it.

01:15:27   Where the stable diffusion one actually looks more like a photograph.

01:15:30   Yeah, the stable diffusion one looks like it's a real person.

01:15:33   Yeah, I get it.

01:15:34   Now, here's the key thing.

01:15:35   At this stage already, what we're talking about is like incredibly subtle, but I think

01:15:40   it is interesting and important that stable diffusion isn't trying to be an interesting

01:15:46   art project stable diffusion is it seems like they're really aiming for make an image of

01:15:52   what we've typed anything like it and it just looks like an actual image it's not art focused

01:15:59   I'll put all these links in the show notes by the way so people can go because you really

01:16:02   should go see these images. Yeah, so I think this is

01:16:06   it's really caught my attention because I cannot think of an area of technology

01:16:15   recently that I have seen progress as fast as it has as this area of a machine is able to make an

01:16:26   image that you couldn't possibly tell that a human had not made.

01:16:32   And I think this is one of these things that is going to have an absolutely massive impact

01:16:39   on the world very quickly.

01:16:42   Like if you scroll down, there's one that's particularly to me really captures the like

01:16:47   what's going to happen with this.

01:16:50   And the prompt is "low poly game asset, Cthulhu monster, isometric view".

01:16:57   And so this has generated what is a very good model for a computer game if you were going

01:17:05   to have a little computer monster and you wanted it to be in this low poly art style.

01:17:10   Well I mean the stable diffusion one didn't do a good job.

01:17:13   Yeah the stable diffusion one there, I think this is actually where you see the very fact

01:17:17   that stable diffusion is trying to do real images is like oh the stable diffusion one

01:17:22   isn't low poly it actually looks much more like a real thing. Dali too has done the best

01:17:26   job there I think of what you've asked for. But that is also it's the most artistic style

01:17:32   like it is the most artistic interpretation mid journey is another one of these projects

01:17:37   that's doing a similar kind of thing and I think the mid journey one is quite interesting

01:17:41   But it's like, man, I have seen this go from, like, very good to getting like incredibly

01:17:49   specific and it seems to be getting much, much faster so quickly.

01:17:53   And I look at this and I think, man, if I was in the graphic design world in particular,

01:18:01   I would be terrified.

01:18:03   And stable diffusion, I couldn't find any good examples this morning.

01:18:07   There's sort of one directly below there,

01:18:09   but they also have a lot of generate clip art

01:18:13   about whatever, and one of the areas of weakness

01:18:16   is often human stuff, but man,

01:18:19   they're coming for stock photography here.

01:18:21   I think this stuff is gonna just absolutely blow through

01:18:24   the whole industry of stock photography

01:18:27   and just destroy its existence very fast

01:18:31   if someone can, instead of searching on a stock photo site

01:18:34   for an image that happens to be the thing that they want,

01:18:37   like actually just type in, you know,

01:18:39   I need three people in a room with this kind of laptop

01:18:42   looking at a presentation that is this sort of thing.

01:18:45   What do you think about all this, Myke?

01:18:46   Like what's your reaction looking at these images?

01:18:49   - I don't like it.

01:18:52   - Okay, why don't you like it?

01:18:53   - I don't think this is a good precedent.

01:18:57   I feel this way about deepfake technology.

01:19:01   I feel this way about audio AI technology,

01:19:05   which people are always trying to pitch me on.

01:19:08   - Oh yeah?

01:19:09   - We get pitches from companies that are like,

01:19:11   why even read your ads anymore?

01:19:13   - Oh.

01:19:14   - Let us just feed the ad copy

01:19:17   into a AI-generated version of your voice,

01:19:21   and you can save all this time.

01:19:23   - Wow, okay.

01:19:24   - Just like, how about fundamentally,

01:19:26   let me tell you why I think that's a terrible thing.

01:19:29   But there are people that want to do that to my voice.

01:19:33   I don't like the idea that somebody could take my voice

01:19:37   and make me say whatever they want, right?

01:19:41   And so that's the concern I have

01:19:43   for this type of technology,

01:19:45   that it can be used to create fake materials

01:19:48   and that moon landing thing is part of it, right?

01:19:51   How will anybody in the future know what's true?

01:19:56   when in seconds you could create an image

01:20:00   which looks real and share it.

01:20:02   We already have enough of a problem

01:20:05   with people misunderstanding what an image means

01:20:08   or misunderstanding what a sentence means.

01:20:12   What are we gonna do when it is impossible

01:20:15   to work out what's true by looking at something?

01:20:19   When someone can force you to have that misunderstanding

01:20:23   based on showing you something

01:20:24   you're supposed to believe of your eyes

01:20:26   because your eyes tell you what's true, right?

01:20:29   You see this image.

01:20:31   Like if you said to me, oh my God,

01:20:32   look at this image that was released of the moon landing.

01:20:35   It was fake.

01:20:36   I would be like, oh my God, he's right.

01:20:39   Because it just looks like that, right?

01:20:43   And so that's part one.

01:20:44   And then part two is, what I care about is artists,

01:20:48   individuals trying to make a living.

01:20:50   And what they want to do is they want to be illustrators

01:20:55   and they want to illustrate things for newspapers or whatever.

01:20:58   And I worry about that entire industry of people that want to create graphic design

01:21:05   people who work with you.

01:21:06   Oh my god, how easy would it be?

01:21:09   Because you have such a defined art style, there's so much of it, you could just make

01:21:14   all of your animations in theory based on feeding it prompts.

01:21:20   And I just don't like any.

01:21:23   I don't like the idea that all of these creative people would be put at risk because I don't

01:21:33   believe in the idea of a computer being able to have the artistic vision of a human. I

01:21:42   just don't think that that's possible. And the bigger issue is all of this is feeding

01:21:47   on the back of people's human beings already existing work right so go back to that Kermit

01:21:53   thing we talk about the Miyazaki movie well it can only generate that because the AI has

01:22:00   been pointed to the Miyazaki movies because there's no way it could know what that means

01:22:04   you can't say to an AI create put Kermit the Frog in Spirited Away if it's never seen what

01:22:10   spirited away looks like. So it can create this... it's like we were to be in a

01:22:17   situation where there is no more new inspiration anymore. We'll draw a line at

01:22:22   2045 that was the last time that human beings were allowed to make anything of

01:22:26   their own and now computers just make it all and we're in this continued refresh

01:22:31   of content that's just made based on everything that came in the hundred

01:22:35   years before it. I don't like this. This is a very fun thing to play around with because

01:22:41   it's fun right now. I don't like the future ramifications of this computer generated content

01:22:48   where there's no human interaction outside of here's a string of text, now it's done.

01:22:56   It makes me uncomfortable. I worry about the future of creative endeavors because of technology

01:23:02   like this?

01:23:03   Yeah, when you say a sentence like you don't think that the computer will be able to generate

01:23:11   new things and it will always be a kind of remix.

01:23:15   I don't go along with that as a general statement.

01:23:19   But that is how AI works, right?

01:23:21   So like, you couldn't create an AI system without it having sucked in all this other

01:23:26   information.

01:23:27   Yeah, yeah, no, don't remember, there's two separate issues here.

01:23:29   Like, there's a lot in this conversation, and there's one side that often comes out

01:23:35   where people are like, "Ah, it'll never have the inspiration of a human."

01:23:39   And I do think that is true now, and that probably will be true for a while, but I don't

01:23:45   think that's a fundamental truth.

01:23:47   I'm not sure we should pursue that anyway.

01:23:52   How about we stop before the point where we allow computers to think on their own?

01:23:57   How about that?

01:23:58   (laughs)

01:24:00   Yes, I didn't expect us to go like this hard

01:24:04   and this fast into this topic,

01:24:05   but I guess I also think why is this so interesting

01:24:09   is because I think this is something

01:24:10   that you can show to people, and I have shown to people,

01:24:14   if you try to have the conversation about,

01:24:16   "Oh, actually, I think AI is a terrifying existential threat

01:24:20   for the human race."

01:24:21   People are like, "That's dumb."

01:24:23   And you're like, "Okay, tell me what you think

01:24:25   computer can't do. And creativity is often very high up on that list. This stuff is doing a bunch

01:24:33   of remixing. There is no denying about that. But this is also just like blows past what a lot of

01:24:39   people think computers would even be possible of doing, creating this kind of art just immediately.

01:24:45   And the thing that I see happen in these conversations is people over the past few months

01:24:51   have been pointing out, they're going, "Oh yeah, the dolly stuff is really good, but

01:24:55   it's not great at textures." You know, "Oh, you can see that the textures don't work perfectly

01:25:00   on a bunch of the art that it generates." It's like, "Okay, yeah, cool." You wait two

01:25:04   months and they release the new version and you go, "Oh, it's way better at textures."

01:25:07   Like it keeps getting better at all of the things. They go, "Oh, it's not good at creating

01:25:11   text on these images." And it's like, "Okay, I've already seen it keep getting better and

01:25:15   better at text at every iteration." And I just think the final version of that for people

01:25:20   is always "oh yeah, but ultimately it can't be creative in a new way" and I just don't

01:25:28   think that that argument will ultimately hold. Like, I think there will be a version of this

01:25:32   that generates new stuff that isn't just remixes of the old things in different kinds

01:25:41   of ways. And there's huge concerns about that.

01:25:44   I see the argument of like human beings create this way.

01:25:49   We create based on what we've seen,

01:25:52   but I just don't understand,

01:25:53   why can't we just continue to let humans do that?

01:25:55   Why do we now need to have machines do it?

01:25:57   Why do we need to have an AI platform

01:26:00   that can create artwork with little effort put in?

01:26:05   I don't know why that's needed in the world, right?

01:26:14   outside of it being a curiosity.

01:26:16   If that's all it ever is, is it just exists

01:26:18   as it currently does, like people share these things

01:26:20   on Twitter, it's like, "Ha ha, look at this funny thing

01:26:22   that I made."

01:26:22   But that's not why people were building this.

01:26:25   - Yeah.

01:26:26   - And I am not comfortable with the idea of

01:26:31   suggesting that the work of artistic people

01:26:40   should be replaced by AI systems.

01:26:45   I'm just not comfortable with that.

01:26:47   - Yeah, I'll give you something

01:26:49   that makes me really uncomfortable.

01:26:50   So I'm gonna send you a link.

01:26:53   There's so many projects in this world,

01:26:55   like it's hard to keep all of the different groups straight,

01:26:57   so I'm simply not even gonna try.

01:26:58   Like there's another group that is working

01:27:01   in this same field, and it's a project

01:27:04   that they're calling Textual Inversion.

01:27:08   And so right now with all of these systems, you kind of type in, oh, Kermit the Frog,

01:27:13   but he's in Blade Runner, and you try Kermit the Frog and a bunch of different things.

01:27:18   But what they what they don't have is any kind of sense of continuity.

01:27:25   So all of these things are just individual, one off prompts.

01:27:29   And either you write a prompt and you get a good image or you don't.

01:27:32   But you can't create like a series of related images because everything just exists independently.

01:27:40   So of course, people were like, "Aha, that's why like this won't go everywhere because

01:27:44   the machine doesn't really understand what you mean by like Kermit the Frog.

01:27:49   You can't create a little animation with a stable Kermit the Frog.

01:27:53   All of these things are just one off."

01:27:54   So it's like, "Oh, well, textual inversion is working to solve that problem where they're

01:28:00   They're working to really nail down a particular concept that can then be expressed by these

01:28:07   machines.

01:28:08   And so they have a bunch of examples on this page of like, hey, if you give the computer

01:28:15   a bunch of images, it's really going to try to nail down pulling out a particular concept

01:28:22   that you want to repeat, and to make this process easier.

01:28:26   And so like, okay, you scroll down on their page, and like, there's a bunch of AI generated

01:28:30   art images, but I just happened to click on one and follow it through.

01:28:36   And I think it gets to the heart of like, "Boy, this is really uncomfortable."

01:28:40   So they have a section called "Learning to represent styles."

01:28:45   You feed it a bunch of somebody's art and you try to teach it,

01:28:50   "This is a style that this person has made."

01:28:54   And so they say like, "Okay, here's these four images that we've inputted."

01:28:57   And now we can say like, "Oh, what if this person was to paint the streets of Paris? What if this

01:29:04   person was to paint an adorable corgi?" And they have this art style that's kind of like a colorful,

01:29:10   psychedelic style. And they happen to credit whose art style that is. And so it's this person called

01:29:17   Queenie Art. And I clicked and went to follow this person. And it's someone on DeviantArt who died

01:29:27   from cancer.

01:29:28   "Great.

01:29:29   Yeah, this is what we want."

01:29:30   Oh!

01:29:31   Like, I presume that this was someone whose permission you got to be part of the project?

01:29:37   No, this is a person who was an artist on DeviantArt and, like, had a very unique style.

01:29:44   And their last post on DeviantArt is from several years ago saying they have stage four

01:29:51   cancer and they're not gonna live for another year.

01:29:53   I'm sorry!

01:29:54   What the f*ck just happened here?

01:29:57   We could talk about like the machines copying people's styles, but this is a particularly

01:30:02   awful example of like, "oh, did the people who put together this project even know?"

01:30:09   Like, are they just-

01:30:10   But don't worry, Gray.

01:30:11   It says "image reproduction authorized for non-commercial use only."

01:30:15   Yeah.

01:30:16   So it's okay.

01:30:17   Yeah, that's the reason it caught my attention is they had this explicit disclaimer like,

01:30:22   We've made new art in this dead person style.

01:30:25   It's for non-commercial use only."

01:30:27   That was the whole reason I clicked the link.

01:30:28   I was like, "Oh, what's the deal with this artist?

01:30:30   I wonder, did they put a bunch of stuff in the creative comment?"

01:30:33   Like I just, I just didn't know.

01:30:34   I was curious.

01:30:35   And then I just-

01:30:36   No copyright intended.

01:30:37   It's fine.

01:30:38   It's like I stumbled upon like, "Oh, they're dead."

01:30:40   I feel like the world has been bumping up against this concept in a variety of different

01:30:47   ways.

01:30:48   The first place you start to see this is like, oh, having dead actors in movies and like,

01:30:53   okay, there's a bunch of ways in which you can kind of clear the rights on this.

01:30:57   But now we have computer programs that are just like, hoovering up everything that exists

01:31:04   on the internet.

01:31:06   And then also being able to like, target an individual.

01:31:10   Okay, that person, we want the machine to make more of that person, and they're dead,

01:31:17   do it and it that feels really awful it feels really awful in in so many ways

01:31:22   it's like an abuse of some kind well like I'm trying to think of the word

01:31:27   but it's like wrong mm-hmm it just feels wrong it's like taking advantage of them

01:31:34   yeah in a way like I don't it's very uncomfortable that's a terrible example

01:31:39   yeah I know right it's like it's one of the worst things I've stumbled upon in

01:31:43   this whole world like and I don't you know I don't know the backstory like you

01:31:46   know, maybe there's some like really reasonable explanation for this.

01:31:50   Hey look, maybe this person will contribute to this system, whatever.

01:31:54   But like, yeah, the point is not necessarily this one.

01:31:57   It is just a proof of the point that this can and will happen.

01:32:01   Yeah.

01:32:02   Right.

01:32:02   Yeah.

01:32:03   And that there's, there's like teams that are trying to solve the very specific

01:32:08   problem of, oh, we want to be able to replicate an exact thing consistently,

01:32:14   a bunch with like a stable concept of like this person's art style like or this object

01:32:20   and and maintain it over time.

01:32:22   This is so messed up.

01:32:24   Yeah it's incredibly messed up and I just keep wondering when one of these systems is

01:32:32   going to come across some kind of copyright constraint and like the Miyazaki one is really

01:32:39   the one that that kind of caught my attention because you feel like Miyazaki's movies, they're

01:32:45   beloved. And as often the case with like, someone who creates art that is beloved, it's

01:32:51   partly because that creator has very specific ideas about how they want their artwork created,

01:32:59   very often how they want it distributed, what circumstances under which they view it. I

01:33:05   I think another good example of this is that I always think of and give huge

01:33:09   respect to is, is the author of Calvin and Hobbes, like Calvin and Hobbes is a

01:33:13   hugely successful strip comic book strip.

01:33:16   And part of the reason is because the creator was incredibly picky

01:33:21   about how is this reproduced?

01:33:24   He like, he never wanted it merchandised.

01:33:26   He never wanted like a bunch of different things done with it.

01:33:29   And that's part of the reason why like, Oh, these comics.

01:33:34   They're really untouched and magical in a way that very few things are in the modern world.

01:33:39   And like, I cannot express how sad I would be if someone was like, "Oh,

01:33:43   hey, I made an AI that just spits out brand new Calvin and Hobbs strips all

01:33:47   the time because I'm sad that the author isn't making them anymore."

01:33:51   It's like, no, please, please don't do that.

01:33:53   That's, that's terrible.

01:33:55   And also kind of like ruins the thing that you're trying to do.

01:33:58   Like, I don't know.

01:34:00   I feel like there's incredible and scary economic impacts for everyone who works in the art

01:34:09   world in any way.

01:34:11   And there's also just massive society impacts coming from this kind of thing in like every

01:34:19   possible direction that you can look.

01:34:21   Like I said, there's this weird kind of taking advantage of people who've put their art out

01:34:27   in the world.

01:34:29   And then there's just the, like you said, the mere question of, man, like if you thought

01:34:34   it was hard to know what things are real before when you can fake photographic, and we all

01:34:41   know, very soon, video evidence of anything is like, man, that is terrible for the world.

01:34:49   I mean, look, if this was such a thing that could be constrained to "isn't this a

01:34:56   fun little curiosity. I have no problem with it. I think it's fun to see this stuff produced,

01:35:03   right? Like, to see people creating Kermit the Frog in the style of Twin Peaks. It's

01:35:10   like this is really funny, right? Like this is like a funny little thing, share it on

01:35:13   Twitter and everybody says, "Hey, look what I put into Dali," right? I have no problem

01:35:18   with that because it's like, well, here's the thing we're doing and it's like, like,

01:35:22   not no problem. I barely have any problems with it. It's just like whatever. We know

01:35:25   it's a fake image, etc.

01:35:28   Yeah, in the same way that if an actual human had created like, "Hey, I'm a real fan

01:35:33   of Kermit the Frog and I'm a real fan of Blade Runner, and look at me, I put them together."

01:35:38   I don't think anyone would really have a problem with that.

01:35:41   It doesn't matter.

01:35:42   It's just like, "Oh, this is fun fan art of things that you like."

01:35:46   But it's the next phase of it that I just can't reconcile in my brain of like, what

01:35:53   comes after this then. What happens to special effects artists? What happens to illustrators?

01:36:01   What happens to podcasters? What happens if we just decide that we're all good with the

01:36:11   AI just doing this for us? I just don't think it's a good thing and we don't need it. Right?

01:36:22   And so I don't really understand what the point of it is, especially when there's going

01:36:29   to be a selection of people that take this technology, package it, and sell it to companies

01:36:40   to be like, "You don't need this anymore, like all these people.

01:36:44   Give us some money, which is less, and you can use our software."

01:36:48   Yeah.

01:36:49   Yeah.

01:36:50   What I have just seen develop recently with this stuff is the beginning

01:36:54   of the commercialization of it.

01:36:56   And so there's a weird website, which I don't entirely understand like how

01:37:00   this, how or why this works, but there's a, there's a website already where people

01:37:04   can like buy or sell good prompts for generating art because right now, since

01:37:11   it is so costly to generate this stuff, like you don't want to mess around

01:37:15   unnecessarily, it's like already people are doing this thing of like, "Oh, I can sell you a pack of

01:37:21   prompts that will help you cut down on the difficulty of figuring out what looks cool."

01:37:27   Actually, there's a good article I found, which is trying to get an image of like a llama playing

01:37:30   basketball. Like the key phrases that you want are dramatic backlighting. Like that's one of

01:37:34   the things that makes this photo of a llama playing basketball awesome. But so I think

01:37:38   that this website is trying to commercialize that, but the most directly commercial thing

01:37:44   that I've seen, which I'll send you, is a Photoshop plugin that connects to stable diffusion.

01:37:52   And so it's like, okay, this is already being worked into professional tools. And you can see

01:38:00   like, oh, using this plugin, a Photoshop expert who doesn't have any artistic skill, can just

01:38:06   start selecting images of the screen and ask stable diffusion to fill in what they actually

01:38:13   want in this image. And I just think like, oh my god, like, here we go. Like, this is

01:38:17   totally the beginning of it has real commercial purposes. It is not just a, like you said,

01:38:25   a fun toy. And given how fast this stuff has developed, like I cannot imagine where this

01:38:30   is going to be a year from now. I keep seeing people say like, Oh, you know, you're not

01:38:35   going to be able to make video out of this. Like, it's just going to be still images.

01:38:39   Like guys, video, I hate to break it to you.

01:38:42   It's just a sequence of still images.

01:38:45   Like it's not magic.

01:38:46   It's just way more computationally intensive.

01:38:49   But I've already seen videos on YouTube that I am absolutely sure are 100% AI

01:38:55   generated where it's like, there's a voice.

01:38:58   It sounds pretty good, but it makes me suspicious.

01:39:00   Like I don't think it's actually a person.

01:39:02   It basically sounds like it's summarizing through language models,

01:39:08   like a Wikipedia article about a topic and they have, you know, a bunch of stock images,

01:39:15   which are mostly connected to the narration, but not entirely.

01:39:19   And it's like, that stuff exists on YouTube already.

01:39:21   Like I'm very certain that those things are AI generated and they're just going to keep

01:39:25   getting better and better and better.

01:39:28   And I was like, I met a guy at a conference who was working on some AI research stuff

01:39:35   And he said that he had instituted this guideline that he was no longer reading or watching

01:39:43   anything that he wasn't absolutely sure was produced by a real person.

01:39:49   At the time, I thought this was kind of crazy, right?

01:39:53   I thought like, "Whoa, dude, you know, you're really like, that's a real stance to take."

01:39:59   And you know, the more I see of the stuff that stable diffusion is doing, the more I

01:40:03   feel like, "Man, I don't know, like maybe that just is a good policy very soon? Like,

01:40:08   only read stuff that you know is written by a person?" Because we just haven't touched on it

01:40:14   here, but I think the art stuff really visually catches people's attention. But there are projects

01:40:21   that do this with language, where they write things. And can you tell that a person did or

01:40:26   did not write this sentence? No, you can't tell. And like the language models are also getting very,

01:40:32   very good, very fast, and they're also really starting to understand human concepts in a

01:40:41   way that's like, "Boy, this is all just absolutely terrifying."

01:40:44   So yeah, I don't know.

01:40:46   Maybe it's significantly less crazy now than it was at the time, or maybe like, "Oh,

01:40:51   this guy just could kind of see what was coming," but yeah, I don't know how we're going

01:40:57   to survive in a world where it is incredibly cheap to produce a huge amount of content

01:41:07   and distribute that content widely compared to what humans can create and distribute.

01:41:14   Like we have enough problems sorting through the world, but it's just dealing with stuff

01:41:20   that humans make.

01:41:21   I don't know how this is going to go down when at some point the majority of material

01:41:26   being generated is not being generated by humans.

01:41:30   I think that is going to be a very confusing world to live in.

01:41:33   - Before we go.

01:41:37   - On that cheery note, for action you can take today for something that is unambiguously

01:41:42   good, Saint Jude.

01:41:45   - Saintjude.org/relay.

01:41:49   Go there now, donate, find out more about fundraising, tune into the Podcast-A-Thon

01:41:53   on September 16th. Be a part of this. Let's cure childhood cancer together.