YouTube2Text

Let's Open Source the World — Brandon Thomas, ArtCraft — Transcript

by AI Engineer · 3,619 words · 562 segments · language en · Watch on YouTube

Full transcript

  1. 0:01[music]
  2. 0:12>> Hey everybody. Thanks for attending my
  3. 0:14talk. I had to put this together late
  4. 0:16last night. I've been in a whirlwind
  5. 0:18after this project went viral of
  6. 0:20unblocking thousands of open source
  7. 0:22developers. We've had VC's approaching
  8. 0:25us. I've been in constant meetings. Like
  9. 0:28this has been
  10. 0:29a sleepless month of activity since this
  11. 0:32happened. So if you're not familiar,
  12. 0:34I kind of cloned Adobe Photoshop and a
  13. 0:37bunch of other apps and that went super
  14. 0:39viral.
  15. 0:40My phone, my inbox, everything has just
  16. 0:43been absolutely lit up and so I I I
  17. 0:46apologize for the structure of the the
  18. 0:48the the talk. This was thrown together
  19. 0:50as quickly as I can, but I've got a lot
  20. 0:51of really exciting stuff to talk about
  21. 0:54and I think it talks about the future of
  22. 0:57not only the industry, but just
  23. 0:59technology as a whole. So
  24. 1:02we're going to open source the entire
  25. 1:03world. Like why not? It's it's something
  26. 1:06that I've always wanted to do as a
  27. 1:08developer. That's something that
  28. 1:09a lot of us in the open source community
  29. 1:11are interested in. Open source powers
  30. 1:15a lot of the you know, technology that
  31. 1:18we're familiar with, but it's never been
  32. 1:19at the forefront. Like Linux has always
  33. 1:21been like a second class citizen. We've
  34. 1:23been pulled into a whole bunch of
  35. 1:26platforms where we don't really have
  36. 1:27ownership and AI kind of gives us a
  37. 1:30chance to take the web back, if you
  38. 1:33will. So
  39. 1:36I might have to do this manually.
  40. 1:39Cool. So a little bit about myself. I've
  41. 1:41been an engineer for about 15 years. I
  42. 1:43spent eight years at Square doing heavy
  43. 1:47hitting infra. So online payments,
  44. 1:50the the authentication stack, basically
  45. 1:52anything that could take down the
  46. 1:54company my team was kind of in charge of
  47. 1:56and it was you know, eight years of
  48. 1:57carrying the pager, a lot of really
  49. 1:58heavy hitting infrastructure
  50. 2:00Um and then I've spent seven years doing
  51. 2:02startups, and that's a whole different
  52. 2:03world. It's a lot of risk, a lot of
  53. 2:05learning like what the market needs, a
  54. 2:08lot of correcting your own assumptions.
  55. 2:11And I've also been um I'm based in
  56. 2:13Atlanta. It's a big film film state.
  57. 2:15I've been very big in the arts. So out
  58. 2:17of college, made lots of films, built
  59. 2:19laser projectors that play video games
  60. 2:21on the side of skyscrapers, that kind of
  61. 2:22thing. And those are the sort of
  62. 2:25customers that our startups or the
  63. 2:27startups that I built have have catered
  64. 2:28to, creatives, artists, filmmakers, and
  65. 2:30so forth.
  66. 2:32Um
  67. 2:33I want to transport everybody back in
  68. 2:34time, back to the very beginnings of the
  69. 2:37web and the internet, the '90s, the the
  70. 2:40the early 2000s. So for us millennials,
  71. 2:42we're going to feel old. And for
  72. 2:46um everybody that's that's younger, this
  73. 2:48is just let me kind of set the stage for
  74. 2:49what the internet was like.
  75. 2:51Um we had a bunch of chat applications
  76. 2:53that were completely open. So IRC is
  77. 2:56like Discord, but you could spin up your
  78. 2:57own servers that you actually own. You
  79. 2:59can use whatever programs to automate
  80. 3:01them and interrogate them without rules.
  81. 3:03All of the commercial
  82. 3:05messaging apps were basically they
  83. 3:07didn't care if you used third-party
  84. 3:09tools to connect to them.
  85. 3:11Every website had data exports. They had
  86. 3:13APIs. You could literally build really
  87. 3:15cool mashups with Reddit and Flickr and
  88. 3:18Twitter. And these were like trending on
  89. 3:21slashdot before Hacker News was a thing.
  90. 3:23And then, you know, in in the more
  91. 3:25modern times, these kind of trickle down
  92. 3:27to
  93. 3:29a much slower pace because the APIs got
  94. 3:31got removed.
  95. 3:33There was a bunch of file sharing. So
  96. 3:35not not just file sharing, you could
  97. 3:37actually buy MP3s and feel a sense of
  98. 3:39ownership in in everything that you were
  99. 3:41doing on the web. And now it's more
  100. 3:43licensing.
  101. 3:45Um
  102. 3:46there was so much open creativity
  103. 3:48because largely the people on the
  104. 3:49internet were technical. They were early
  105. 3:51adopters. So Flash kids were making all
  106. 3:53kinds of animations. There were diffuse
  107. 3:56communities with PHP
  108. 3:58PHPBB and Envision and just just a ton
  109. 4:01of creativity that was happening at that
  110. 4:02time.
  111. 4:04Wikipedia got started sort of in this
  112. 4:06ethos, and we don't really see this kind
  113. 4:09of stuff anymore in the current era and
  114. 4:12incarnation of tech.
  115. 4:14Websites were highly customizable, like
  116. 4:16MySpace at the time. You could actually
  117. 4:18change the HTML and CSS and make it look
  118. 4:20however you want. So this was a totally
  119. 4:23different alien time compared to what we
  120. 4:26have today.
  121. 4:30When Web 2.0 happened, broadband,
  122. 4:31smartphones, it brought everybody into
  123. 4:33the fold. So the technical people got
  124. 4:35sort of pushed to the sidelines. And
  125. 4:37it's not a bad thing. Like we needed
  126. 4:38our, you know, our parents and our
  127. 4:39grandparents to be able to use these
  128. 4:40products. But we sort of we rounded the
  129. 4:43edges on things. We took away all the
  130. 4:44technical pieces. We made everything
  131. 4:46easy for people to adapt because they
  132. 4:49couldn't contribute from a technical
  133. 4:51footing. They didn't want to unpackaged
  134. 4:53things. They didn't know what APIs were.
  135. 4:55And so the platforms and the the the
  136. 4:57protocol like protocols got replaced by
  137. 4:59platforms. And they sort of grabbed as
  138. 5:03many users as they could. And sort of
  139. 5:04the diffuse, experimental early web
  140. 5:06started to disappear.
  141. 5:08The APIs got taken away from us.
  142. 5:11And then we started seeing distribution
  143. 5:13being gatekept by the big providers.
  144. 5:16Like we never once had a phone ecosystem
  145. 5:19where we could freely distribute to like
  146. 5:21like we could on the desktop.
  147. 5:25Things started to cost more and more as
  148. 5:27as startups began to like win market
  149. 5:30share. They realized that they had
  150. 5:31pricing power and
  151. 5:33suddenly your Netflix actually became
  152. 5:34cable once you started subscribing to
  153. 5:36all the streaming services.
  154. 5:38And then they started to get very
  155. 5:40extractive. For companies that maybe
  156. 5:43maybe hit the limits of their growth,
  157. 5:45they started coming up with crazy crazy
  158. 5:47gotchas. So the one that got me is Adobe
  159. 5:50has this
  160. 5:51you can subscribe to a monthly plan and
  161. 5:54they don't tell you this. It's in like
  162. 5:55the very small print. If you cancel your
  163. 5:58monthly plan before the 12 months or 18
  164. 6:00months, they actually charge you up to
  165. 6:02like $200 to unsubscribe, which is
  166. 6:05really inconvenient If you're going to a
  167. 6:06film festival like once a year, making a
  168. 6:09short film, and then suddenly you've got
  169. 6:11an extra $200 bill for an application
  170. 6:13that you were just using over a weekend.
  171. 6:14And that's that's hit so many people and
  172. 6:16it's really frustrating
  173. 6:18that the web that was once really really
  174. 6:21promising and open and free
  175. 6:24has turned into this
  176. 6:27extractive
  177. 6:28owned rented rent-seeking world.
  178. 6:33So it it used to be a very special place
  179. 6:35in the '90s in the 2000s
  180. 6:37and
  181. 6:39more and more has been extracted.
  182. 6:42This is a term that's commonly used in
  183. 6:45shitification. I think that aptly
  184. 6:47describes everything that happened.
  185. 6:51Our company started with the ethos of
  186. 6:53let's build everything open source. Like
  187. 6:55the applications that we built in the AI
  188. 6:56space were
  189. 6:58100% open, not only the desktop apps but
  190. 7:01the actual infrastructure as well. So
  191. 7:03Artcraft originally started as sort of
  192. 7:05like a professionally minded
  193. 7:07Higuysfield. So all of the the
  194. 7:09filmmakers that I know and that I've
  195. 7:10I've I've grown up with, we were making
  196. 7:12a product for them to use rather than a
  197. 7:14UGC centered approach like like
  198. 7:17Higuysfield or or Open Art or so forth
  199. 7:19have. And that enabled us to give them a
  200. 7:22desktop tool that lived on their machine
  201. 7:25where all the files that they were
  202. 7:27generating lived right next to their
  203. 7:28editor.
  204. 7:29And one of the things that we observed
  205. 7:32and my advisors thought I was absolutely
  206. 7:34crazy that artists, you know, they don't
  207. 7:35know what open source is. They don't
  208. 7:36they don't need it. They don't talk
  209. 7:38about it. You know, isn't really
  210. 7:40used by artists. We actually observed
  211. 7:42artists vibe coding their own workflows
  212. 7:45on top of the software that we gave
  213. 7:46them. And that is something that like
  214. 7:48was brand new to all of us. Like
  215. 7:50there were micro drama studios that were
  216. 7:52doing order fulfillment for some of the
  217. 7:54top micro drama startups. They would do
  218. 7:5730 to 50 grand of fulfillment every
  219. 7:59single month and they had big teams
  220. 8:01and no product out there fits their
  221. 8:03needs. Like nobody's going to go and
  222. 8:04workshop and build that specifically for
  223. 8:07the type of delivery that they need. Yet
  224. 8:09they in house have people that know
  225. 8:11cloud code and they can take the APIs
  226. 8:13and they can generate an interface
  227. 8:15that's like that. So we started
  228. 8:16experimenting with completely open
  229. 8:18source, giving it to our customers, and
  230. 8:20they really responded to that. And then
  231. 8:23we thought,
  232. 8:24what if we could turn back time on
  233. 8:26everything? Like all of these things
  234. 8:28that they're consuming and building
  235. 8:29would be much more friendly and and and
  236. 8:32give them way more capacity to create if
  237. 8:35all of the infrastructure was open
  238. 8:37source, if all of the application shapes
  239. 8:40that they wanted to leverage were open
  240. 8:41source. Like they use Premiere, they use
  241. 8:44Adobe. Like what could what could we do
  242. 8:46to give them something that they could
  243. 8:47actually take away all of the surface
  244. 8:49area that they don't use and constrain
  245. 8:51the apps to their specific workflow?
  246. 8:54What would it look like if they could do
  247. 8:55that, you know, five or 10 times for
  248. 8:58each different artist on their team
  249. 8:59that's doing a particular part of that
  250. 9:01workflow? And so we wanted to kind of
  251. 9:04undo what had happened during Web 2.0
  252. 9:07and everything became owned and and
  253. 9:09consolidated and and
  254. 9:11gate kept and just give people back
  255. 9:13their applications. And AI really is the
  256. 9:17means for us to do that. So we took a
  257. 9:20stab at the Adobe Creative Suite.
  258. 9:24We looked at Photoshop first. We
  259. 9:25basically prompted, you know,
  260. 9:27this is the basic shape of a photo
  261. 9:29editing app. You you've consumed and
  262. 9:31trained on all of the tutorials about
  263. 9:33what Photoshop is. That's in your
  264. 9:35training data.
  265. 9:36We we laid out a large large prompt to
  266. 9:39say let's let's build out the the
  267. 9:42application. Let's separate the UI UX
  268. 9:44from that. Let's make everything MCP and
  269. 9:46CLI compatible. Let's build out plans
  270. 9:49for every single tool, how all of those
  271. 9:51work. Let's talk about the data
  272. 9:52structures, the file format importers.
  273. 9:54We basically built these large prompts
  274. 9:56and within 24 hours we were looking at a
  275. 10:00pretty good approximation of what
  276. 10:02Photoshop was. And it it kind of blew
  277. 10:04our minds away.
  278. 10:06It was functional. It was very familiar
  279. 10:08to us. It started 10 times
  280. 10:11than the original Photoshop. Now we
  281. 10:12didn't like actually cover the entire
  282. 10:14surface area of Photoshop. There's a lot
  283. 10:16of work. We're still not at 100% parity,
  284. 10:18but we're measuring ourselves against
  285. 10:20what has what needs to be done and we're
  286. 10:22getting closer and closer. We were at
  287. 10:2450%. We've just passed 60% and in the
  288. 10:27fullness of time like Pareto principle,
  289. 10:29we could reach 99.99%
  290. 10:32parity. Um
  291. 10:34I'm going to take a little bit of a sort
  292. 10:35of a diatribe. Our team has talked about
  293. 10:37whether we take this further now or
  294. 10:39whether we just go for parity. And right
  295. 10:41now we're just we want to have we want
  296. 10:43to fulfill the muscle memory of all the
  297. 10:45artists that use Adobe. So we are going
  298. 10:47for Adobe like their sort of interface
  299. 10:49and look and feel directly. In the
  300. 10:51fullness of time we will build net new
  301. 10:53apps that have new interfaces that are
  302. 10:54more agentic. Um but we wanted to do
  303. 10:57that also as sort of like a proving
  304. 10:58function that we have actually
  305. 10:59accomplished something real completely
  306. 11:01agentically.
  307. 11:03Uh so we did Photoshop. We did um
  308. 11:06basically all of their apps. We did
  309. 11:09InDesign. We did Premiere. Um and uh
  310. 11:13this
  311. 11:14I I did not expect this to go viral. I
  312. 11:16was withholding posting it on Hacker
  313. 11:18News. It got out on Reddit and then like
  314. 11:20two days later like the like New York
  315. 11:22Times was asking me for an interview and
  316. 11:23our Discord went from 3,000 people to
  317. 11:26It's now got more people in our Discord
  318. 11:27than Jev, Suno, Higgfield. It
  319. 11:31I I'm still struggling to bear with what
  320. 11:33has happened. This was a It struck a
  321. 11:34nerve with people. Uh they've felt sort
  322. 11:37of maligned by Adobe and other tech
  323. 11:38companies and uh what we've built has
  324. 11:41has really resonated with them and it's
  325. 11:43sort of
  326. 11:44built up sort of this mission to build
  327. 11:45more. Uh so all of the anti-AI artists
  328. 11:49who have been pushing back against AI
  329. 11:50all of this time have you know they've
  330. 11:52sort of turned around and they're like
  331. 11:54we actually love what you're doing.
  332. 11:56They've started using these apps.
  333. 11:57They've started giving us like feature
  334. 11:59requests and bug reports and we're
  335. 12:01engaging them and I think we're going to
  336. 12:03be able to teach them this is what an
  337. 12:05MCP is. This is how you agentically
  338. 12:07create things and maybe we'll start
  339. 12:08pulling away some of these bloated
  340. 12:10interfaces and giving them things that
  341. 12:12are faster and more fluid, and let them
  342. 12:13do their work faster. Um, so we didn't
  343. 12:16stop with those apps. We've done all
  344. 12:18kinds of stuff. We've done CAD, we've
  345. 12:20done um uh uh uh uh Flash, we've done uh
  346. 12:25a bunch of digital audio workstations.
  347. 12:27It's not just us, it's the community
  348. 12:29doing these things as well. And there's
  349. 12:31really no limit in terms of what can be
  350. 12:33done. Uh so we've done the entire
  351. 12:35Microsoft well the core apps within the
  352. 12:36Microsoft Office suite. There's been
  353. 12:38Google Docs. We have uh companies and
  354. 12:41universities and all sorts of people
  355. 12:43approaching us wanting to use these
  356. 12:45things. And we're we're basically trying
  357. 12:47to fulfill all of these different asks.
  358. 12:50Um they're being used on Linux where
  359. 12:51they've never been used before. Um PDF
  360. 12:54Craft is probably the best app. You
  361. 12:55should all go download it right now.
  362. 12:56You'll never have to use Acrobat again.
  363. 12:58This is like really fast,
  364. 13:01really smooth.
  365. 13:03[laughter]
  366. 13:06It works on Linux, it works on NetBSD.
  367. 13:08It's it's brilliant. Like this one's
  368. 13:10ready to go. So go ahead and get this
  369. 13:11one. Um we did Figma. We've actually
  370. 13:14done the server component for this too
  371. 13:17and you can basically have your team
  372. 13:19collaborate on on documents. It's got
  373. 13:21like the raft paxos like multi-user
  374. 13:24setup. So we're starting to go into
  375. 13:25infrastructure too. And you know I the
  376. 13:27entire infrastructure and data model all
  377. 13:29of it is open source. We're not keeping
  378. 13:30any of it proprietary.
  379. 13:32Um
  380. 13:33there's a community that's working on
  381. 13:35Solid Craft that's led by engineers.
  382. 13:37They've actually got uh an open
  383. 13:40uh engineering kernel that manages all
  384. 13:42of this. Uh and it's it's really solid.
  385. 13:44Like the MCP you can actually do 3D
  386. 13:46printing projects really quickly with
  387. 13:48Solid Craft.
  388. 13:49Um
  389. 13:51we've had so so much happen. We've got
  390. 13:53over 3,000 300,000 uh GitHub stars.
  391. 13:564,000 different people have been
  392. 13:58contributing to this project and it's
  393. 14:00picking up. It's it's it's actually
  394. 14:02become sort of a movement if you will.
  395. 14:04And we're we're trying to adapt to how
  396. 14:06much influx of of people trying to
  397. 14:08contribute to this. I'm going to pick up
  398. 14:09the pace cuz I've been kind of lagging
  399. 14:11on this. But um uh we're We're about 1.7
  400. 14:14million downloads a day and this is
  401. 14:16accelerating. We've got a launcher app
  402. 14:18that puts us on the desktop, and like
  403. 14:19you can you can basically manage
  404. 14:21everything here. It works on all
  405. 14:22operating systems.
  406. 14:24A quarter of our users are on Linux,
  407. 14:26which is really fascinating. Like Linux
  408. 14:28has been historically really, really
  409. 14:30underserved, and now they have super
  410. 14:33high quality professional apps for
  411. 14:36creative and and office work, which is
  412. 14:39it's really nice. So
  413. 14:41we've you know, we we love and respect
  414. 14:43all of the work that's been done there
  415. 14:45in the past, but people are really
  416. 14:46excited to use the apps that they didn't
  417. 14:48have access to.
  418. 14:51All of these things are MCP and CLI like
  419. 14:53first class. You can totally command
  420. 14:55these, and you can watch them build and
  421. 14:57assemble creative projects right in
  422. 14:58front of you.
  423. 15:00They've been localized into languages
  424. 15:02that were never like officially
  425. 15:04supported by Adobe or Microsoft. Like we
  426. 15:06were told that Arabic was not supported
  427. 15:08in InDesign. We've also supported these
  428. 15:11apps past their support lifecycle. Like
  429. 15:13Adobe Animate was being
  430. 15:16end of life or it was being like put on
  431. 15:18life support. We've we've picked that
  432. 15:19up. We've had basically animation
  433. 15:20studios come to us saying, we use this
  434. 15:23for our jobs. Like this is important.
  435. 15:25Please please support this. So a lot of
  436. 15:28a lot of really fascinating things are
  437. 15:29happening where we're getting this into
  438. 15:31more places than they've ever been
  439. 15:32before. It's it's definitely not me.
  440. 15:34Like I I started this and now we've got
  441. 15:36a community of like 4,000 developers.
  442. 15:38This is taken on a life of its own. They
  443. 15:40are doing everything from CI/CD. It's
  444. 15:43it's incredible. This this is way bigger
  445. 15:45than just me.
  446. 15:46I really got to pick up the pace here.
  447. 15:47I've made way too much content. So
  448. 15:51there are 300 apps being developed by
  449. 15:52the community. We've developed an app
  450. 15:54where non-technical people can start
  451. 15:56contributing and use the basis of the
  452. 15:58apps that we've created like the UI/UX
  453. 16:00libraries. We're decomposing this into
  454. 16:02all of the the components and primitives
  455. 16:03so that you can reuse them. So you'll
  456. 16:05have nice
  457. 16:073D libraries, nice 2D libraries. It's
  458. 16:09all great. I'll get into the technical
  459. 16:11stuff. So
  460. 16:13Rust was probably essential in this
  461. 16:15product's product success. We'd seen
  462. 16:16people doing it in Electron before, but
  463. 16:18for high-quality creative apps, you
  464. 16:20really need something that's native. Um,
  465. 16:23there's just a lot of really, really
  466. 16:24good quality things about Rust.
  467. 16:26Google had papers about like the low
  468. 16:28defect rate. It starts really fast. It's
  469. 16:31it's the best language for this.
  470. 16:33Uh,
  471. 16:35yeah. Um,
  472. 16:37it supports all all different platforms
  473. 16:40and operating systems. We've got this on
  474. 16:42NetBSD. We've got this on tablets. It
  475. 16:44natively deploys to the web via Wasm, so
  476. 16:47it can get anywhere and everywhere.
  477. 16:50Um, do you need to know Rust to start
  478. 16:52using Rust today as an engineer? Uh,
  479. 16:55definitely not. Uh, I've been using Rust
  480. 16:57since pre-1.0 times, but everybody on
  481. 16:59the team uh it's a mix of people that
  482. 17:01have never experienced the language
  483. 17:02before. Your agents will drive it just
  484. 17:04fine. You're not going to have any
  485. 17:05trouble. It is a fantastic language. The
  486. 17:07one big thing you're going to run into
  487. 17:08is compile times. Um,
  488. 17:11is it GUI yet? There's this famous
  489. 17:12website or series of websites in the
  490. 17:14Rust Rust world. Does it support like
  491. 17:16modern features? Yes. You can build GUI
  492. 17:18rich GUI apps in in Rust, UI UX. Uh,
  493. 17:21we've run into a lot of gotchas like
  494. 17:22permissions, GitHub, scaling, CI. A lot
  495. 17:24of people have talked about that. I'm
  496. 17:25going to just speed through this cuz I'm
  497. 17:26kind of running out of time. Uh, CI is
  498. 17:29like falling over.
  499. 17:30[laughter] We've been spinning up our
  500. 17:32own. We found a good place to get RAM on
  501. 17:34co-located servers. We're made of that.
  502. 17:35Um, this is a good one. So like agentic
  503. 17:38develop versus human uh review. If
  504. 17:40you're doing backend work or things that
  505. 17:41require schema migrations, API design,
  506. 17:44let your humans do that, but for like
  507. 17:45desktop apps and games and anything
  508. 17:47that's basically immutable, throw your
  509. 17:49agents at it and give them like full
  510. 17:50confidence. You don't need to be human
  511. 17:52in the loop fully. Like the upstream
  512. 17:54like directing where things go, just
  513. 17:55just let it run. Tell it to build
  514. 17:57things. Somebody in our community is is
  515. 17:58building Uh, they've got 40 machines
  516. 18:00rebuilding Unreal Engine in Rust. So you
  517. 18:04you don't even have to mind these
  518. 18:05things. Um, Autocraft is a completely
  519. 18:09autonomous software factory that we're
  520. 18:11going to open source. It ingests issues,
  521. 18:13it triages them, it does agentic
  522. 18:14development, it post-trains any vowels
  523. 18:16on all of the decisions. This will be
  524. 18:17open source soon. Um, basically we've
  525. 18:20got an entire app factory that's
  526. 18:22designed in code. You can express this.
  527. 18:24We'll open source this soon. I'm just
  528. 18:25going to have to hurry cuz I'm overtime.
  529. 18:27Uh
  530. 18:28injustice issues.
  531. 18:31This is open source code rabbit open
  532. 18:33source all the things. We're we're doing
  533. 18:34a lot here and
  534. 18:36I I thought I would save more time for
  535. 18:37the technical, but this is really good
  536. 18:39and please please look out for this. Um
  537. 18:41it will support all of your agentic
  538. 18:43development. It supports all of the sort
  539. 18:45of crafting and curating how you uh
  540. 18:47approach a project. It keeps your
  541. 18:48engineers on task. Um again fully open
  542. 18:51source. All of the decisions made by it
  543. 18:53are training data that we're going to
  544. 18:55use for evals and post training on the
  545. 18:56decision trees. Um
  546. 18:59let's take back the entire internet.
  547. 19:00Let's rebuild everything. All software,
  548. 19:03all platforms.
  549. 19:10We can we can turn all of these things
  550. 19:12into proper protocols. Um we can open
  551. 19:15source the entire world. Let's own our
  552. 19:17apps, our infra, our protocols, our
  553. 19:19models, all of it.
  554. 19:20Um
  555. 19:22we are hiring. So if anybody is
  556. 19:24interested and wants to work on 100%
  557. 19:26open source, none of our code is closed
  558. 19:27source. All of it is in the open. Um uh
  559. 19:30you get time to work on your own
  560. 19:31projects. Just reach out to me at hello
  561. 19:33at storyteller.ai and thank you so much
  562. 19:36everybody.

About this transcript

This page contains the full transcript of Let's Open Source the World — Brandon Thomas, ArtCraft by AI Engineer, generated from the public captions YouTube serves with the video. The transcript has 3,619 words across 562 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.