1
01:20:56,280 --> 01:20:58,400
I had to put this together late last night.

2
01:20:58,520 --> 01:21:01,460
I've been in a whirlwind after this project went viral

3
01:21:01,460 --> 01:21:04,980
of unblocking thousands of open source developers.

4
01:21:05,500 --> 01:21:07,720
We've had VCs approaching us.

5
01:21:07,840 --> 01:21:09,420
I've been in constant meetings.

6
01:21:09,680 --> 01:21:14,420
This has been a sleepless month of activity since this happened.

7
01:21:14,600 --> 01:21:18,440
So if you're not familiar, I kind of cloned Adobe Photoshop

8
01:21:18,440 --> 01:21:21,800
and a bunch of other apps, and that went super viral.

9
01:21:22,240 --> 01:21:26,560
My phone, my inbox, everything has just been absolutely lit up.

10
01:21:26,680 --> 01:21:30,820
And so I apologize for the structure of the talk.

11
01:21:31,060 --> 01:21:32,920
This was thrown together as quickly as I can,

12
01:21:33,000 --> 01:21:35,560
but I've got a lot of really exciting stuff to talk about.

13
01:21:35,740 --> 01:21:40,500
And I think it talks about the future of not only the industry,

14
01:21:40,580 --> 01:21:42,140
but just technology as a whole.

15
01:21:42,400 --> 01:21:45,520
So we're going to open source the entire world.

16
01:21:45,880 --> 01:21:46,720
Like, why not?

17
01:21:46,720 --> 01:21:50,140
It's something that I've always wanted to do as a developer.

18
01:21:50,300 --> 01:21:53,800
That's something that a lot of us in the open source community are interested in.

19
01:21:54,140 --> 01:22:00,520
Open source powers a lot of the technology that we're familiar with,

20
01:22:00,520 --> 01:22:01,840
but it's never been at the forefront.

21
01:22:02,200 --> 01:22:04,220
Linux has always been a second-class citizen.

22
01:22:04,660 --> 01:22:09,560
We've been pulled into a whole bunch of platforms where we don't really have ownership.

23
01:22:09,960 --> 01:22:15,300
And AI kind of gives us a chance to take the web back, if you will.

24
01:22:15,300 --> 01:22:19,600
So I might have to do this manually.

25
01:22:20,820 --> 01:22:21,280
Cool.

26
01:22:21,420 --> 01:22:22,240
So a little bit about myself.

27
01:22:22,600 --> 01:22:24,940
I've been an engineer for about 15 years.

28
01:22:25,080 --> 01:22:29,300
I spent eight years at Square doing heavy-hitting infra,

29
01:22:29,520 --> 01:22:33,840
so online payments, the authentication stack,

30
01:22:34,000 --> 01:22:36,060
basically anything that could take down the company,

31
01:22:36,060 --> 01:22:37,680
my team was kind of in charge of,

32
01:22:37,680 --> 01:22:39,960
and it was eight years of carrying the pager,

33
01:22:39,960 --> 01:22:41,720
a lot of really heavy-hitting infrastructure work.

34
01:22:42,400 --> 01:22:44,500
And then I've spent seven years doing startups,

35
01:22:44,500 --> 01:22:45,940
and that's a whole different world.

36
01:22:46,160 --> 01:22:49,280
It's a lot of risk, a lot of learning what the market needs,

37
01:22:49,720 --> 01:22:52,100
a lot of correcting your own assumptions.

38
01:22:52,560 --> 01:22:55,120
And I've also been, I'm based in Atlanta.

39
01:22:55,120 --> 01:22:56,760
It's a big film state.

40
01:22:56,920 --> 01:22:58,660
I've been very big in the arts.

41
01:22:58,880 --> 01:23:01,220
So out of college, made lots of films,

42
01:23:01,420 --> 01:23:03,200
built laser projectors that play video games

43
01:23:03,200 --> 01:23:04,800
on the side of skyscrapers, that kind of thing.

44
01:23:04,940 --> 01:23:08,700
And those are the sort of customers that our startups

45
01:23:08,700 --> 01:23:10,480
or the startups that I built have catered to,

46
01:23:10,620 --> 01:23:13,240
creatives, artists, filmmakers, and so forth.

47
01:23:14,500 --> 01:23:16,900
I want to transport everybody back in time,

48
01:23:17,060 --> 01:23:19,840
back to the very beginnings of the web and the Internet,

49
01:23:19,880 --> 01:23:23,420
the 90s, the early 2000s.

50
01:23:23,600 --> 01:23:26,200
So for us millennials, we're going to feel old.

51
01:23:26,400 --> 01:23:29,420
And for everybody that's younger,

52
01:23:29,720 --> 01:23:32,600
just let me kind of set the stage for what the Internet was like.

53
01:23:33,480 --> 01:23:36,680
We had a bunch of chat applications that were completely open.

54
01:23:36,980 --> 01:23:38,560
So IRC is like Discord,

55
01:23:38,620 --> 01:23:40,900
but you can spin up your own servers that you actually own.

56
01:23:40,940 --> 01:23:44,420
You can use whatever programs to automate them and interrogate them.

57
01:23:44,500 --> 01:23:50,440
without rules. All of the commercial messaging apps were basically, they didn't care if you

58
01:23:50,440 --> 01:23:56,060
used third-party tools to connect to them. Every website had data exports. They had APIs.

59
01:23:56,060 --> 01:24:01,740
You could literally build really cool mashups with Reddit and Flickr and Twitter. And these

60
01:24:01,740 --> 01:24:07,060
were like trending on Slashdot before Hacker News was a thing. And then, you know, in the

61
01:24:07,060 --> 01:24:13,140
more modern times, these kind of trickled down to a much slower pace because the APIs

62
01:24:13,140 --> 01:24:19,420
got got removed. There's a bunch of file sharing. So not not just file sharing, you can actually

63
01:24:19,420 --> 01:24:23,940
buy mp3s and feel a sense of ownership and and everything that you were doing on the

64
01:24:23,940 --> 01:24:30,800
web. And now it's more licensing. There was so much open creativity because largely the

65
01:24:30,800 --> 01:24:35,100
people on the internet were technical. They were early adopters. So flash kids were making

66
01:24:35,100 --> 01:24:41,600
all kinds of animations. There were diffuse communities with PHP, PHP BB and envision

67
01:24:41,600 --> 01:24:45,060
and just a ton of creativity that was happening at the time.

68
01:24:46,200 --> 01:24:49,040
Wikipedia got started sort of in this ethos,

69
01:24:49,040 --> 01:24:52,600
and we don't really see this kind of stuff anymore

70
01:24:52,600 --> 01:24:54,940
in the current era and incarnation of tech.

71
01:24:55,960 --> 01:24:58,020
Websites were highly customizable.

72
01:24:58,340 --> 01:25:01,600
Like MySpace at the time, you could actually change the HTML and CSS

73
01:25:01,600 --> 01:25:03,180
and make it look however you want.

74
01:25:03,360 --> 01:25:07,080
So this was a totally different alien time

75
01:25:07,080 --> 01:25:08,960
compared to what we have today.

76
01:25:11,440 --> 01:25:15,640
When Web 2.0 happened, broadband, smartphones, it brought everybody into the fold.

77
01:25:15,800 --> 01:25:18,580
So the technical people got sort of pushed to the sidelines.

78
01:25:18,740 --> 01:25:19,940
And it's not a bad thing.

79
01:25:20,080 --> 01:25:23,080
Like, we needed our parents and our grandparents to be able to use these products.

80
01:25:23,500 --> 01:25:25,960
But we sort of, we rounded the edges on things.

81
01:25:26,080 --> 01:25:27,580
We took away all the technical pieces.

82
01:25:27,740 --> 01:25:33,640
We made everything easy for people to adapt because they couldn't contribute from a technical footing.

83
01:25:33,780 --> 01:25:35,320
They didn't want to unpackage things.

84
01:25:35,320 --> 01:25:36,820
They didn't know what APIs were.

85
01:25:36,820 --> 01:25:41,940
And so the platforms and the protocols got replaced by platforms,

86
01:25:41,940 --> 01:25:46,200
and they sort of grabbed as many users as they could.

87
01:25:46,360 --> 01:25:49,560
And sort of the diffuse experimental early web started to disappear.

88
01:25:49,900 --> 01:25:52,400
The APIs got taken away from us,

89
01:25:52,440 --> 01:25:57,920
and then we started seeing distribution being gatekept by the big providers.

90
01:25:57,920 --> 01:26:03,020
Like, we never once had a phone ecosystem where we could freely distribute to,

91
01:26:03,080 --> 01:26:04,460
like we could on the desktop.

92
01:26:06,460 --> 01:26:13,160
things started to cost more and more as startups began to like win market share they realized that

93
01:26:13,160 --> 01:26:18,240
they had pricing power and suddenly your Netflix actually became cable once you started subscribing

94
01:26:18,240 --> 01:26:24,240
to all the streaming services and then they started to get very extractive for companies

95
01:26:24,240 --> 01:26:29,760
that maybe maybe hit the limits of their growth they started coming up with crazy crazy gotchas

96
01:26:29,760 --> 01:26:37,340
So the one that got me is Adobe has this, you can subscribe to a monthly plan, and they don't tell you this.

97
01:26:37,420 --> 01:26:38,680
It's in like the very small print.

98
01:26:38,860 --> 01:26:46,480
If you cancel your monthly plan before the 12 months or 18 months, they actually charge you up to like $200 to unsubscribe,

99
01:26:46,480 --> 01:26:51,520
which is really inconvenient if you're going to a film festival like once a year making a short film,

100
01:26:51,560 --> 01:26:56,280
and then suddenly you've got an extra $200 bill for an application that you were just using over a weekend.

101
01:26:56,280 --> 01:26:59,380
And that's hit so many people, and it's really frustrating.

102
01:26:59,760 --> 01:27:05,020
that the web that was once really, really promising and open and free

103
01:27:05,020 --> 01:27:13,260
has turned into this extractive, owned, rented, rent-seeking world.

104
01:27:14,420 --> 01:27:19,040
So it used to be a very special place in the 90s and the 2000s,

105
01:27:19,120 --> 01:27:22,740
and more and more has been extracted.

106
01:27:23,660 --> 01:27:27,180
This is a term that's commonly used in shitification.

107
01:27:27,180 --> 01:27:30,380
I think that aptly describes everything that happened.

108
01:27:32,480 --> 01:27:36,440
Our company started with the ethos of let's build everything open source.

109
01:27:36,660 --> 01:27:41,520
Like the applications that we built in the AI space were 100% open,

110
01:27:41,700 --> 01:27:44,560
not only the desktop apps, but the actual infrastructure as well.

111
01:27:44,780 --> 01:27:49,600
So Artcraft originally started as sort of like a professionally-minded Higgs field.

112
01:27:49,700 --> 01:27:53,360
So all of the filmmakers that I know and that I've grown up with,

113
01:27:53,360 --> 01:27:57,820
we were making a product for them to use rather than a UGC-centered approach,

114
01:27:57,820 --> 01:28:01,840
like Higgs field or OpenArt or so forth have.

115
01:28:02,100 --> 01:28:07,340
And that enabled us to give them a desktop tool that lived on their machine

116
01:28:07,340 --> 01:28:10,720
where all the files that they were generating lived right next to their editor.

117
01:28:11,260 --> 01:28:16,220
And one of the things that we observed, and my advisors thought I was absolutely crazy,

118
01:28:16,360 --> 01:28:18,220
that artists, you know, they don't know what open source is.

119
01:28:18,300 --> 01:28:20,400
They don't need it. They don't talk about it.

120
01:28:20,800 --> 01:28:22,680
You know, GIMP isn't really used by artists.

121
01:28:22,680 --> 01:28:28,320
we actually observed artists vibe coding their own workflows on top of the software that we gave them.

122
01:28:28,400 --> 01:28:30,980
And that is something that was brand new to all of us.

123
01:28:31,620 --> 01:28:37,680
There were micro drama studios that were doing order fulfillment for some of the top micro drama startups.

124
01:28:38,140 --> 01:28:41,740
They would do 30 to 50 grand of fulfillment every single month.

125
01:28:41,840 --> 01:28:45,340
And they had big teams and no product out there fits their needs.

126
01:28:45,460 --> 01:28:50,740
Nobody's going to go and workshop and build that specifically for the type of delivery that they need.

127
01:28:50,740 --> 01:28:57,600
Yet they in-house have people that know cloud code and they can take the APIs and they can generate an interface that's like that

128
01:28:57,600 --> 01:29:01,700
So we started experimenting with completely open source giving it to our customers

129
01:29:01,700 --> 01:29:05,600
And they really responded to that and then we thought

130
01:29:05,600 --> 01:29:13,880
What if we could turn back time on everything like all of these things that they're consuming and building would be much more friendly and

131
01:29:13,880 --> 01:29:19,720
And give them way more capacity to create if all of the infrastructure was open source

132
01:29:19,720 --> 01:29:23,940
if all of the application shapes that they wanted to leverage were open source?

133
01:29:24,260 --> 01:29:26,240
Like, they use Premiere. They use Adobe.

134
01:29:26,400 --> 01:29:30,300
Like, what could we do to give them something that they could actually take away

135
01:29:30,300 --> 01:29:32,520
all of the surface area that they don't use

136
01:29:32,520 --> 01:29:35,400
and constrain the apps to their specific workflow?

137
01:29:35,600 --> 01:29:37,740
What would it look like if they could do that, you know,

138
01:29:38,400 --> 01:29:41,340
five or ten times for each different artist on their team

139
01:29:41,340 --> 01:29:43,700
that's doing a particular part of that workflow?

140
01:29:43,940 --> 01:29:48,880
And so we wanted to kind of undo what had happened during Web 2.0

141
01:29:48,880 --> 01:29:54,080
and everything became owned and consolidated and gatekept

142
01:29:54,080 --> 01:29:56,340
and just give people back their applications.

143
01:29:56,640 --> 01:30:00,000
And AI really is the means for us to

144
01:30:00,000 --> 01:30:00,440
do that.

145
01:30:00,640 --> 01:30:07,840
So we took a stab at the Adobe Creative Suite. We looked at Photoshop first. We basically

146
01:30:07,840 --> 01:30:13,540
prompted, you know, this is the basic shape of a photo editing app. You've consumed and

147
01:30:13,540 --> 01:30:17,620
trained on all of the tutorials about what Photoshop is. That's in your training data.

148
01:30:18,940 --> 01:30:24,980
laid out a large, large prompt to say, let's build out the application. Let's separate

149
01:30:24,980 --> 01:30:26,400
the UI UX from that.

150
01:30:26,580 --> 01:30:29,460
Let's make everything MCP and CLI compatible.

151
01:30:29,820 --> 01:30:32,240
Let's build out plans for every single tool,

152
01:30:32,320 --> 01:30:33,360
how all of those work.

153
01:30:33,500 --> 01:30:34,860
Let's talk about the data structures,

154
01:30:35,020 --> 01:30:36,080
the file format importers.

155
01:30:36,220 --> 01:30:38,080
We basically built these large prompts.

156
01:30:38,160 --> 01:30:40,520
And within 24 hours, we were looking

157
01:30:40,520 --> 01:30:44,820
at a pretty good approximation of what Photoshop was.

158
01:30:44,960 --> 01:30:47,480
And it kind of blew our minds away.

159
01:30:47,860 --> 01:30:49,020
It was functional.

160
01:30:49,060 --> 01:30:50,440
It was very familiar to us.

161
01:30:50,600 --> 01:30:53,800
It started 10 times faster than the original Photoshop.

162
01:30:53,800 --> 01:30:57,500
Now, we didn't actually cover the entire surface area of Photoshop.

163
01:30:57,600 --> 01:30:59,880
There's a lot of work. We're still not at 100% parity.

164
01:31:00,140 --> 01:31:03,740
But we're measuring ourselves against what needs to be done,

165
01:31:03,780 --> 01:31:05,700
and we're getting closer and closer.

166
01:31:05,720 --> 01:31:08,220
We were at 50%. We've just passed 60%.

167
01:31:08,220 --> 01:31:10,740
And in the fullness of time, like Pareto principle,

168
01:31:11,240 --> 01:31:14,040
we could reach 99.99% parity.

169
01:31:15,960 --> 01:31:18,100
I'm going to take a little bit of a diatribe.

170
01:31:18,100 --> 01:31:20,500
Our team has talked about whether we take this further now

171
01:31:20,500 --> 01:31:21,920
or whether we just go for parity.

172
01:31:21,920 --> 01:31:28,240
And right now, we want to fulfill the muscle memory of all the artists that use Adobe.

173
01:31:28,360 --> 01:31:32,240
So we are going for Adobe, like their sort of interface and look and feel directly.

174
01:31:32,560 --> 01:31:37,300
In the fullness of time, we will build net new apps that have new interfaces that are more agentic.

175
01:31:37,800 --> 01:31:40,440
But we wanted to do that also as sort of like a proving function,

176
01:31:40,560 --> 01:31:43,980
that we have actually accomplished something real, completely agentically.

177
01:31:44,680 --> 01:31:50,480
So we did Photoshop. We did basically all of their apps.

178
01:31:50,480 --> 01:31:51,520
We did InDesign.

179
01:31:51,600 --> 01:31:52,520
We did Premiere.

180
01:31:54,240 --> 01:31:58,080
And this, I did not expect this to go viral.

181
01:31:58,240 --> 01:31:59,980
I was withholding posting it on Hacker News.

182
01:31:59,980 --> 01:32:00,860
It got out on Reddit.

183
01:32:01,000 --> 01:32:02,620
And then, like, two days later,

184
01:32:02,880 --> 01:32:05,460
like, New York Times was asking me for an interview.

185
01:32:05,520 --> 01:32:07,640
And our Discord went from 3,000 people to,

186
01:32:07,840 --> 01:32:09,400
it's now got more people in our Discord

187
01:32:09,400 --> 01:32:11,940
than Jev, Suno, Higgsfield.

188
01:32:13,020 --> 01:32:15,480
I'm still struggling to bear with what has happened.

189
01:32:15,580 --> 01:32:17,280
This was, it struck a nerve with people.

190
01:32:17,880 --> 01:32:19,840
They've felt sort of maligned by Adobe

191
01:32:19,840 --> 01:32:20,760
be in other tech companies.

192
01:32:21,000 --> 01:32:24,680
And what we've built has really resonated with them.

193
01:32:24,760 --> 01:32:27,720
And it's sort of built up this mission to build more.

194
01:32:28,560 --> 01:32:31,760
So all of the anti-AI artists who have been pushing back

195
01:32:31,760 --> 01:32:35,440
against AI all of this time, they've sort of turned around.

196
01:32:35,840 --> 01:32:37,920
We actually love what you're doing.

197
01:32:38,140 --> 01:32:39,500
They've started using these apps.

198
01:32:39,520 --> 01:32:42,360
They've started giving us feature requests and bug reports.

199
01:32:42,600 --> 01:32:44,020
And we're engaging them.

200
01:32:44,140 --> 01:32:46,080
And I think we're going to be able to teach them

201
01:32:46,080 --> 01:32:47,700
this is what an MCP is.

202
01:32:47,700 --> 01:32:49,640
This is how you agentically create things.

203
01:32:49,640 --> 01:32:52,760
And maybe we'll start pulling away some of these bloated interfaces

204
01:32:52,760 --> 01:32:54,980
and giving them things that are faster, more fluid,

205
01:32:55,020 --> 01:32:56,460
and let them do their work faster.

206
01:32:57,760 --> 01:32:59,280
So we didn't stop with those apps.

207
01:32:59,320 --> 01:33:00,680
We've done all kinds of stuff.

208
01:33:00,780 --> 01:33:01,480
We've done CAD.

209
01:33:01,800 --> 01:33:05,340
We've done Flash.

210
01:33:05,620 --> 01:33:09,100
We've done a bunch of digital audio workstations.

211
01:33:09,300 --> 01:33:10,420
It's not just us.

212
01:33:10,420 --> 01:33:12,240
It's the community doing these things as well.

213
01:33:12,380 --> 01:33:15,340
And there's really no limit in terms of what can be done.

214
01:33:15,860 --> 01:33:17,400
So we've done the entire Microsoft,

215
01:33:17,640 --> 01:33:19,460
well, the core apps within the Microsoft Office,

216
01:33:19,460 --> 01:33:21,520
Suite, there's been Google Docs.

217
01:33:21,620 --> 01:33:25,240
We have companies and universities and all sorts of people

218
01:33:25,240 --> 01:33:27,300
approaching us wanting to use these things.

219
01:33:27,760 --> 01:33:30,020
And we're basically trying to fulfill

220
01:33:30,020 --> 01:33:31,440
all of these different asks.

221
01:33:31,860 --> 01:33:34,020
They're being used on Linux, where they've never

222
01:33:34,020 --> 01:33:34,900
been used before.

223
01:33:36,100 --> 01:33:37,540
PDFcraft is probably the best app.

224
01:33:37,560 --> 01:33:38,580
You should all go download it right now.

225
01:33:38,620 --> 01:33:40,360
You'll never have to use Acrobat again.

226
01:33:40,380 --> 01:33:43,660
This is really fast, really smooth.

227
01:33:47,640 --> 01:33:49,160
It works on Linux.

228
01:33:49,300 --> 01:33:50,420
It works on NetBSD.

229
01:33:50,600 --> 01:33:51,360
It's brilliant.

230
01:33:51,540 --> 01:33:52,560
This one's ready to go.

231
01:33:52,660 --> 01:33:53,700
So go ahead and get this one.

232
01:33:55,040 --> 01:33:56,080
We did Figma.

233
01:33:56,280 --> 01:33:58,500
We've actually done the server component for this, too.

234
01:33:58,700 --> 01:34:02,880
And you can basically have your team collaborate on documents.

235
01:34:02,960 --> 01:34:06,040
It's got the wrapped Paxos multi-user setup.

236
01:34:06,260 --> 01:34:08,440
So we're starting to go into infrastructure, too.

237
01:34:08,540 --> 01:34:11,640
And the entire infrastructure and data model, all of it, is open source.

238
01:34:11,700 --> 01:34:13,220
We're not keeping any of it proprietary.

239
01:34:15,220 --> 01:34:18,660
There's a community that's working on SolidCraft that's led by engineers.

240
01:34:18,660 --> 01:34:24,100
They've actually got an open engineering kernel that manages all of this.

241
01:34:24,380 --> 01:34:25,940
And it's really solid.

242
01:34:26,040 --> 01:34:30,580
Like the MCP, you can actually do 3D printing projects really quickly with SolidCraft.

243
01:34:32,400 --> 01:34:34,620
We've had so, so much happen.

244
01:34:34,800 --> 01:34:37,660
We've got over 300,000 GitHub stars.

245
01:34:38,060 --> 01:34:42,520
4,000 different people have been contributing to this project, and it's picking up.

246
01:34:43,260 --> 01:34:45,880
It's actually become sort of a movement, if you will.

247
01:34:46,040 --> 01:34:50,100
And we're trying to adapt to how much influx of people

248
01:34:50,100 --> 01:34:50,920
trying to contribute to this.

249
01:34:51,020 --> 01:34:52,280
I'm going to pick up the pace, because I've

250
01:34:52,280 --> 01:34:53,360
been kind of lagging on this.

251
01:34:53,660 --> 01:34:57,320
But we're getting about 1.7 million downloads a day.

252
01:34:57,380 --> 01:34:58,420
And this is accelerating.

253
01:34:58,720 --> 01:35:00,920
We've got a launcher app that puts us on the desktop.

254
01:35:00,940 --> 01:35:03,480
And you can basically manage everything here.

255
01:35:03,560 --> 01:35:05,260
It works on all operating systems.

256
01:35:06,380 --> 01:35:09,320
A quarter of our users are on Linux, which is really

257
01:35:09,320 --> 01:35:09,940
fascinating.

258
01:35:09,940 --> 01:35:12,940
Linux has been historically really, really underserved,

259
01:35:12,940 --> 01:35:17,180
and now they have super high-quality professional apps

260
01:35:17,180 --> 01:35:21,760
for creative and office work, which is really nice.

261
01:35:22,040 --> 01:35:26,300
So we love and respect all of the work

262
01:35:26,300 --> 01:35:27,620
that's been done there in the past,

263
01:35:27,680 --> 01:35:29,560
but people are really excited to use the apps

264
01:35:29,560 --> 01:35:31,200
that they didn't have access to.

265
01:35:32,960 --> 01:35:35,980
All of these things are MCP and CLI first class.

266
01:35:35,980 --> 01:35:37,580
You can totally command these,

267
01:35:37,580 --> 01:35:41,200
and you can watch them build and assemble creative projects right in front of you.

268
01:35:41,940 --> 01:35:47,760
They've been localized into languages that were never officially supported by Adobe or Microsoft.

269
01:35:47,760 --> 01:35:51,260
We were told that Arabic was not supported in NDesign.

270
01:35:51,460 --> 01:35:54,760
We've also supported these apps past their support lifecycle.

271
01:35:55,180 --> 01:36:00,180
Adobe Animate was being end of life, or it was being put on life support.

272
01:36:00,300 --> 01:36:01,240
We've picked that up.

273
01:36:01,340 --> 01:36:03,480
We've had basically animation studios come to us saying,

274
01:36:03,620 --> 01:36:05,740
we use this for our jobs.

275
01:36:05,920 --> 01:36:06,920
This is important.

276
01:36:06,920 --> 01:36:08,620
Please, please support this.

277
01:36:08,880 --> 01:36:11,940
So a lot of really fascinating things are happening

278
01:36:11,940 --> 01:36:13,380
where we're getting this into more places

279
01:36:13,380 --> 01:36:14,160
than they've ever been before.

280
01:36:14,240 --> 01:36:15,620
And it's definitely not me.

281
01:36:16,020 --> 01:36:18,180
Like, I started this, and now we've got a community

282
01:36:18,180 --> 01:36:20,440
of, like, 4,000 developers.

283
01:36:20,660 --> 01:36:22,140
This has taken on a life of its own.

284
01:36:22,300 --> 01:36:24,560
They are doing everything from CI, CD.

285
01:36:24,900 --> 01:36:25,640
It's incredible.

286
01:36:25,860 --> 01:36:27,220
This is way bigger than just me.

287
01:36:27,760 --> 01:36:29,320
I really got to pick up the pace here.

288
01:36:29,340 --> 01:36:30,560
I've made way too much content.

289
01:36:30,800 --> 01:36:34,860
So there are over 300 apps being developed by the community.

290
01:36:34,860 --> 01:36:39,300
We've developed an app where non-technical people can start contributing and use the

291
01:36:39,300 --> 01:36:42,080
basis of the apps that we've created, like the UI, UX libraries.

292
01:36:42,240 --> 01:36:46,680
We're decomposing this into all of the components and primitives so that you can reuse them.

293
01:36:46,780 --> 01:36:50,760
So you'll have nice 3D libraries, nice 2D libraries.

294
01:36:50,960 --> 01:36:51,500
It's all great.

295
01:36:51,760 --> 01:36:53,160
I'll get into the technical stuff.

296
01:36:53,520 --> 01:36:58,000
So Rust was probably essential in this product's success.

297
01:36:58,160 --> 01:37:02,560
We'd seen people do it at Electron before, but for high-quality, creative apps, you really

298
01:37:02,560 --> 01:37:03,620
need something that's native.

299
01:37:04,580 --> 01:37:07,200
There's just a lot of really, really good quality things

300
01:37:07,200 --> 01:37:07,760
about Rust.

301
01:37:08,360 --> 01:37:11,200
Google had papers about the low defect rate.

302
01:37:11,200 --> 01:37:12,220
It starts really fast.

303
01:37:12,480 --> 01:37:14,300
It's the best language for this.

304
01:37:18,940 --> 01:37:23,200
It supports all different platforms and operating systems.

305
01:37:23,220 --> 01:37:24,560
We've got this on NetBSD.

306
01:37:24,560 --> 01:37:25,580
We've got this on tablets.

307
01:37:25,860 --> 01:37:29,000
It natively deploys to the web via Wasm.

308
01:37:29,000 --> 01:37:31,400
so it can get anywhere and everywhere.

309
01:37:32,840 --> 01:37:36,400
Do you need to know Rust to start using Rust today as an engineer?

310
01:37:37,160 --> 01:37:37,820
Definitely not.

311
01:37:38,360 --> 01:37:40,340
I've been using Rust since pre-1.0 times,

312
01:37:40,440 --> 01:37:41,580
but everybody on the team,

313
01:37:41,700 --> 01:37:44,500
it's a mix of people that have never experienced the language before.

314
01:37:44,760 --> 01:37:46,360
Your agents will drive it just fine.

315
01:37:46,440 --> 01:37:47,320
You're not going to have any trouble.

316
01:37:47,440 --> 01:37:48,540
It is a fantastic language.

317
01:37:48,680 --> 01:37:50,980
The one big thing you're going to run into is compile times.

318
01:37:52,540 --> 01:37:53,460
Is it GUI yet?

319
01:37:53,540 --> 01:37:56,900
There's this famous series of websites in the Rust world.

320
01:37:57,040 --> 01:37:58,580
Does it support modern features?

321
01:37:58,580 --> 01:38:02,140
Yes, you can build rich GUI apps in Rust, UI, UX.

322
01:38:02,680 --> 01:38:04,740
We've run into a lot of gotchas, like permissions,

323
01:38:04,960 --> 01:38:06,000
GitHub, scaling, CI.

324
01:38:06,080 --> 01:38:07,040
A lot of people talked about that.

325
01:38:07,100 --> 01:38:08,180
I'm going to just speed through this,

326
01:38:08,240 --> 01:38:09,200
because I'm kind of running out of time.

327
01:38:09,900 --> 01:38:11,700
CI is falling over.

328
01:38:12,660 --> 01:38:14,040
We've been spinning up our own.

329
01:38:14,100 --> 01:38:16,580
We found a good place to get RAM on co-located servers.

330
01:38:16,700 --> 01:38:17,240
We're moving to that.

331
01:38:17,840 --> 01:38:18,720
This is a good one.

332
01:38:18,920 --> 01:38:21,360
So agentic develop versus human review.

333
01:38:21,520 --> 01:38:24,060
If you're doing back-end work or things that require schema

334
01:38:24,060 --> 01:38:26,720
migrations, API design, let your humans do that.

335
01:38:26,720 --> 01:38:33,040
But for like desktop apps and games and anything that's basically immutable, throw your agents at it and give them like full confidence

336
01:38:33,040 --> 01:38:36,880
You don't need to be human in the loop fully like the upstream like directing where things go

337
01:38:36,880 --> 01:38:40,800
Just just let it run tell it to build things somebody in our community is is building

338
01:38:40,800 --> 01:38:47,940
They've got 40 machines rebuilding unreal engine in rust. So you don't even have to mind these things

339
01:38:49,340 --> 01:38:55,660
Autocraft is a completely autonomous software factory that we're going to open source it ingest issues it triages them

340
01:38:55,660 --> 01:38:56,780
It does agentic development.

341
01:38:56,800 --> 01:38:59,020
It post trains and evals on all of the decisions.

342
01:38:59,260 --> 01:39:00,560
This will be open source soon.

343
01:39:01,580 --> 01:39:04,860
Basically, we've got an entire app factory that's designed in code.

344
01:39:04,860 --> 01:39:05,920
You can express this.

345
01:39:06,060 --> 01:39:07,120
We'll open source this soon.

346
01:39:07,120 --> 01:39:08,620
I'm just going to have to hurry because I'm over time.

347
01:39:09,860 --> 01:39:10,500
Injust issues.

348
01:39:12,500 --> 01:39:15,520
This is open source code rabbit, open source all the things.

349
01:39:15,660 --> 01:39:16,600
We're doing a lot here.

350
01:39:16,720 --> 01:39:20,000
And I thought I would save more time for the technical.

351
01:39:20,100 --> 01:39:20,880
But this is really good.

352
01:39:20,880 --> 01:39:22,460
And please, please look out for this.

353
01:39:23,020 --> 01:39:25,140
It will support all of your agentic development.

354
01:39:25,140 --> 01:39:28,000
It supports all of the crafting and curating

355
01:39:28,000 --> 01:39:29,940
how you approach a project.

356
01:39:29,940 --> 01:39:31,600
It keeps your engineers on task.

357
01:39:32,000 --> 01:39:33,540
Again, fully open source.

358
01:39:33,780 --> 01:39:36,180
All of the decisions made by it are training data

359
01:39:36,180 --> 01:39:38,340
that we're going to use for evals and post-training

360
01:39:38,340 --> 01:39:39,220
on the decision trees.

361
01:39:40,760 --> 01:39:42,040
Let's take back the entire internet.

362
01:39:42,160 --> 01:39:43,360
Let's rebuild everything.

363
01:39:43,540 --> 01:39:45,720
All software, all platforms.

364
01:39:52,160 --> 01:39:55,640
We can turn all of these things into proper protocols.

365
01:39:56,300 --> 01:39:58,260
We can open source the entire world.

366
01:39:58,400 --> 01:40:00,000
Let's own our apps, our infra,

367
01:40:00,280 --> 01:40:02,060
our protocols, our models, all of it.

368
01:40:04,140 --> 01:40:08,300
We are hiring, so if anybody's interested and wants to work on 100% open source,

369
01:40:08,420 --> 01:40:10,840
none of our code is closed source. All of it is in the open.

370
01:40:12,100 --> 01:40:16,800
You get time to work on your own projects, just reach out to me at hello at storyteller.ai.

371
01:40:16,960 --> 01:40:18,380
And thank you so much, everybody.
