WEBVTT

NOTE Sentence-level transcript of https://www.youtube.com/watch?v=xUnRQ9vLXxo

NOTE One cue per sentence. Cue ids are the line anchors on /transcripts/xUnRQ9vLXxo.html. A cue ends where the next begins, or 2 s after its last word.

s1
00:00:01.309 --> 00:00:03.309
[music]

s2
00:00:17.400 --> 00:00:18.360
Hello, hello.

s3
00:00:18.360 --> 00:00:20.120
Fantastic to see you guys here.

s4
00:00:20.120 --> 00:00:31.160
I still can't believe they're letting me take a stage at something like this, a YouTuber apparently, but can't wait to share a bit about how I've been thinking because if I'm being real,

s5
00:00:31.160 --> 00:00:33.200
kind of going through some AI psychosis.

s6
00:00:33.200 --> 00:00:37.240
Who here would classify how they feel right now as some form of AI psychosis?

s7
00:00:37.240 --> 00:00:39.520
I want to see some hands.

s8
00:00:39.520 --> 00:00:42.200
The Those who don't have their hands up yet, don't worry.

s9
00:00:42.200 --> 00:00:45.440
We'll get you there by the end of this talk if I do everything right.

s10
00:00:45.440 --> 00:00:49.440
In order to talk about this, I want to start with a bit of a personal journey of my own.

s11
00:00:49.440 --> 00:00:54.800
And I'm going to go through this the way anybody does in modern timelines, with the models.

s12
00:00:54.800 --> 00:01:01.800
Who here used Sonnet 3.5 when it was the creme de la creme, the cream of the crop model available to us?

s13
00:01:01.800 --> 00:01:04.920
It was unbelievably better than what we had used before, right?

s14
00:01:04.920 --> 00:01:14.360
Like having used all of these different models and trying them in tools, Sonnet 3.5 was a big moment for me because it felt like these models could suddenly complete much more end-to-end

s15
00:01:14.360 --> 00:01:18.760
tasks, like actually get real work done that takes multiple steps.

s16
00:01:18.760 --> 00:01:23.480
And then we got Opus 4.5. Who felt or I'll go a different way differently here.

s17
00:01:23.480 --> 00:01:28.040
Who didn't feel a big jump when they switched over to Opus 4.5?

s18
00:01:28.200 --> 00:01:28.800
That's a relief.

s19
00:01:28.800 --> 00:01:34.880
There were not too many hands because Opus 4.5 is probably when my psychosis started in November and December of last year.

s20
00:01:34.880 --> 00:01:39.520
Having a model that couldn't just write the code and call tools, but could go way further.

s21
00:01:39.520 --> 00:01:46.240
A model that could test the work and actually get it into a good state and complete tasks that take hours instead of minutes.

s22
00:01:46.240 --> 00:01:48.320
It was unbelievable.

s23
00:01:48.320 --> 00:01:49.880
And then we got Mythos.

s24
00:01:49.880 --> 00:01:54.120
Who here has had a chance to play with Mythos and Fable so far?

s25
00:01:54.120 --> 00:01:57.400
We agree it's a pretty damn good model, right?

s26
00:01:57.400 --> 00:01:58.360
But why?

s27
00:01:58.360 --> 00:01:59.840
It's not just better at coding.

s28
00:01:59.840 --> 00:02:05.320
If you handed a prompt that you would have handed to these other models before, it's not going to feel that different.

s29
00:02:05.320 --> 00:02:09.360
I think of these almost as eras now, where Sonnet 3.5 is the tool call era.

s30
00:02:09.360 --> 00:02:20.840
Not that it was the first model that could do tool calls, rather it was the first one that did them consistently and reliably enough in context of a code base where you could use this for day-to-day coding work.

s31
00:02:20.840 --> 00:02:26.360
Then we got Opus 4.5, which was able to do much longer running tasks without losing track of what it's working on.

s32
00:02:26.360 --> 00:02:28.240
It's no longer, "Okay, build step one."

s33
00:02:28.240 --> 00:02:28.840
and then it does it.

s34
00:02:28.840 --> 00:02:30.560
Then you say, "Okay, can you build this next part?"

s35
00:02:30.560 --> 00:02:31.560
and then the next part.

s36
00:02:31.560 --> 00:02:35.200
You can just tell it what you want, and it could figure it out a lot of the time.

s37
00:02:35.200 --> 00:02:39.320
Mythos is another jump to orchestration.

s38
00:02:39.320 --> 00:02:44.760
It feels to me like it's the first model that doesn't just understand your code base, but it understands itself.

s39
00:02:44.760 --> 00:02:51.800
And it knows how to spawn additional models and break up work in a way where it could be completed more reliably and then verified afterwards.

s40
00:02:51.800 --> 00:02:53.880
And if you tell the model to do that, it will just do it.

s41
00:02:53.880 --> 00:02:58.040
You don't need some custom tooling, some custom system, some fancy software factory.

s42
00:02:58.040 --> 00:03:00.080
You just need to prompt it to go a little further.

s43
00:03:00.080 --> 00:03:03.280
I think you'll be surprised how far it can go.

s44
00:03:03.280 --> 00:03:06.680
What I'm trying to say here is we need to go bigger.

s45
00:03:06.680 --> 00:03:12.800
You're not going to see the benefits going forward if you're not pushing the model further, you're not pushing yourself further with what you're building.

s46
00:03:12.800 --> 00:03:22.560
Most of the Jira tickets I closed in my previous job could be trivially solved with a model like Opus 4.5. My previous work would not benefit from a model like Mythos.

s47
00:03:22.560 --> 00:03:29.480
If the models are going to keep getting better, and at this point I'm confident saying we they I was wrong when I claimed that we were hitting a wall before.

s48
00:03:29.480 --> 00:03:31.720
The models are getting better faster than we are.

s49
00:03:31.720 --> 00:03:33.040
So, we can't necessarily get better.

s50
00:03:33.040 --> 00:03:35.280
So, instead we have to go bigger.

s51
00:03:35.280 --> 00:03:37.720
In order to do that, we have to get over ourselves.

s52
00:03:37.720 --> 00:03:42.800
This was really hard for me as someone who spent a long time writing software.

s53
00:03:42.800 --> 00:03:45.360
Who here has written code for more than 10 years?

s54
00:03:45.360 --> 00:03:46.840
I want to see hands.

s55
00:03:46.840 --> 00:03:49.480
That's the majority of the people here.

s56
00:03:49.480 --> 00:03:51.760
I don't want to think about how long I've been writing code for.

s57
00:03:51.760 --> 00:03:55.000
But I have been building up all of these strong opinions since I started.

s58
00:03:55.000 --> 00:03:59.240
I was using GNU screen and eventually tmux back in the day.

s59
00:03:59.240 --> 00:04:02.920
I learned how to use those tools and SSH and get before I even wrote code.

s60
00:04:02.920 --> 00:04:06.320
And those have all been really ingrained in my workflow.

s61
00:04:06.320 --> 00:04:09.640
I think back to the old days in a weird way.

s62
00:04:09.640 --> 00:04:10.720
Hear me out.

s63
00:04:10.720 --> 00:04:12.520
Let me talk about iOS for a second.

s64
00:04:12.520 --> 00:04:15.000
Who owned an iPhone back when they looked like this?

s65
00:04:15.000 --> 00:04:17.640
iOS 6 or earlier?

s66
00:04:17.640 --> 00:04:21.959
How have you guys written code for 10 years when a fourth of you are that old?

s67
00:04:21.959 --> 00:04:22.960
I'm confused.

s68
00:04:22.960 --> 00:04:25.080
Hopefully, you're all Android people or something.

s69
00:04:25.080 --> 00:04:28.000
This is how iPhone apps used to look.

s70
00:04:28.000 --> 00:04:30.080
You might notice it's different from how they look now.

s71
00:04:30.080 --> 00:04:35.560
It looks less like an app and more like somebody took a picture of a compass and put it in the phone.

s72
00:04:35.560 --> 00:04:38.919
This is how apps used to look.

s73
00:04:38.919 --> 00:04:40.919
But, now they look like this.

s74
00:04:40.919 --> 00:04:44.200
And most people look at that and they're like, "Oh, that's an obvious downgrade.

s75
00:04:44.200 --> 00:04:45.480
That's so much worse.

s76
00:04:45.480 --> 00:04:46.400
Why would Apple ever do that?

s77
00:04:46.400 --> 00:04:49.960
This is the downfall of Apple and the beauty of their design."

s78
00:04:49.960 --> 00:04:51.880
I'm going to fight you guys on that.

s79
00:04:51.880 --> 00:05:00.600
iOS 7 was Apple moving away from trying to convince you that these devices can replace the old tools we used to rely on.

s80
00:05:00.600 --> 00:05:07.640
The compass had to look like a compass because the compass had to replace the physical compass that you relied on.

s81
00:05:07.640 --> 00:05:17.600
The books app had to look like a bookshelf with real pages that turned because I had to convince you it was a reasonable alternative to buying a book and reading the paper version.

s82
00:05:17.600 --> 00:05:22.440
Apps had to be designed to convince you to use them, not to be useful.

s83
00:05:22.440 --> 00:05:28.000
And iOS 7 represents the shift to not focusing on convincing you anymore.

s84
00:05:28.000 --> 00:05:29.240
Apple won.

s85
00:05:29.240 --> 00:05:32.320
By that point, everyone knew their iPhone could do all of these different things.

s86
00:05:32.320 --> 00:05:37.120
The point of iOS 7 was to stop convincing and start embracing.

s87
00:05:37.120 --> 00:05:38.440
Start making a better interface.

s88
00:05:38.440 --> 00:05:42.200
And this interface, as much as we might not like how it looks, it's so much more useful.

s89
00:05:42.200 --> 00:05:46.760
You have clear indications of the difference between what direction you're you're on and where you're currently pointing.

s90
00:05:46.760 --> 00:05:51.520
That red block is a super nice way to know that you're not in the direction you intend.

s91
00:05:51.520 --> 00:05:56.120
The current direction you're facing is way clearer, too, with the giant 228 at the bottom.

s92
00:05:56.120 --> 00:05:59.840
You just get way more info here than you did before.

s93
00:05:59.840 --> 00:06:01.400
It's so much clearer.

s94
00:06:01.400 --> 00:06:08.280
Even if we don't like it because it's not the thing we're used to, we got over it.

s95
00:06:08.480 --> 00:06:11.800
We're currently in our skeuomorphic phase as software developers.

s96
00:06:11.800 --> 00:06:21.480
Skeuomorphism is this design aesthetic trying to represent the way things used to look, the physical goods that we relied on, and try to make them digital.

s97
00:06:21.480 --> 00:06:23.760
We're doing this right now with software.

s98
00:06:23.760 --> 00:06:27.680
We're pretending our terminals are the ultimate interface when they're not even good interfaces.

s99
00:06:27.680 --> 00:06:30.960
And I'm saying this as someone who loves their terminal deeply.

s100
00:06:30.960 --> 00:06:35.480
Natural language has no place in a terminal, but we pretend it does because the terminal's familiar.

s101
00:06:35.480 --> 00:06:37.000
It's what we It's what we're used to.

s102
00:06:37.000 --> 00:06:37.680
It's what we love.

s103
00:06:37.680 --> 00:06:41.880
It's where we like to think of ourselves when we're thinking about coding.

s104
00:06:41.880 --> 00:06:47.360
Who here's an aspiring Vim user that like wishes you could use Vim, or even does?

s105
00:06:47.360 --> 00:06:51.040
I know we've all had that Vim phase where we all tried.

s106
00:06:51.120 --> 00:06:52.400
This is just how we are as devs.

s107
00:06:52.400 --> 00:06:54.040
We care so deeply about these things.

s108
00:06:54.040 --> 00:06:58.440
We care so deeply about our tools, the systems we pick, the frameworks, the languages.

s109
00:06:58.440 --> 00:07:00.840
We like to think it all matters.

s110
00:07:00.840 --> 00:07:02.880
And we've blinded ourselves in this.

s111
00:07:02.880 --> 00:07:06.880
We think things that just don't make sense when you take one step back.

s112
00:07:06.880 --> 00:07:10.320
Like, why can't we commit our environment files?

s113
00:07:10.320 --> 00:07:14.760
It sounds stupid when I put it on a slide like this, but I want you to really think about this for a second.

s114
00:07:14.760 --> 00:07:24.720
When I have a team of engineers that are working on a project, why do I have to build another system to share this specific file, but all the other files can go and get just fine?

s115
00:07:24.720 --> 00:07:25.480
It's dumb.

s116
00:07:25.480 --> 00:07:33.680
It's just how Git was built because it was built for a very specific thing, and then it took over our industry, and it took over our brains, and we aren't letting go of that.

s117
00:07:33.680 --> 00:07:38.600
There's a lot of these things in our heads that we have to start fighting.

s118
00:07:38.600 --> 00:07:46.200
We have to take the step back and think, is this how we do things cuz it's right, or is this how we do things cuz it's just how we've always done it?

s119
00:07:46.200 --> 00:07:51.920
And as you start to think more about this, you'll realize there are so many things that we do this with.

s120
00:07:51.920 --> 00:07:55.000
Like, why do we qualify ourselves by the languages we know?

s121
00:07:55.000 --> 00:07:56.160
I used to think this was a junior thing.

s122
00:07:56.160 --> 00:08:00.080
Like, I could can't tell you how many times I had a junior engineer I was talking to who was like, "Oh, you're a coder?

s123
00:08:00.080 --> 00:08:01.040
What languages do you write?

s124
00:08:01.040 --> 00:08:02.760
I write JavaScript."

s125
00:08:02.760 --> 00:08:06.520
I thought this was a junior problem, but then you talk to senior engineers who are like, "Oh, he writes JavaScript.

s126
00:08:06.520 --> 00:08:08.160
He's not a real developer."

s127
00:08:08.160 --> 00:08:09.320
We care too much.

s128
00:08:09.320 --> 00:08:11.680
We pride ourselves in these things.

s129
00:08:11.680 --> 00:08:12.560
They're our identity.

s130
00:08:12.560 --> 00:08:20.440
These weird facts, these weird choices, these things that feel essential just don't matter that much anymore.

s131
00:08:20.440 --> 00:08:22.560
And they didn't then, and they matter less now.

s132
00:08:22.560 --> 00:08:31.240
We got away with it because it was so hard to find engineers that we could just tell the company what we were doing, and they couldn't really say no cuz the alternative is spend 6 months trying to hire someone else.

s133
00:08:31.240 --> 00:08:33.680
They're not going to do that.

s134
00:08:33.680 --> 00:08:37.200
And along that note, why are we so scared of deleting code?

s135
00:08:37.200 --> 00:08:47.440
I cannot tell you how many times I've been in a conversation with someone where the solution is to just delete it and reset, but we have such a bad sunk cost

s136
00:08:47.440 --> 00:08:49.120
mindset in this industry.

s137
00:08:49.120 --> 00:09:00.360
We care so much about the code we wrote, and we care so much about it still being there that I feel bad working with my team sometimes when somebody files a PR that isn't quite the right solution, but they spent a week or two on it.

s138
00:09:00.360 --> 00:09:02.960
Like, who here has guilt merged a PR before?

s139
00:09:02.960 --> 00:09:11.360
Where you just felt bad because somebody put a lot of work in, it's not quite the right thing, but you merge it anyways cuz the alternative was a conversation you didn't want to have.

s140
00:09:12.600 --> 00:09:14.480
Why do we do this to ourselves?

s141
00:09:14.480 --> 00:09:20.760
One of the nice things about ages you don't have to feel bad when you shut down their work, but we we just care too much.

s142
00:09:20.760 --> 00:09:21.640
It's the point I'm trying to make.

s143
00:09:21.640 --> 00:09:29.080
And the things we care about are not necessarily the things that matter anymore, and I hope we can finally start to challenge some of these.

s144
00:09:29.080 --> 00:09:36.760
Going to get a little more personal here by showing some of the ideas I've built cuz the goal here is that when you guys go out of this talk,

s145
00:09:36.760 --> 00:09:44.520
you have a better mindset for coming up with ideas that make sense now by rejecting the things that made sense before.

s146
00:09:44.520 --> 00:09:47.920
These are three of the things that I have built or are currently working on.

s147
00:09:47.920 --> 00:09:49.800
We're going to go from the bottom up.

s148
00:09:49.800 --> 00:09:55.440
I built a Reddit scraper because making good memes is hard and I would rather just steal them from Reddit and went pretty well.

s149
00:09:55.440 --> 00:09:56.280
It was a side project.

s150
00:09:56.280 --> 00:09:57.280
It would take me two to three days.

s151
00:09:57.280 --> 00:10:04.480
Would just scrape Reddit, top posts on programming humor, put them in a nice format for me so I could go copy paste them onto Twitter.

s152
00:10:04.480 --> 00:10:07.480
Zoom for streamers was a startup I went through Y Combinator with.

s153
00:10:07.480 --> 00:10:08.360
It was called Ping.

s154
00:10:08.360 --> 00:10:15.920
I wanted to make it easier for live content creators to do high-quality collaborations in the software they already used, OBS.

s155
00:10:15.920 --> 00:10:20.200
The full stack cloud is let's just imagine Vercel, but it goes further each direction.

s156
00:10:20.200 --> 00:10:22.960
They have off built in, but they also have databases built in.

s157
00:10:22.960 --> 00:10:29.160
These are all things I've wanted that I benefit from existing, enough so that I tried to build all of them.

s158
00:10:29.160 --> 00:10:31.080
The bottom two I built in 2021.

s159
00:10:31.080 --> 00:10:33.640
The top one I'm working on right now.

s160
00:10:33.640 --> 00:10:38.920
These are also kind of tiers, different levels that we can build at.

s161
00:10:38.920 --> 00:10:46.400
If I was to try and categorize them, I would call the bottom one side project, call the middle one startup, and call the top one too big.

s162
00:10:46.400 --> 00:10:48.240
It just doesn't make sense.

s163
00:10:48.240 --> 00:10:52.960
Well, this is how I would have categorized this even just a year ago.

s164
00:10:52.960 --> 00:10:54.680
But things have changed.

s165
00:10:54.680 --> 00:10:58.200
Now that the models are bigger, the tiers have shifted.

s166
00:10:58.200 --> 00:11:02.160
Everything is now one tier lower.

s167
00:11:02.360 --> 00:11:04.760
And this is a crazy thing for me to process.

s168
00:11:04.760 --> 00:11:08.320
The fact that what used to be a startup is now a side project.

s169
00:11:08.320 --> 00:11:17.000
In fact, some of the startups I've talked to even at an event like this, their whole startup could have arguably been a side project or this bottom tier, which there's a weird gap there.

s170
00:11:17.000 --> 00:11:18.800
What's that?

s171
00:11:18.800 --> 00:11:19.720
It's the G brain tier.

s172
00:11:19.720 --> 00:11:21.280
It's a markdown file.

s173
00:11:21.280 --> 00:11:27.160
Do you know how many companies are at this event where their whole product could just be a markdown file?

s174
00:11:27.160 --> 00:11:28.412
It's insane.

s175
00:11:28.412 --> 00:11:28.839
[applause]

s176
00:11:28.839 --> 00:11:41.480
And like, okay, seriously though, the fact that you can now execute markdown by just piping it to Codex or Claude is unbelievable and I think most of us haven't fully appreciated how insane that is.

s177
00:11:41.480 --> 00:11:47.640
I had a service that would triage all of my PRs, have them all get reviewed with AI and then help me prioritize.

s178
00:11:47.640 --> 00:11:49.280
That service is a markdown file now.

s179
00:11:49.280 --> 00:11:57.200
I just literally wrote like, go to these four GitHub repos, look at all the open PRs, figure out what the current status of the work is, and then help me prioritize it.

s180
00:11:57.200 --> 00:12:02.760
And then when you're done, go update the static HTML file and send it to S3 and give me the URL.

s181
00:12:02.760 --> 00:12:09.840
And every morning at 9:00 a.m. this runs on a cron and around 9:15 to 9:20, my markdown file generates me my work for the day.

s182
00:12:09.840 --> 00:12:11.160
What the hell?

s183
00:12:11.160 --> 00:12:13.160
How are we actually here now?

s184
00:12:13.160 --> 00:12:15.560
I Try that if you have it, by the way.

s185
00:12:15.560 --> 00:12:21.960
You'd be amazed how many of these types of things can exist that are literally just a markdown file running on a cron.

s186
00:12:21.960 --> 00:12:26.320
But what about Okay, two more things I want to change about this, though.

s187
00:12:26.320 --> 00:12:29.200
First is the full stack cloud.

s188
00:12:29.200 --> 00:12:30.000
This is mine.

s189
00:12:30.000 --> 00:12:30.480
Don't do it.

s190
00:12:30.480 --> 00:12:31.440
Like bit coming soon.

s191
00:12:31.440 --> 00:12:32.480
Very excited.

s192
00:12:32.480 --> 00:12:33.880
I wouldn't want to compete with me on this one.

s193
00:12:33.880 --> 00:12:34.320
Trust me.

s194
00:12:34.320 --> 00:12:36.400
It's going to be really cool.

s195
00:12:36.400 --> 00:12:38.480
But there's still something else.

s196
00:12:38.480 --> 00:12:40.560
There's a gap here.

s197
00:12:40.560 --> 00:12:41.440
And I'm going to be real with you guys.

s198
00:12:41.440 --> 00:12:43.960
I don't know what goes in this gap.

s199
00:12:43.960 --> 00:12:46.120
I don't know what too big means anymore.

s200
00:12:46.120 --> 00:12:48.440
Is it training your own model from scratch?

s201
00:12:48.440 --> 00:12:50.360
Is it building your own operating system?

s202
00:12:50.360 --> 00:12:53.920
Is it trying to compete with NPM and Node directly?

s203
00:12:53.920 --> 00:12:54.800
I don't know.

s204
00:12:54.800 --> 00:12:56.760
I don't know what too big is right now.

s205
00:12:56.760 --> 00:12:59.120
And that's scary, but it's also exciting.

s206
00:12:59.120 --> 00:13:06.360
It means I need to keep pushing myself to go bigger than makes sense in order to find these limits.

s207
00:13:06.360 --> 00:13:07.680
But what does that even mean?

s208
00:13:07.680 --> 00:13:12.080
What does it mean to think bigger in this scenario?

s209
00:13:12.920 --> 00:13:17.480
I would argue that bigger's probably the wrong word for most of how I'm thinking here.

s210
00:13:17.480 --> 00:13:19.920
It's time to think wider.

s211
00:13:20.200 --> 00:13:23.760
What I mean by this is spectrum and I'm sorry I have to do a diagram.

s212
00:13:23.760 --> 00:13:26.680
If you watch my videos, you understand.

s213
00:13:26.839 --> 00:13:29.960
There's breadth and depth to any piece of software.

s214
00:13:29.960 --> 00:13:33.760
The breadth is the range of things that your software covers.

s215
00:13:33.760 --> 00:13:36.920
And the depth is the number of features in a given area.

s216
00:13:36.920 --> 00:13:39.000
Let's look at a company like Vercel.

s217
00:13:39.000 --> 00:13:41.680
Vercel does not offer all of the features that AWS does.

s218
00:13:41.680 --> 00:13:42.280
They never will.

s219
00:13:42.280 --> 00:13:44.080
It doesn't make sense.

s220
00:13:44.080 --> 00:13:52.800
But Vercel offers deeper But Vercel offers deeper features in the space they're in, which is full-stack front-end leaning servers.

s221
00:13:52.800 --> 00:13:58.560
If you're a front-end developer and you're not using Vercel, you're feeling some amount of pain because they're just further ahead with this.

s222
00:13:58.560 --> 00:14:00.920
So much so that even the agents prefer it.

s223
00:14:00.920 --> 00:14:03.120
And this was kind of how you had to build your startups.

s224
00:14:03.120 --> 00:14:08.520
Because if you were competing with a company like AWS, you're never going to have all of the features they have.

s225
00:14:08.520 --> 00:14:11.520
You're never going to cover the range that they cover.

s226
00:14:11.520 --> 00:14:18.480
It'd make no sense to try because you don't have the thousands of engineers they do doing that.

s227
00:14:18.640 --> 00:14:20.880
At least you didn't.

s228
00:14:21.120 --> 00:14:22.960
But now things have changed.

s229
00:14:22.960 --> 00:14:27.920
All of a sudden, that range is viable in a way that it never was before.

s230
00:14:27.920 --> 00:14:31.160
I'm not saying you can build something as reliable as RDS.

s231
00:14:31.160 --> 00:14:37.960
I'm saying that you can build a database platform into your product in a day or two of work with enough prompting and enough effort.

s232
00:14:37.960 --> 00:14:48.000
And if you build your stuff right, if you play your cards correctly, and you think about things the right way, you'll realize that you can build enough across a spectrum of things you care about

s233
00:14:48.000 --> 00:14:51.720
to enable most users to at least start trying the thing.

s234
00:14:51.720 --> 00:14:56.880
And when they have features they need in a given vertical that you don't support, it's not your problem.

s235
00:14:56.880 --> 00:14:58.320
As long as you build it right.

s236
00:14:58.320 --> 00:15:01.280
Because they can build the features that are missing themselves.

s237
00:15:01.280 --> 00:15:08.240
If you architect your systems and you architect your products in such a way that users can do things that they you never would have guessed.

s238
00:15:08.240 --> 00:15:13.520
Like Slack accidentally did this because Slack is now the platform people run their agents in half the time, which is crazy.

s239
00:15:13.520 --> 00:15:14.280
Slack sucks.

s240
00:15:14.280 --> 00:15:24.000
It's not a good product, but it's the right shape for people to build the features they want into it through the somewhat functional Slack bot APIs.

s241
00:15:24.440 --> 00:15:28.160
This is all crazy cuz I'm basically sitting here telling you like it's time to compete with Slack.

s242
00:15:28.160 --> 00:15:29.440
It's time to build your own AWS.

s243
00:15:29.440 --> 00:15:32.360
It's time to challenge Salesforce directly.

s244
00:15:32.360 --> 00:15:35.200
It sounds stupid, but I'm going to be real.

s245
00:15:35.200 --> 00:15:40.320
If your idea doesn't feel stupid, it's cuz your idea's not big enough.

s246
00:15:40.440 --> 00:15:41.640
That's all I have to say.

s247
00:15:41.640 --> 00:15:44.480
Thank you so much, AIE.