WEBVTT

00:00:00.000 --> 00:00:12.240
The last time you're on this podcast, you had this hot take that people were sleeping on Claude code. You were so unbelievably right. The premise of this episode is we're gonna go through what else you predict will happen. The AI jobpocalypse

00:00:12.240 --> 00:00:15.280
is not really a thing. I am

00:00:14.485 --> 00:00:15.205
super,

00:00:15.445 --> 00:01:31.560
super bullish on PMs and full stack designers. You guys are hiring double than people in the past year, which is not what people would have expected from a company that is so AI forward. I'm simultaneously extremely AI pilled and very bullish on humans. Automation is a lie. Every agent needs a human. We have so much automation, so much AI, and I also work way more. Creativity. It just feels like it's gonna be more and more valuable to stand out from all the slop that people are shipping and launching constantly. What models do in general is they make yesterday's human competence cheap, and so it becomes commoditized. It's not valuable anymore. What humans do is we go in there and we're like, yeah. We we have all this frozen human competence from yesterday. How do I use this to, like, make something new and interesting? What are some predictions for how the way we work is gonna change? It's going to bifurcate in two main ways. One is everyone's gonna have at least one agent that they talk to that they can offload work to. Second is that most of the work that you do is actually going to happen on your computer in an environment like Codex or Cloud Cowork. What you're predicting here is the SaaS tools will run within Codex or Cloud Code. I think the SaaS apocalypse is dumb. I would buy SaaS docs right now. What agents do is increase the number of users of SaaS, not get rid of it. A lot of people are moving to CLI and trying to work from the terminal. We speed ran the CLI era. It was nice while it lasted, but I think CLIs are over.

00:01:33.480 --> 00:02:44.210
Today, my guest is Dan Shipper, CEO and founder of Every. Dan and his team are building maybe the most AI forward startup out there. And as a result, are very much living in the future of how work is going to look as AI becomes a bigger and bigger part of our day to day. Everybody at their company, including every non technical person, uses Codecs and CoWork and Cloud Code to get much of their work done. And this is why way before anybody else, Dan saw the rise of Cloud Code and what is now CoWork, which he predicted almost a year ago when he was on the podcast last time. So I asked Dan to come back on the podcast to share his current biggest predictions for how work is going to change over the coming year for most people. We chat about what work will look like at most companies at the end of this year, how the shape of the work we do will change, and who will do best in this coming future slash what you need to be working on right now. Hint hint, product managers and designers are going to do very well. Dan makes a lot of bold predictions and many quite contrarian takes that I was not expecting him to say, and we are going to revisit this conversation exactly a year from today to see how much he got right. Before we get into it, do not forget to check out lennysproductpass.com

00:02:44.210 --> 00:02:51.010
for a free year of the hottest and most well crafted AI products in the world available exclusively to Lenny's newsletter subscribers.

00:02:51.090 --> 00:02:53.410
With that, I bring you Dan Shipper.

00:02:56.770 --> 00:03:00.905
Dan, thank you so much for being here, and welcome back to the podcast.

00:03:00.985 --> 00:03:10.025
Thanks for having me. Always a pleasure to be with you. The last time you were on this podcast, you had this kind of it was almost like an offhand hot take that people were sleeping

00:03:10.105 --> 00:03:22.110
on Claude code, and in particular Claude code for non engineering work, for just, like, uh, fixing files, sorting your hard drive, just all these things that people hadn't thought about. Nobody was talking about this. This was a year ago.

00:03:22.430 --> 00:03:34.255
Uh, you were so unbelievably right about this. It's just, like, unreal what has happened since then. They built Cowork, which was this whole they built on this very specific idea using Cloud Code for nontechnical

00:03:34.255 --> 00:03:41.370
work. Codex is getting into this now. I imagine you've been seeing this. They're, like, leaning into this nontechnical use of basically coding agents.

00:03:41.690 --> 00:03:46.650
I feel like this has also been a big part of Anthropic success over the past year, just like how do nontechnical

00:03:46.650 --> 00:03:47.850
people use this stuff.

00:03:48.410 --> 00:03:48.970
So

00:03:49.290 --> 00:04:02.495
you were just so ahead on this stuff. I I I even wrote a newsletter post building on this idea. I'm like, hey. This is interesting. I should dig into this. I asked people how to use Cloud Code for non engineering work, and I just had, like, so many examples, and it's, like, my second most popular post. So

00:04:03.535 --> 00:04:10.970
clearly, you you you have a unique glimpse into where things are heading. So the premise of this episode is we're gonna go through

00:04:11.610 --> 00:04:17.130
what else you predict will happen in the future, how things will change for people building products.

00:04:17.690 --> 00:04:23.355
And I think it would be helpful to start with giving people a brief glimpse into just how you operate and how your team operates.

00:04:23.675 --> 00:04:31.755
That gives you this unique lens into where things are going. So just give us a sense of how you how you work. Thank you. I I really appreciate the introduction.

00:04:33.270 --> 00:04:43.910
And, yeah, I think one one of the things about predicting the future or or the way that we think about predicting the future at every is that you what you don't wanna do is prognosticate.

00:04:44.390 --> 00:04:46.150
What do you what you wanna do instead

00:04:46.425 --> 00:04:47.145
is

00:04:48.185 --> 00:04:50.025
is just live in it together.

00:04:50.185 --> 00:04:58.985
So everybody at Every is an AI early adopter. We're almost 30 people now. I think when when we did our interview, we were 15. So we've doubled in size in the in the last year.

00:04:59.890 --> 00:05:06.050
We're all early adopters, and we have engineers. We have designers. We have writers. We have editors. We have

00:05:07.090 --> 00:05:09.250
salespeople. We have customer service people.

00:05:10.050 --> 00:05:10.690
And

00:05:10.850 --> 00:05:11.810
everybody

00:05:11.810 --> 00:05:13.010
has a little bit of that.

00:05:13.935 --> 00:05:20.015
Whatever that thing is where you're just like, I like to explore. I like to experiment. I'm very curious, and I'm, like, super all in on AI.

00:05:20.255 --> 00:05:37.620
And what I what that does, I think, is it creates this, like, little pocket of the future where we're all living in it together, and we get to be a little bit further ahead because at any other company, there's, like, a mix of people. There's really adopters. There's, like there's sort of, like, the middle of the pack people, and there are people who are the, like, very anti.

00:05:38.020 --> 00:05:42.100
And another thing that happens, which is really cool, is we get to because of our role,

00:05:43.405 --> 00:05:45.565
you know, reviewing models and and

00:05:45.965 --> 00:05:50.925
being a little bit of a pacemaker in AI, we get access to stuff before it comes out. So

00:05:51.405 --> 00:05:57.165
we get to beta test and alpha test and kind of help help steer the direction of where things are going a little bit, which is very, very cool.

00:05:57.830 --> 00:06:00.710
And so when when I think about predicting the future,

00:06:01.350 --> 00:06:07.030
it's actually when you create an environment like that, it's actually just about noticing what's going on.

00:06:08.150 --> 00:06:12.785
And and I think what a core part of it too is writing about it. I think articulating

00:06:12.785 --> 00:06:17.505
what you're noticing, articulating the future kind of brings it about in this way that

00:06:18.945 --> 00:06:23.825
makes it real for you and your team and then anybody else who's, like, on the Internet who's reading it.

00:06:24.305 --> 00:06:26.385
And so the Claude code thing,

00:06:26.705 --> 00:06:29.930
it was this it's this very organic thing where,

00:06:30.410 --> 00:06:31.370
for us,

00:06:33.130 --> 00:06:41.610
we tried Cloud Code when it came out. That's sort of our job. We we try all the new stuff from all the new all the new model all we try all the new stuff from the model companies.

00:06:42.330 --> 00:07:06.520
And at the time, it was, like, a little bit early. But right around, I think, like, Sonnet three five or Sonnet three seven, we were testing that to do our vibe check on it, and we were like, holy shit. This is crazy. This is, like, really you can they got rid of the code editor. And so from that point on, we just basically, we we run at this point now, we run, like, six products software products internally. At at that time, we ran, like, maybe two or three.

00:07:07.000 --> 00:07:11.720
And from that point on, we just started shifting to a a world where everybody was

00:07:12.200 --> 00:07:14.655
no one was looking at the code. Everybody was,

00:07:15.055 --> 00:07:19.135
you know, talking to their computer in English using Cloud Code in the terminal.

00:07:19.455 --> 00:07:22.735
And so I was able to see, like, oh, this is starting to happen.

00:07:24.335 --> 00:07:26.415
And then

00:07:25.640 --> 00:07:36.040
because my job is a little bit to just, like, push and play with stuff, I was like, I wonder if I could use this for, like, my writing. Like, how could I do that? And then it just, like, starts to unfold, you're like, okay. This is

00:07:36.680 --> 00:07:37.640
not ready yet,

00:07:38.245 --> 00:07:49.525
but it's obviously useful for me. You know? My like, one of the things that we talk about internally is what I call the reach test, which is like, do you just, when you wake up in the morning, do like, reach for it organically? I love this combination of

00:07:50.085 --> 00:07:56.310
you are using the latest stuff, and I think this is, as you said, maybe an underrate underrated skills. You're you're good at

00:07:56.790 --> 00:08:12.075
being self aware of here's what's weird and new and different and interesting. So that's a really cool combination, partly because you have to write about it and you write about it. So I think that's, like, the perfect recipe for someone having a sense of where things are going. This episode is brought to you by our season's presenting sponsor WorkOS.

00:08:12.235 --> 00:08:14.315
What do OpenAI, Anthropic,

00:08:14.315 --> 00:08:22.320
Cursor, Vercel, Replit, Sierra, Clay, and hundreds of other winning companies all have in common? They are all powered by WorkOS.

00:08:22.480 --> 00:08:39.325
If you're building a product for the enterprise, you've felt the pain of integrating single sign on, SCIM, RBAC, audit logs, and other features required by large companies. WorkOS turns those deal blockers into drop in APIs with a modern developer platform built specifically for b to b SaaS.

00:08:39.565 --> 00:08:45.325
Literally every startup that I'm an investor in that starts to expand upmarket ends up working with WorkOS,

00:08:45.620 --> 00:09:00.515
And that's because they are the best. Whether you are a seed stage startup trying to land your first enterprise customer or a unicorn expanding globally, WorkOS is the fastest path to becoming enterprise ready and unblocking growth. It's essentially Stripe for enterprise features.

00:09:00.595 --> 00:09:01.955
Visit workos.com

00:09:01.955 --> 00:09:13.430
to get started or just hit up their Slack where they have actual engineers waiting to answer your questions. Work o s allows you to build faster with delightful APIs, comprehensive docs, and a smooth developer experience.

00:09:13.670 --> 00:09:15.030
Go to workos.com

00:09:15.030 --> 00:09:17.350
to make your app enterprise ready today.

00:09:17.590 --> 00:09:22.150
So the way that I'm gonna structure this conversation, there's gonna be basically three buckets of predictions.

00:09:22.805 --> 00:09:45.630
One is how the way we work is gonna change in the coming years. Two is how what the shape of the work we're gonna be doing is gonna look like and change. And then three is who is gonna be most successful in this future slash what should you be doing and working on now to be successful in this future? Lenny, my only ask is we come on a year from now and then you score it. I wanna score.

00:09:45.790 --> 00:09:50.430
Okay. So this is a year from now. Okay. Okay. So is this that's actually

00:09:51.165 --> 00:10:10.540
is this, like, your predictions for in a year, this is what it's gonna look like, or this is, like, the emerging future? I think like, I don't I will probably say I don't have, like, an exact timeline. I think most of the stuff that I'm I'm gonna talk about will be pretty apparent within a year, but it it probably it may it may take longer than that. Okay. But I think it will it should within at least a year be, like,

00:10:11.020 --> 00:10:17.300
not obviously wrong. Like, it it seem it could it should seem like it's moving in that direction to count. Okay. May

00:10:17.300 --> 00:10:18.220
2027,

00:10:18.725 --> 00:10:20.805
we will review your predictions.

00:10:22.085 --> 00:10:38.020
Amazing. Okay. I love this. Okay. So let's dive in. What are some predictions for how the way we work is gonna change in the coming year? One of my favorite questions. Because I think if you look at the benchmarks, you're just looking at, okay. Like, yeah, AI is gonna just take all of our jobs, basically. You know?

00:10:39.140 --> 00:10:47.935
Meter has this really cool benchmark where it's like, it measures how long it can like, the newest models can do tasks autonomously, and it's like, oh, it's like, it can

00:10:49.615 --> 00:11:06.070
what's it called? Oh, myth like, mythos preview, the, like, big anthropic model that everyone's, like, so worried about. It can do tasks of seventeen hours at 50% accuracy. It's like, holy shit. That's crazy. And I think it is real. It's true. And and and the the progress, like, model progress is

00:11:06.630 --> 00:11:08.150
going up exponentially.

00:11:09.190 --> 00:11:09.750
And

00:11:10.230 --> 00:11:14.070
my experience and my feeling is that we will look back in a year and

00:11:14.390 --> 00:11:14.790
say,

00:11:15.375 --> 00:11:18.895
we actually have a lot more work to do. Humans have a lot more work to do

00:11:19.695 --> 00:11:24.095
even as models get better at doing work. And there's, like, a really interesting paradox there.

00:11:24.575 --> 00:11:25.935
And my

00:11:26.255 --> 00:11:27.215
prediction

00:11:27.215 --> 00:11:28.015
for the,

00:11:28.780 --> 00:11:39.660
like, how how work will or my my big prediction of how work will change or or how you will be doing work in a year is it's going to bifurcate in this in two main ways, how you how you use agents.

00:11:41.225 --> 00:11:45.385
One is you're going to be doing I think, like, what we

00:11:45.625 --> 00:11:49.385
figured you would be doing, like, five years ago when we thought about how

00:11:49.625 --> 00:11:51.625
work with AI works,

00:11:51.945 --> 00:11:56.250
which is everyone's gonna have at least in their company, at least one agent

00:11:56.970 --> 00:11:59.130
that they talk to that

00:11:59.610 --> 00:12:05.290
can do work, that they can offload work to. And we'll talk about, like, what that looks like, but it's essentially like OpenClaw.

00:12:05.610 --> 00:12:06.170
Second

00:12:06.565 --> 00:12:09.205
is that most of the work that you do

00:12:10.165 --> 00:12:12.565
is actually going to happen on your computer

00:12:13.685 --> 00:12:17.525
in an environment like Codex or Cloud Cowork

00:12:17.605 --> 00:12:20.885
that becomes the sort of operating system for

00:12:21.960 --> 00:12:30.600
it becomes a sort of operating system for how you do all of your work, whether that's your email, the documents you create, like, all that kind of stuff. It's gonna be on

00:12:31.000 --> 00:12:35.160
that kind of a surface. It's that's becoming the the clear competitive landscape.

00:12:36.435 --> 00:12:44.275
So there's I wanna go in order of those two. So the first one is you're gonna have agents you delegate to probably in Slack, but, you know, anywhere.

00:12:44.675 --> 00:12:46.915
First thing that's interesting about that one

00:12:47.650 --> 00:12:50.290
is it's not clear what the architecture

00:12:50.370 --> 00:12:53.890
is gonna be like for that. Is everyone gonna have an agent?

00:12:54.450 --> 00:13:04.125
Is every team gonna have an agent? Is it gonna be, like, just one agent? Is it, like, do agent specialize? Is there this, like, parallel shadow org chart?

00:13:05.085 --> 00:13:05.725
And

00:13:06.525 --> 00:13:11.965
when OpenCloud first came out, everyone internally at every adopted it, and I was very convinced that

00:13:12.365 --> 00:13:14.365
it would be a

00:13:14.605 --> 00:13:18.320
everyone has their own agent. And there's, like, some real really interesting

00:13:18.720 --> 00:13:20.480
things about that world

00:13:21.280 --> 00:13:23.280
of, you know, a parallel

00:13:23.440 --> 00:13:33.325
a parallel org chart agents in that world sort of become little reflections of you, which is, like, really cool and really interesting. It's like if you ever did you ever read the Golden Compass?

00:13:33.565 --> 00:13:41.085
It's like having a little demon on your shoulder. You know? That's a little part of your soul. Yeah. I I really think like, that's sort of what it looked like was happening.

00:13:42.605 --> 00:13:47.520
And so I was very into personal agents, and I have completely flipped.

00:13:47.920 --> 00:13:50.320
And I I really think that

00:13:50.800 --> 00:13:56.560
the the model for now is going to be a super agent, like, one agent for the entire company.

00:13:56.960 --> 00:13:59.600
And I you're you're starting to see this in some companies. So, like,

00:14:00.365 --> 00:14:11.645
Shopify very famously has one. Ramp has one now. And and I think there's some, like, really interesting reasons for that. I actually still think that the personal agent thing is coming,

00:14:12.445 --> 00:14:13.085
but

00:14:13.325 --> 00:14:30.105
what we found is there's all this hype with OpenClaw. Everyone's like, I'm gonna set it up. I'm it's so cool or whatever. And then everyone realizes it's, like, way too much work. This thing breaks all the time. I gotta, like, fumble around with it. I gotta be able to SSH into my server and, like, blah blah blah. And most people,

00:14:31.625 --> 00:14:35.785
to do work at least, just don't wanna spend that time or can't.

00:14:38.185 --> 00:14:38.905
And

00:14:39.145 --> 00:14:40.905
the the, like, fundamental

00:14:40.905 --> 00:14:43.460
underlying thing that drives that is

00:14:43.620 --> 00:14:45.940
whether it's OpenClaw or any other harness,

00:14:46.180 --> 00:14:47.460
in order for

00:14:47.860 --> 00:14:55.380
an AI agent to be useful right now, it really needs a human who cares about it. It really needs, like, a human personal connection with someone who's, like,

00:14:56.615 --> 00:15:10.535
watching what it does and make sure that it's doing the right thing and that it's useful for people. And the minute you, like, sever that connection so the minute someone's like, ugh. Like, I don't I don't wanna, like, maintain this, like, dumb open claw is the minute the agent is, like, not really that useful anymore.

00:15:10.775 --> 00:15:11.255
And

00:15:11.970 --> 00:15:15.570
that's why it I think it has started to shift to a more

00:15:17.010 --> 00:15:22.130
one agent per company model because for now, like, the the the ideal is

00:15:23.730 --> 00:15:31.195
you you basically set up a forward deployed engineer or someone with that sort of profile who's responsible for making sure that that agent is working for the whole company.

00:15:31.515 --> 00:15:34.315
And then maybe you have some, like, some little team agents.

00:15:34.715 --> 00:15:44.580
And I think as the models get better at being more independent, that will, like, shift down, and you'll it'll be more likely that we'll have more personal agents because we don't have to fuck around with all the internals.

00:15:44.740 --> 00:15:45.380
But

00:15:45.940 --> 00:15:54.095
the model that I see working for us and for a lot of other companies, including the model companies, the model companies themselves are starting to see this, is

00:15:54.895 --> 00:15:58.575
when it comes to the sort of, like, async agents, it's really a

00:15:58.815 --> 00:16:04.415
you know, you have one agent at the top that's, like, doing sometimes it's everything. A lot of times it's

00:16:04.735 --> 00:16:10.410
a a particular kind of job that you've decided that everyone in the company needs an agent for, like, data requests.

00:16:11.370 --> 00:16:12.010
And

00:16:12.250 --> 00:16:20.170
and then I think it will start to it starts top at at the top, and then it sort of starts to trickle down where you may get more specialized agents and teams and and all that kind of stuff.

00:16:20.890 --> 00:16:53.165
And the mechanism is agents need people who care about them. That is so interesting, that point about you need to, like, garden your agent because there's context you have to keep adding to it. There's, like it breaks, as you said, and it's just, like once it's just too much work, you're like, okay. Forget this thing. I'm gonna go back to Codex or Cloud or something like that. Exactly. Okay. Cool. So this is a cool opportunity. So the idea so what you're predicting here is companies will have this super agent that everyone can talk to. As you said, Shopify has got River, I think it's called. What's the ramp one called? I can't remember. Okay. It's probably gotta fund it. Okay. So

00:16:53.965 --> 00:16:55.645
so that's the prediction. Okay.

00:16:56.125 --> 00:16:58.365
Use it. That's the first prediction. That's the first We

00:16:59.965 --> 00:17:03.325
will start with agents at the top that

00:17:04.230 --> 00:17:12.870
that are more general and are used by more people in the company, and then it will start to kind of grow down as the as people get more used to these use cases,

00:17:13.110 --> 00:17:14.390
they get more specialized,

00:17:14.470 --> 00:17:16.390
and agents become

00:17:18.470 --> 00:17:19.670
less

00:17:19.935 --> 00:17:21.135
less fiddly.

00:17:21.135 --> 00:17:30.335
Like, they just work better. And is this mostly gonna be in Slack? Do you predict? For work? Yeah. It seems to make sense. I think people people love having the green bubbles on OpenClaw.

00:17:30.415 --> 00:17:40.470
Like sorry. The the blue bubbles on OpenClaw, like, if you can use it with your iPhone. But I think there's this little thing in people's heads where they really like to keep their personal and work agents separate.

00:17:40.710 --> 00:17:41.910
Mhmm. And

00:17:42.710 --> 00:17:45.110
I think there's a whole there's a whole territory.

00:17:45.110 --> 00:18:00.385
Our our COO, Brandon Gell, calls this computer errands. There's, like, a this whole territory of using personal agents for your computer errands. It's like, order my groceries or whatever, and it's like, there's so much of that that I think this is gonna it's gonna be huge for, but

00:18:00.865 --> 00:18:03.345
I focus we focus mostly on the work stuff,

00:18:04.730 --> 00:18:13.050
and I think that's gonna happen mostly in Slack. Sweet. Go Slack. Should we do you wanna talk about the, uh, the other work surface? Absolutely.

00:18:13.050 --> 00:18:19.130
Codex, co work? Okay. This is the Let's do it. I'm so excited about this one. I think it's the coolest thing. So, basically,

00:18:19.905 --> 00:18:21.345
what happened was

00:18:21.745 --> 00:18:22.705
Anthropic

00:18:23.105 --> 00:18:23.985
realized

00:18:24.305 --> 00:18:25.825
at some point that

00:18:26.145 --> 00:18:29.105
with Cloud Code, if you put an agent on your computer

00:18:29.425 --> 00:18:31.185
and it runs on your computer,

00:18:31.505 --> 00:18:43.960
it has everything it has access to everything that you have access to. It uses the terminal, so it has, like, basically super powered access to it. And not only that, it really these agents really understand how to use the terminal because there's so much

00:18:44.360 --> 00:18:46.280
content online about about that.

00:18:47.335 --> 00:18:51.815
And it it created this, like, super powerful coding paradigm, which is,

00:18:52.135 --> 00:18:54.855
you know, Anthropic was really doing it first.

00:18:54.855 --> 00:19:02.960
OpenAI for a while was, I I I, in my opinion, like, very, very behind on this, and then in my opinion has surpassed them recently. It's really interesting.

00:19:03.680 --> 00:19:04.480
But they were

00:19:05.120 --> 00:19:06.560
very early on this.

00:19:07.120 --> 00:19:12.720
When people were still thinking about coding agents or coding models as being really pair programmers,

00:19:12.720 --> 00:19:14.880
they were among the first to be like,

00:19:15.200 --> 00:19:19.335
no, and do it successfully. Like, there are people before them like Devin who,

00:19:19.735 --> 00:19:24.695
I think, had had a big had the big, like, cloud environment and and OpenAI tried this too, but

00:19:25.175 --> 00:19:29.460
the the the real adoption seems to have happened when you

00:19:29.780 --> 00:19:31.060
put it on your computer.

00:19:31.620 --> 00:19:36.260
So they figured that out. And then I think they figured out, along with their community,

00:19:36.900 --> 00:19:44.675
that once you have a coding agent on your computer that can build anything, it's actually really good for any kind of work you wanna do. And people started

00:19:44.755 --> 00:19:50.195
just hacking Cloud Code essentially to do all their work. So Anthropic then built Cowork,

00:19:50.595 --> 00:19:55.715
which is, you know, a little bit of a nicer wrapping around Cloud Code, but it's fundamentally the same thing.

00:19:56.595 --> 00:19:57.475
And then

00:19:57.795 --> 00:19:58.275
I think,

00:19:59.450 --> 00:20:03.290
you know, I think OpenAI made a couple of different bets, but

00:20:03.850 --> 00:20:14.410
their main bet on a programming agent was the the the the earlier version of the codex were, like, very technical, and they were, like, super smart, but they were, like, a little bit autistic. Like, it was a little hard to

00:20:14.855 --> 00:20:19.015
they didn't quite get what you meant. They get they got exactly what you said.

00:20:19.495 --> 00:20:25.895
And I think maybe, like, three or four months ago around the time that they launched, uh, 5.3,

00:20:26.215 --> 00:20:26.535
they

00:20:27.240 --> 00:20:35.480
started to move in this direction of, oh, no. We get it. Like, it's, um, this model is fast. It's, like, really good for general purpose knowledge work type tasks.

00:20:35.720 --> 00:20:40.360
And then they launched the Codex desktop app. And I think the Codex desktop app takes

00:20:40.605 --> 00:20:46.765
if you look at all the lessons that, like, Anthropic learned, they went from Claude code to Cowork,

00:20:47.005 --> 00:20:51.725
and you can kinda see that in the tabs on the on the Anthropic desktop app UI.

00:20:52.125 --> 00:20:55.725
I think opening out was just like, we we see where this is going. Like, let's just skip to that.

00:20:56.370 --> 00:21:01.890
And so I think Codex right now this is a horse race. Like, they're gonna have different positions.

00:21:03.170 --> 00:21:16.965
But I I think Codex right now, it's my daily driver. I, like, spend all all my time in it, basically. I flip the card every once in a while, but I think they're getting the paradigm right. And it's clear to me that whoever is in the lead, because I I again, I think it'll change.

00:21:17.285 --> 00:21:21.205
Whoever's in the lead, it feels very obvious to me that

00:21:21.690 --> 00:21:26.410
all of the work that you do is going to be in one of those surfaces where,

00:21:26.570 --> 00:21:28.650
uh, for example, when I'm writing a document,

00:21:28.970 --> 00:21:30.490
Codex has a browser

00:21:30.890 --> 00:21:33.690
in, uh, in the app. It has an in app browser.

00:21:34.215 --> 00:21:40.375
And when I'm writing a document, I just go into one of my one of my codex threads, which I have one thread for every project.

00:21:40.775 --> 00:21:48.055
And I just open the in app browser. I go to the document. I usually do it in proof, which is this online mark markdown editor I built.

00:21:48.935 --> 00:22:02.510
And then I just have Codex running and watching me in proof, and Codex can see what I'm doing. I can see what Codex is doing. It's all kind of in one place, which is the an extension of the same thing that made Cloud Code work really well originally.

00:22:03.195 --> 00:22:06.395
And I basically feel like I have this parallel

00:22:06.955 --> 00:22:16.875
work buddy that not only can it, like, respond and write in the document, but then it can go do research. It can go it can use my computer to basically do anything that I can do on my computer, and that's, like,

00:22:17.780 --> 00:22:19.300
incredibly powerful.

00:22:20.420 --> 00:22:26.340
Um, and I do this with everything. Like, I've been in I've been at Inbox Zero for, like, ten days straight now, which,

00:22:26.740 --> 00:22:32.660
if you know me, is crazy. I'm never like this. And that's because I literally just have

00:22:33.735 --> 00:22:34.615
codex,

00:22:35.415 --> 00:22:38.135
gather all my emails with Quora, which is our email agent,

00:22:38.375 --> 00:22:39.095
and then

00:22:39.415 --> 00:22:47.415
it it renders a little page. And I I think I showed you this at the in at the Anthropic event. It renders a little page, and I just, like, monologue

00:22:47.415 --> 00:22:51.420
into it and just talk at each email. I'm like, okay. Go go research this.

00:22:51.740 --> 00:23:06.935
Oh, here's a question from our lawyers. Can you go, like, collect all of the, you know, documents from the last, like, four years and then put them into a report and send them? And it just does it. And so all the stuff that I would procrastinate on, I don't really procrastinate procrastinate on on anymore. And so I feel like there's this

00:23:07.895 --> 00:23:14.935
for a long time, we thought I thought too that the optimal experience of AI was gonna be take AI

00:23:15.095 --> 00:23:16.535
and put it in a browser.

00:23:17.530 --> 00:23:23.850
And I think the reverse is actually starting to happen and be, like, really, really valuable in a way that I did not expect, which is

00:23:24.650 --> 00:23:37.645
take the AI agent that you use all the time on your computer and put a browser in it so it can see everything you're doing. And that is just like a magical combination that I think will be is very uncommon now. You can't even do this in cloud in Cloud Code

00:23:38.205 --> 00:23:49.450
because they they don't let you browse external websites inside of Cloud Code. So it's very uncommon now, but I think it will be super common in a year. This is more profound than it may even sound. What I'm hearing is

00:23:49.930 --> 00:23:52.330
instead of AI being baked into

00:23:52.650 --> 00:23:58.330
SaaS tools, what you're predicting here is, uh, you will the the SaaS tools will run within

00:23:58.650 --> 00:24:00.170
Codecs or Cloud Code.

00:24:01.625 --> 00:24:04.985
That that is that is one, uh, really important,

00:24:05.065 --> 00:24:07.865
uh, second order effect of this is okay.

00:24:10.345 --> 00:24:10.985
So

00:24:11.785 --> 00:24:15.785
yeah. Like, I'm I'm using Proof or or really any website, maybe Posthog or whatever,

00:24:17.590 --> 00:24:24.870
and I'm doing it inside of my agent. And the agent has access to the website. So it has access to everything that I have access to, and it has access to my whole computer.

00:24:25.110 --> 00:24:28.390
When I run the agent on that website, I'm using my tokens.

00:24:28.870 --> 00:24:35.695
I'm not using the the vendor's tokens. I'm not using the app's tokens. And so it puts SaaS back in this place where,

00:24:35.935 --> 00:24:41.135
yeah, you wanna make it friendly for an agent. And everyone's got a CLI now. Um, you wanna make the HTML,

00:24:41.135 --> 00:24:50.070
uh, really, uh, really usable. You wanna make sure that what anything that happens in the CLI shows up for the user immediately, all that kind of stuff. There are a lot of issues to to deal with.

00:24:50.470 --> 00:24:56.390
But, um, once you do that, you actually don't really need to think about having a

00:24:56.710 --> 00:25:11.955
an AI surface that's primarily gonna be the thing that users use in the sense that you don't need to build an agent necessarily into your product. I think you can, and there's there's another really interesting bifurcation of this that that that we should talk about, um, which is that having two agents is better than one.

00:25:12.435 --> 00:25:13.475
Um, but

00:25:14.115 --> 00:25:17.475
I think for now, there's this really cool thing where,

00:25:18.140 --> 00:25:19.660
with Proof, for example,

00:25:20.380 --> 00:25:25.900
anyone who uses it, I don't pay for tokens because they're just bringing they bring their AI to to Proof.

00:25:26.220 --> 00:25:29.500
And so it changes what you build as a SaaS company,

00:25:29.740 --> 00:25:39.225
and you build it now for both humans and agents to use at the same time. And it changes your margins back to, well, I don't really have to pay for tokens anymore because the user is gonna bring the AI.

00:25:39.385 --> 00:25:46.345
So I think this is a huge deal. So what you're describing here is, uh, more and more work that we do, more and more professional work, is it just gonna happen within

00:25:46.745 --> 00:25:48.025
Codex or Cloud Code?

00:25:48.740 --> 00:25:55.620
Where does Cursor fit into this? Is that one of the is is there a potential there? That's a good question. I think that Cursor

00:25:55.620 --> 00:25:57.300
sees a lot of the same stuff.

00:25:58.100 --> 00:26:04.420
And there and in some ways, they have some of the same stuff, but it's better. Like, I think that Cursor's

00:26:03.465 --> 00:26:04.745
cloud implementation

00:26:04.745 --> 00:26:08.185
is better than either it or OpenAI's or Anthropix and is more advanced.

00:26:08.665 --> 00:26:09.225
And

00:26:09.785 --> 00:26:11.385
I think that Cursor

00:26:11.385 --> 00:26:11.945
has,

00:26:12.425 --> 00:26:13.785
at least so far,

00:26:14.025 --> 00:26:19.830
more distinctly chosen a lane. Like, they're more distinctly choosing to be for programmers,

00:26:20.550 --> 00:26:21.030
and

00:26:21.670 --> 00:26:33.725
that may limit how far they get in here. Like, I think the definition programmer is expanding enough that they'll have a big market, but I don't know that they're gonna jump into, like, okay. Use this to make a slide deck or whatever. But it is really clear

00:26:34.605 --> 00:26:35.165
that

00:26:35.565 --> 00:26:43.165
every model company is starting to realize how important it is to have a harness to get the most out of the the model.

00:26:43.405 --> 00:26:46.285
And so where the where all the platforms are moving

00:26:46.700 --> 00:26:48.860
is to a world where

00:26:49.020 --> 00:26:51.340
you you're not just doing prompt and response.

00:26:51.420 --> 00:26:56.140
When you call the the model at at on on the OpenAI platform, the Anthropic platform,

00:26:56.300 --> 00:27:03.495
you are they're literally, like, running the model on a computer that that is in the cloud that they run and then giving you the result out of it.

00:27:03.815 --> 00:27:08.935
And they know that they in order to get the best results of the model, they need to offer that. And so you see,

00:27:09.255 --> 00:27:10.455
you know, Anthropic's

00:27:10.455 --> 00:27:11.815
got cloud managed agents.

00:27:13.140 --> 00:27:28.260
OpenAI does not have a have a response yet, but I assume that that's gonna happen. And now Cursor was just essentially acquired by SpaceX. It's not, like, a full acquisition, but it's close. So I think people are starting to realize, like, I can't just do the, like, model part of it. I have to have this, like,

00:27:29.395 --> 00:27:47.970
harness above it. And I think the ultimate form of that harness is, like, I can do any kind of knowledge work. Cursor itself is feels like one of the things that it's gonna be a hard decision for them whether to stay just for coders or not. So people building products that aren't OpenAI or Anthropic, if this proves to be true,

00:27:48.370 --> 00:27:54.130
the prediction here is they're gonna be using your product over time inside of one of these agents.

00:27:55.075 --> 00:27:58.755
Is there something you would do if you're one of those companies to prepare for that future?

00:27:59.075 --> 00:28:02.755
I I would just prepare for that. So, like, you know, for for example,

00:28:03.875 --> 00:28:04.595
every

00:28:04.835 --> 00:28:08.995
more classic piece of productivity software, whether it's Slack or

00:28:09.850 --> 00:28:14.970
Word docs or PowerPoints or whatever. It's really mostly meant for

00:28:15.210 --> 00:28:16.650
a human to use.

00:28:18.010 --> 00:28:20.970
And now people are doing CLI, so it's, like, meant for

00:28:21.975 --> 00:28:24.615
an agent to use independently of a human.

00:28:25.095 --> 00:28:31.655
And we're moving into this new paradigm, I think, where the human and the agent are on the same piece of work

00:28:31.895 --> 00:28:41.690
together, and they're both doing things. And you need to have I I need to have visibility into what the agent is doing. The agent has to have visibility into what I'm doing. We have to go back and forth in this sort of, like, seamless way.

00:28:42.570 --> 00:28:47.210
And the kind of software that you make for that is gonna be very different. So for example,

00:28:49.685 --> 00:29:07.520
like, there's a lot of stuff that Proof doesn't have. I don't have to have a lot of the, like, Word document kinda, like, formatting or page breaks or, like, you know, making tables or whatever because the agent just does it. I don't need to worry about that. It can do all the formatting for me. So you can make the products a lot simpler and faster to start than the legacy products are.

00:29:08.000 --> 00:29:15.680
And then there's all these other affordances that you need to start to have because the way agents interact with software is very different. So, for example,

00:29:17.205 --> 00:29:24.485
agents can do a lot at once. They can just do, like, a billion different things to your document or your slide deck or your code base or whatever.

00:29:24.805 --> 00:29:25.365
And

00:29:25.765 --> 00:29:35.980
how you display that to the user is gonna be very different than the way you might display a human being concurrent in your document and doing stuff. You need you need, like, approval.

00:29:36.140 --> 00:29:42.220
You need a sort of inbox that sort of summarizes here's all the stuff that's going to happen or has happened. You need

00:29:42.700 --> 00:29:44.860
you need logs and the ability to roll it back real quick.

00:29:45.905 --> 00:29:48.705
So there's all those kinds of considerations that,

00:29:49.025 --> 00:29:49.585
um,

00:29:49.985 --> 00:29:55.505
that change the actual product, and then the underlying UX of it or the underlying infrastructure you need is different too because,

00:29:55.825 --> 00:30:01.020
you know, agents can make a billion requests in, like, three seconds. So how are you gonna deal with that? Right?

00:30:01.100 --> 00:30:10.300
Um, this is exactly why, you know, GitHub is having problems right now because because the the number of people using GitHub has is skyrocketing exponentially, and it's really just people's agencies in GitHub.

00:30:10.700 --> 00:30:18.805
So I I think it's a this whole new world that is just starting you're just starting to see, like, a little peak of it. But there's so many cool things about it. So, for example,

00:30:19.045 --> 00:30:20.005
in proof,

00:30:21.125 --> 00:30:22.805
in some of our other products too,

00:30:23.205 --> 00:30:25.045
uh, when someone has a problem,

00:30:25.525 --> 00:30:27.605
they don't email support.

00:30:27.765 --> 00:30:28.885
Their agent

00:30:29.445 --> 00:30:30.725
sends a bug report,

00:30:31.800 --> 00:30:35.880
and an agent bug report is way better than a human bug report.

00:30:36.120 --> 00:30:49.005
Um, it has, like, here's exactly what I did. Here's the exact repro steps. Here's, like proof is open source. So here's what I think is going on in the code base. And then we just get that. It becomes a GitHub issue, and then we can just, like, send off an agent to fix it. And

00:30:49.885 --> 00:30:52.925
you can't do that with everything, but it's so much better.

00:30:53.805 --> 00:30:56.125
And you can see the, like, the glimmers of

00:30:56.365 --> 00:31:00.910
this this very fast, like, closed loop between

00:31:01.230 --> 00:31:09.870
I ran into something, a paper cut, a little feature I want, a little bug, and my agent just goes off and talks to the company agent, and then the company agent just goes and

00:31:10.190 --> 00:31:33.670
fixes it. That, I think, is incredibly cool. So is a part of this that you a lot of people are moving to CLI and trying to work from the terminal. Is the part of this prediction that people shift away from that and back to actual UX with agents kind of running alongside them? CLIs are over. We we speed ran the CLI, uh, era. It was nice while it lasted, but I think it's pretty it's pretty clear.

00:31:33.990 --> 00:31:47.205
It's not that CLI sorry. It's not that CLIs are going to completely go away. Obviously, they've been around for the last, like, thirty years or forty years or fifty years or whatever. They will continue to be around. And I think there is this moment when Cloud Code was, like, so

00:31:47.525 --> 00:31:48.325
popular,

00:31:48.965 --> 00:31:52.325
uh, or or when when Cloud Code was really starting to gain in popularity

00:31:52.580 --> 00:31:56.900
that people were like, the the thing that's working is the fact that it's the CLI,

00:31:56.900 --> 00:32:02.660
and I don't think that's what it is. And when you move into an actual UI for this, you start to realize

00:32:04.100 --> 00:32:05.620
we made GUIs for a reason,

00:32:06.575 --> 00:32:07.135
and

00:32:07.455 --> 00:32:12.415
it's just nicer to be in a GUI. And you can get all the same benefits

00:32:12.415 --> 00:32:16.415
inside inside of GUI, especially for nonprogrammer work. But I would

00:32:16.975 --> 00:32:18.975
I would estimate that

00:32:19.855 --> 00:32:21.935
definitely

00:32:21.260 --> 00:32:32.220
the majority of the technical people inside of every are not using CLIs anymore as their main work surface. I think a lot of programmers are still flipping into it every once in a while, but it's more or less they're using

00:32:32.620 --> 00:32:45.395
codecs, Cloud Code, cursor, um, that kind of thing. Awesome. Okay. I I I would I definitely wanted to make that part clear. So coming back to kind of the the big picture of the prediction here, there's kind of these two modes of work that you're anticipating.

00:32:45.555 --> 00:32:56.320
One is this kind of super agent within a company that you chat with through Slack, most likely, that can go off and do work and answer questions. And then there's on your computer running Codex or ClockCode.

00:32:56.560 --> 00:33:11.595
And within that, all the work that you'd normally do kind of on your computer is now gonna be living within Codex or ClockCode or maybe some third party that emerges that we're not even aware of yet. Yes. And you're going to use apps inside of the internal browser of those,

00:33:11.835 --> 00:33:13.755
uh, of those tools.

00:33:14.555 --> 00:33:16.155
Wow. Okay. Like,

00:33:16.475 --> 00:33:32.830
listening to you talk about it, it's like, it may not feel as profound as it is, because this is a big change to how we work. We don't currently have an AI that we talk to regularly in Slack, and we also don't work currently mostly in Codex or GlarCode. So this is actually a pretty massive shift. I think so.

00:33:34.715 --> 00:33:38.315
Is there anything else along these lines before we get into our next prediction?

00:33:38.395 --> 00:33:49.290
Well, a few things. I am definitely not an agent maximalist. Like, I really think we're gonna have a lot of different agents that we use. Seems pretty clear to me. And I really do think that two agents are better than one.

00:33:50.010 --> 00:33:50.570
So

00:33:51.450 --> 00:33:52.650
that's a good example.

00:33:53.370 --> 00:33:55.290
When I have codex

00:33:55.530 --> 00:33:56.330
interact

00:33:56.330 --> 00:33:57.290
with another agent,

00:33:57.785 --> 00:33:58.905
it can

00:33:59.065 --> 00:34:04.345
give so much more context about me and what I want than I would be able to type.

00:34:05.225 --> 00:34:10.985
And it can go back and forth talking about things that would take a long time for me to express directly to an agent

00:34:11.370 --> 00:34:14.650
that you get this, like, speed up effect when

00:34:15.290 --> 00:34:23.210
you assume that your users are are using Codex or Cloud Code or Cowork as their as their basic way they access your app.

00:34:24.225 --> 00:34:24.865
And

00:34:25.425 --> 00:34:26.945
a a really simple example,

00:34:27.585 --> 00:34:28.625
we have this

00:34:29.505 --> 00:34:34.705
hosted OpenClaw product, which we we had it we had on waitlist. We actually had to pause it because

00:34:34.865 --> 00:34:41.270
we start taking to all the waitlist. OpenClaw is just a very hard agent harness to to make work. It's like,

00:34:41.590 --> 00:34:43.830
it's moving so incredibly fast.

00:34:44.630 --> 00:34:50.485
Uh, and if you're, like, a platform for it, it just it's like when things break, you can't fix it. It's very hard.

00:34:51.765 --> 00:34:54.005
But one of the things that we learned in that process

00:34:54.485 --> 00:35:00.325
is if you're let's say you're building an an agent product or or a news any new software experience,

00:35:01.045 --> 00:35:11.260
what you would assume, let's say, to set up an agent is you need to build, like, a little, like, web interface or a little Slack workflow that ask people about, okay. Like,

00:35:11.500 --> 00:35:12.860
who are you, and

00:35:13.260 --> 00:35:16.780
what are you gonna use this for? And, like, what's your what's your ideal,

00:35:17.505 --> 00:35:22.305
you know, dream outcome? Or what whatever the things you are that you would put on an onboarding checklist.

00:35:23.985 --> 00:35:25.025
If instead,

00:35:25.025 --> 00:35:30.225
you you just you just make a hard line of we are only going to service users

00:35:31.185 --> 00:35:31.585
who

00:35:32.080 --> 00:35:34.240
use Codex or Cowork.

00:35:34.800 --> 00:35:41.440
Um, what happens is you just paste something into you just paste a prompt into Codex or Cowork.

00:35:41.840 --> 00:35:46.960
It goes and talks to the app, and the app can be either just a regular server or or it can be its own agent.

00:35:47.955 --> 00:35:48.675
And

00:35:48.755 --> 00:35:53.715
Codex has so much information about you that it can just give it here's

00:35:53.715 --> 00:35:55.475
all the stuff I've been working on with Dan.

00:35:56.115 --> 00:36:03.380
Here's all the ways that, you know, he might he might wanna use this app and then bring it back to me. And it's this very custom experience.

00:36:03.860 --> 00:36:07.620
And, also, for a technical product like an agent, when something goes wrong,

00:36:08.100 --> 00:36:26.235
I can just tell Codex, go fix it. And Codex will go talk to the app and figure out what's going on for me. And so I think the whole paradigm starts to change when you assume that everyone's got an agent, and those agents are talking to other agents in this, like, really magical and important way. There's a couple of more things I wanna touch on before we get started because there's, like, so much to talk about.

00:36:26.910 --> 00:36:32.030
One is you made this point about SaaS tools not using like, you can use tokens from

00:36:32.270 --> 00:36:42.030
the, uh, model companies, basically, when using a SaaS tool. Talk a bit more about that because that may change the business model for SaaS companies in the future. That feels like a big deal. Well, I think it actually

00:36:42.955 --> 00:36:43.755
may,

00:36:44.235 --> 00:36:46.315
uh, save their margins.

00:36:46.955 --> 00:36:51.835
Because right now, all the all these companies are rushing to, like, add a agent to their offering

00:36:52.315 --> 00:36:56.395
and thinking, oh, the agent is gonna be the main way that I that people interact with me.

00:36:57.200 --> 00:36:57.760
And

00:36:58.160 --> 00:36:59.440
I think that,

00:36:59.600 --> 00:37:06.800
uh, and that cost tokens, obviously. And I actually think once I have once I have codex or co work as my main work surface,

00:37:07.280 --> 00:37:12.005
I still wanna use SaaS. So this is another good prediction. I would buy SaaS stocks right now.

00:37:12.485 --> 00:37:15.125
I would I think the SaaS apocalypse is dumb,

00:37:15.605 --> 00:37:21.365
and SaaS stocks will be up majorly in the next couple years. Not not investment advice, but,

00:37:22.005 --> 00:37:23.685
you know, I would buy SaaS stocks.

00:37:25.790 --> 00:37:26.430
So

00:37:28.270 --> 00:37:29.070
so

00:37:29.950 --> 00:37:33.150
so I think it saves your margin because now

00:37:33.310 --> 00:37:38.465
what you're what the way that you're thinking then is not have to build AI into this. It's it's more like I

00:37:39.025 --> 00:37:42.305
have to make a piece of software that humans and AI wanna collaborate on together.

00:37:44.145 --> 00:37:44.785
And

00:37:45.345 --> 00:37:49.505
that's hard, but it's once you build it, it's a lot cheaper than

00:37:49.905 --> 00:37:51.425
assuming everyone's spending tokens.

00:37:52.390 --> 00:37:53.030
And

00:37:54.630 --> 00:38:00.070
it's I think it's a I think it's a good business. And and part of the reason I'm so bullish on SaaS is,

00:38:00.230 --> 00:38:00.870
a,

00:38:01.510 --> 00:38:04.470
everybody internally here is

00:38:04.125 --> 00:38:05.245
like

00:38:05.245 --> 00:38:11.885
I said, we have all got agents, and we're all using codecs and whatever. And we still pay for a ton of SaaS, and our SaaS spend is up year over year.

00:38:12.685 --> 00:38:16.445
And we're not, like, vibe coding every single, like, little thing. You know?

00:38:18.045 --> 00:38:18.765
And

00:38:19.430 --> 00:38:23.590
I think that what agents do is increase the number of users of SaaS,

00:38:25.030 --> 00:38:26.150
not get rid of it.

00:38:26.630 --> 00:38:36.325
And so I think SaaS companies are going to see, like, an ex insane spike in the amount of demand that they have because there's gonna be tons of agents using these products at, like, a very high volume.

00:38:36.725 --> 00:38:39.285
And like I said, that's a huge infrastructure challenge.

00:38:39.605 --> 00:38:42.325
There's a there's a lot of, like, interesting pricing challenges,

00:38:42.805 --> 00:38:52.900
but, uh, it it it makes me very bullish on SaaS. I love that. If anything else comes out of this conversation, Dan Chipper, SaaS is the future of AI. This

00:38:55.060 --> 00:38:56.340
b to b SaaS.

00:38:56.580 --> 00:38:58.020
Hashtag send tweet.

00:38:58.260 --> 00:39:16.480
I I love just yeah. This is, uh, quite contrarian. And the other interesting piece is that the fact that you guys are hiring, that you doubled in people in the past year, which is not what people would have expected from a company that is so AI forward. Talk about what your experience there of just, okay, we still actually need humans. Automation is a lie,

00:39:16.960 --> 00:39:17.680
um,

00:39:17.840 --> 00:39:19.040
in the sense

00:39:19.200 --> 00:39:19.840
that

00:39:20.240 --> 00:39:27.575
every time you automate something, in order to make sure the automation is working well, you need a human on top of it, like, making sure that it's working well. And so,

00:39:28.055 --> 00:39:36.695
you know, I wrote this piece a couple years ago called the allocation about the allocation economy, like, idea that the way that humans are gonna work with AI is gonna is gonna be,

00:39:36.935 --> 00:39:38.215
like, like, being a manager.

00:39:39.410 --> 00:39:39.970
And

00:39:40.690 --> 00:39:45.170
the thing that you have to remember about managers is, like, managers actually spend a lot of time working.

00:39:45.810 --> 00:39:46.450
Most

00:39:46.610 --> 00:40:03.225
managers are not, like, on the beach. They're, like, checking in with their employees all the time and and and and trying to figure out, How okay. Do we make this work good? How do we make it better? How's it doing? How's this person doing? All that kind of stuff. And I think there's there's some differences between being a human manager and being a model manager, but, um, fundamentally,

00:40:03.225 --> 00:40:05.225
it still requires a lot of time and attention.

00:40:07.140 --> 00:40:07.780
And

00:40:08.100 --> 00:40:13.540
I think that we kind of missed that in the model discourse. And one of the reasons is

00:40:16.020 --> 00:40:19.220
benchmarks make it look like AI is more autonomous than it is.

00:40:20.885 --> 00:40:22.565
And by autonomy

00:40:22.885 --> 00:40:28.965
I mean something specific by autonomy, and I'm gonna try to express this. It's, like, a little hard to express, but I learned this for myself

00:40:29.365 --> 00:40:35.845
because I've been feeling this paradox a little bit. I've been feeling the like, we have so much automation, so much AI, and I also work way more.

00:40:36.550 --> 00:40:38.470
And I think part of the paradox

00:40:38.710 --> 00:40:42.870
part of the paradox started to, like, resolve for me a little bit when I made my own benchmark.

00:40:43.430 --> 00:40:49.830
So I made this senior it's called the the senior engineer benchmark. And it's like, how good is AI versus a human engineer?

00:40:51.295 --> 00:40:53.455
And the way that I built it

00:40:53.695 --> 00:40:54.335
is,

00:40:54.495 --> 00:41:00.735
again, have this app proof. I just vibe coded it on the side and like, while running the rest of every.

00:41:00.975 --> 00:41:12.350
And when we launched it, because it was completely vibe coded, it just started going down, and I couldn't fix it. And it was very embarrassing. I had a lot of egg on my face. And, like, the product worked. We we tested it internally.

00:41:12.750 --> 00:41:13.550
We had

00:41:14.190 --> 00:41:16.990
a lot of beta testers, but, like, the day after launch, it was, like,

00:41:17.635 --> 00:41:27.475
just every, like, ten minutes, the servers would go down, and people were looking at me, and I'd be like, I don't know what's going on. Like, Codex, fix it. And Codex is like, I don't know what's going on. Or, really, Codex was like,

00:41:27.795 --> 00:41:38.140
I do know what's what's going on. I fixed it, and then it it would cause four other errors. And then you're just going around in a circle, and I wasn't sleeping. And I I I vibe coded so hard, I got bursitis on my elbow.

00:41:39.260 --> 00:41:39.900
So

00:41:40.060 --> 00:41:43.420
that's a there's a life lesson in there. Vibe coder elbow.

00:41:43.500 --> 00:41:43.820
Yeah.

00:41:45.965 --> 00:41:51.245
So, anyway, I got a I got actually two different senior engineers to fix it independently.

00:41:51.885 --> 00:41:54.685
So I have two different rewrites

00:41:54.685 --> 00:41:56.445
of the code base that

00:41:57.325 --> 00:42:02.270
tells me how they did it. Right? And so what I get to do is when we get new models,

00:42:02.670 --> 00:42:11.550
I just give the new model a prompt. I say, like, this is Vibe coded slap. If you wanted to rewrite it from first principles, how would you write it? Go do it.

00:42:12.345 --> 00:42:12.985
And

00:42:14.585 --> 00:42:18.665
all the models until GPT 5.5 got, like, a 30 out of a 100,

00:42:19.225 --> 00:42:23.865
and senior like, human senior engineer gets, like, high eighties, low nineties out of a 100.

00:42:24.610 --> 00:42:29.010
So there's a lot to go. And then I tried GPT 5.5, and it got, like, a 62.

00:42:30.130 --> 00:42:33.010
And mind you, the 60 the 60 score was

00:42:33.490 --> 00:42:36.530
GPT 5.5 using an OPUS 4.7

00:42:36.530 --> 00:42:42.445
plan. OPUS 4% the plans are very good. G b d 5.5 is the only model though that has the

00:42:42.765 --> 00:42:58.470
sense of agency and confidence to just, like, rip out old code and just, like, actually rewrite for first principles. Other coding models, they kinda, like, try they, like, end up papering over the edges or around the edges, and they're like, oh, this is a big job. Like, I'll just do a little patch. And you're like, no. I, like, specifically told you not to.

00:42:59.270 --> 00:43:04.870
So g p t 5.5, there's, like, a 30 bump in the score, 60 out of a 100. It's, like,

00:43:05.885 --> 00:43:08.445
very it's very clear that in

00:43:09.165 --> 00:43:10.845
a year or less,

00:43:11.005 --> 00:43:18.525
it's gonna be senior engineer level. Right? And that gives you a certain picture in your mind, especially based on how I named the benchmark, which I think a lot of benchmarks do.

00:43:19.480 --> 00:43:23.000
And I can tell you that when we get to that point,

00:43:23.640 --> 00:43:28.200
I will be it will be very easy for me to change the benchmark to zero out the current model.

00:43:28.600 --> 00:43:30.200
So that gets a zero out of a 100.

00:43:30.680 --> 00:43:31.880
And so, for example,

00:43:34.205 --> 00:43:44.845
it seems like there's no skill or no thought into the prompt, which is this is why I could've slapped, like, fix it from first principles. But, actually, it took me a a while to get to a prompt that

00:43:45.190 --> 00:43:46.870
didn't give away the answer,

00:43:47.430 --> 00:43:47.990
but,

00:43:48.790 --> 00:43:49.750
uh, but

00:43:49.990 --> 00:43:51.590
got the model to reveal

00:43:52.230 --> 00:43:53.670
what it's capable of.

00:43:54.390 --> 00:44:05.525
And the original prompt I gave it was the original prompt that I gave it when, uh, when I was trying to fix the issue when production was going down, which is, like, I'd woke waking up I'd I'd woken up in the morning,

00:44:06.485 --> 00:44:07.605
and I was like,

00:44:08.165 --> 00:44:13.365
okay. We had four or five reported issues yesterday. I want you to go through all the issues and then come to, like,

00:44:14.600 --> 00:44:16.760
make a plan for how to resolve all of them

00:44:17.160 --> 00:44:37.315
and go do it. Right? And every coding model on the market and I I am I'm pretty sure this here's a prediction. I'm pretty sure every coding model on the market will still do this in a year. Every coding model on the market will take that instruction seriously. And if I tell it, here's a bunch of issues. Go fix it. They will just go try to fix the issues.

00:44:37.955 --> 00:44:53.440
What a actual human senior engineer does is they go look at the code base, they're like, this is a piece of shit. This guy doesn't know what he's doing. And then then they say, we're gonna have to, like, actually rewrite a lot of this, and it's gonna be hard and risky. I know you don't wanna hear that, but, like, we're gonna have to do that.

00:44:53.760 --> 00:45:01.885
And if you ask the model, hey. Like, should we do that? It'll it'll probably it'll probably get there, but it's not gonna do it on its own.

00:45:03.245 --> 00:45:10.445
And it and there's a lot of incentives pushing against it doing that. And even if it does that, there's there's always a higher frame for us to go.

00:45:11.140 --> 00:45:13.540
And so I think it's it's really important,

00:45:13.940 --> 00:45:17.380
um, when when we think about benchmark progress to

00:45:17.860 --> 00:45:20.740
think about it from that perspective, which is benchmarks

00:45:20.900 --> 00:45:33.525
rise on problems that we've framed that we can articulate, that we can score. And there's a lot of work that's human work that, uh, it it can't be scored until you write it down, but the act of thinking to prompt it or write it down

00:45:34.565 --> 00:45:35.285
is,

00:45:35.845 --> 00:45:36.405
uh,

00:45:36.645 --> 00:45:46.920
is something that you can't measure, but, like, kind of means that even if the benchmarks get saturated, it doesn't mean the same thing as we you totally replace all senior engineers. And it's I think it's why

00:45:47.560 --> 00:45:49.880
even though the models are getting better at automation,

00:45:50.840 --> 00:45:51.880
I still hire engineers.

00:45:52.375 --> 00:45:56.935
I am so excited to tell you about this season's supporting sponsor Vanta.

00:45:56.935 --> 00:45:59.015
Vanta helps over 15,000

00:45:59.015 --> 00:46:02.535
companies like Cursor, Ramp, Duolingo,

00:46:02.535 --> 00:46:16.770
Snowflake, and Atlassian earn and prove trust with their customers. Teams are building and shipping products faster than ever, thanks to AI. But as a result, the amount of risk being introduced into your product and your business is higher than it's ever been.

00:46:17.010 --> 00:46:22.975
Every security leader that I talk to is feeling the increasing weight of protecting their organization,

00:46:23.135 --> 00:46:25.855
their business, and not to mention their customer data.

00:46:26.095 --> 00:46:29.615
Because things are moving so fast, they are constantly reacting,

00:46:29.615 --> 00:46:33.215
having to guess at priorities, and having to make do with outdated solutions.

00:46:33.670 --> 00:46:39.510
Vanta automates compliance and risk management with over 35 secondurity and privacy frameworks,

00:46:39.670 --> 00:46:42.630
including SOC two, ISO 27,001,

00:46:42.630 --> 00:46:47.190
and HIPAA. This helps companies get compliant fast and stay compliant.

00:46:47.270 --> 00:46:54.035
More than ever before, trust has the power to make or break your business. Learn more at vanta.com/lenny.

00:46:54.035 --> 00:46:56.835
And as a listener of this podcast, you get $1,000

00:46:56.835 --> 00:46:57.955
off Vanta.

00:46:57.955 --> 00:47:00.275
That's vanta.com/lenny.

00:47:00.680 --> 00:47:06.520
One thing I mentioned recently on the podcast, I heard that speaking of the code that you have of, like, humans writing code,

00:47:06.840 --> 00:47:11.480
uh, data labeling companies are buying code that was written before 2021,

00:47:11.480 --> 00:47:17.855
2022, before AI became a thing. It's, like, very valuable data. Archezonal human code. Yeah. Ex

00:47:18.655 --> 00:47:28.415
That's exactly right. And it's so interesting that that's exactly the kind of code used to build this model. Well, what's interesting so I wanna I wanna clarify there. So I did not have a human

00:47:29.375 --> 00:47:34.710
write the code all by hand because I actually think that that's sort of it feels silly to me.

00:47:35.110 --> 00:47:44.815
Like, I don't really care because I know if if an engineer is not using AI, like, I'm not gonna work with them. I don't really care. It's like it's sort of like, am I gonna race

00:47:45.295 --> 00:47:53.695
a human against a car? Like, I probably wouldn't do that. But I would race a human in a car versus another human in a car and say which one's better.

00:47:54.095 --> 00:48:06.960
And in this case, what the the way the benchmark is structured is, yeah, like, these human engineers use AI, but they use it in a way that I could not because I didn't understand it, I didn't have time, and I didn't really wanna, like, go in and try to understand the code base, to be honest.

00:48:07.360 --> 00:48:11.680
And I think that's a really important thing when we think about benchmarks is

00:48:12.235 --> 00:48:19.835
AI is a broadly distributed technology that any human can use. And when we are benchmarking against humans,

00:48:20.075 --> 00:48:27.275
AI against humans, we're actually really always talking about one human using AI versus another human using AI because AI doesn't use itself.

00:48:27.900 --> 00:48:31.820
It it may be able to in this, like, slightly somewhat recursive way, but there's

00:48:32.540 --> 00:48:42.985
in any real use case, there's always a human, like, pretty close to it making sure that it's working. Okay. I wanna try to wrap up our first bucket. There's so much to talk about. I made a little list of things that I think people,

00:48:43.225 --> 00:48:56.745
uh, should do based on your predictions to be successful. We'll talk about this at the end too, but just a few things. Uh, one is start using Codex or ClawCode more and more for the work you're doing, and especially the browser, use tools inside of it.

00:48:58.030 --> 00:49:06.910
Two is allow your a allow agents to be to use your products. If you're building a SaaS tool, make it easy for agents to be a a a user, essentially.

00:49:07.790 --> 00:49:08.750
Three is

00:49:09.070 --> 00:49:23.445
start thinking about some Slack bot that you can work with, like try out tools. Like, I know Slack has their own Slack bot that I think is really good too, and I haven't played with it, but people really like it. So look for, I guess, a tool that could become the AI agent within your company.

00:49:24.405 --> 00:49:26.085
Buy SaaS stock ASAP,

00:49:27.960 --> 00:49:29.720
not investment advice.

00:49:31.640 --> 00:49:34.840
I think that's totally right. I will like, my slight tweak is

00:49:35.000 --> 00:49:36.440
when you're thinking about

00:49:37.400 --> 00:49:39.320
building your software for agents,

00:49:40.035 --> 00:49:50.915
the current model is I'm building a CLI that an agent uses, but they're you're using it in a sort of, like, I'd they're debt being I delegated a task to the agent. The agent's using the CLI.

00:49:51.075 --> 00:49:56.590
And what we what where I think it's going is you and the agent are using the app together.

00:49:56.830 --> 00:50:02.270
The agent's probably using the CLI, but you're using the web interface, and they're they both need to be in sync.

00:50:02.750 --> 00:50:05.630
And that is, I think, a new challenge that's really interesting.

00:50:06.245 --> 00:50:07.125
Awesome.

00:50:07.125 --> 00:50:10.005
Anything else before we get to our next, uh, category?

00:50:10.085 --> 00:50:11.045
By SaaS.

00:50:12.405 --> 00:50:13.845
That's the title.

00:50:14.885 --> 00:50:18.325
Oh, man. Okay. So the second, uh, category

00:50:18.325 --> 00:50:23.380
of predictions is around just the shape of the work that we're gonna be doing is gonna change.

00:50:24.020 --> 00:50:32.580
What do you predict? There's all this interesting stuff in terms of in terms of the shape of work. Like, once you're in this land where you've got, you know, these you've got async

00:50:33.355 --> 00:50:37.835
async agents off that you delegate work to, then you've got your, like, Codex Cloud Code,

00:50:37.995 --> 00:50:41.035
like, work surface, that that starts to happen. So

00:50:41.355 --> 00:50:43.275
one thing that we see a lot internally,

00:50:43.275 --> 00:50:44.955
and you also see this in the big model companies,

00:50:45.690 --> 00:50:46.330
is

00:50:46.570 --> 00:50:50.090
the number of pull requests that you get is like, skyrockets.

00:50:50.570 --> 00:50:57.290
You know, we have people, you know, in consulting or in ops roles or whatever who are or or edit editors just, like, making pull requests.

00:50:59.365 --> 00:51:06.165
And, hey. That's really cool, and it's a very different shape of work where you should you can expect

00:51:06.725 --> 00:51:12.805
that a higher percentage of your company or your users are gonna be doing things that previously only technical users can do.

00:51:14.245 --> 00:51:14.965
And

00:51:15.180 --> 00:51:17.100
what that does is

00:51:17.420 --> 00:51:22.220
it creates all this pressure on the other end for the people who have to deal with

00:51:22.540 --> 00:51:25.500
all of the new code for how to deal with that.

00:51:25.820 --> 00:51:30.540
And so I think there's a lot of there's a lot of interesting things that happen with that. Like, so for example,

00:51:34.245 --> 00:51:37.205
like, OpenClaw. I mentioned that earlier. Pete

00:51:37.685 --> 00:51:48.580
gets, like, thousands of pull requests a day on OpenClaw, and then he has, like and then he just spins up, like, 50,000 codex instances and then sorts through them and then merges, like, a thousand of them.

00:51:49.220 --> 00:51:52.740
It's really crazy. I actually think that that's going to be more and more common.

00:51:53.540 --> 00:51:54.180
There's like

00:51:55.460 --> 00:51:58.180
it brings up a lot of really interesting questions around

00:51:59.635 --> 00:52:00.435
which

00:52:00.675 --> 00:52:02.355
pull request should you merge.

00:52:02.835 --> 00:52:03.475
And,

00:52:04.195 --> 00:52:10.115
you know, when you whenever you add capacity in one part of your process, like, it breaks things.

00:52:10.940 --> 00:52:14.220
It used to be really hard to build things, and now it's very easy. So

00:52:14.620 --> 00:52:22.060
the the point is not, can we build it? It's like, would it make sense with the rest of what we've built, and how do we keep a, like, sense of a coherent whole?

00:52:22.220 --> 00:52:30.635
And, also, what do we delete? I think Anthropic does this really well. Like, they they delete a lot of stuff from Cloud Code to make sure that it's not bloated.

00:52:31.115 --> 00:52:35.835
So I I think there's a there's a lot of that gonna happen on one side. There's a lot of

00:52:36.235 --> 00:52:42.180
nontechnical people can do technical work, and then technical people are in charge of making sure that that work

00:52:42.420 --> 00:52:51.140
gets into a product or into a process in a cohesive, coherent way. And, also, their product people are gonna be doing that too. And I think that's

00:52:51.300 --> 00:53:01.715
that's quite cool. Something I'm hearing from people is that now that everyone can do everything, like, engineers can design, PMs can code, marketing people can ship stuff.

00:53:02.115 --> 00:53:05.315
There's just this, like, confusion about what the hell is my job anymore.

00:53:05.740 --> 00:53:10.940
Yeah. What am I responsible for exactly? Like Yeah. Am I supposed to be shipping stuff? Am I still a marketing person?

00:53:11.180 --> 00:53:16.060
And it's just creating a lot of confusion and uncertainty in the world just like that. For

00:53:16.300 --> 00:53:16.940
real.

00:53:17.180 --> 00:53:36.540
And one of the things that I think is special about every is everyone is sort of a generalist and really loves, like, having their fingers in a lot of different pots or whatever the metaphor is. I think that'll probably settle down at some point, and it'll feel more normal. Like, marketing people are still gonna do marketing even if they're touching the website. Like, that's just part of marketing now.

00:53:36.860 --> 00:53:42.860
But I also think that you can get a lot further being a generalist now, and that's, like, really cool, especially for for smaller companies.

00:53:44.305 --> 00:53:49.825
The the other thing that I think is interesting is there are definitely some new job roles that are a thing.

00:53:51.905 --> 00:53:52.625
And

00:53:52.705 --> 00:53:55.745
the thing that is becoming really clear

00:53:56.545 --> 00:54:00.270
is the whole forward deployed engineer concept, I think, is for real.

00:54:00.670 --> 00:54:02.190
And it comes out of

00:54:02.910 --> 00:54:04.430
every agent needs a human.

00:54:05.950 --> 00:54:15.195
You like, you go to the big model companies, they have they they have these agents that run internally. They have, like, teams of people that run these agents. You know? And I I don't think those teams are going away.

00:54:15.435 --> 00:54:22.235
The models are gonna get more powerful. The agents are gonna get more powerful, and the number of agents is gonna grow, but people are still gonna manage them.

00:54:23.595 --> 00:54:27.450
And so that looks like a very specific kind of person.

00:54:28.010 --> 00:54:28.650
And,

00:54:29.130 --> 00:54:41.690
you know, we have a couple of those people internally here, and it's like the the people who are in charge of making sure your agents are working and doing the right thing. We also do consulting. So we we we lend that out to people, and and I think that's a big

00:54:42.465 --> 00:54:44.865
that's a big thing that that people want. And

00:54:45.185 --> 00:54:48.465
it's another one of those places where you're like, automation

00:54:49.585 --> 00:54:53.825
was supposed to take away jobs, but it looks like it just created one or many.

00:54:54.625 --> 00:54:55.025
You know?

00:54:55.970 --> 00:54:58.690
And there's a specific type of engineer that really loves

00:54:58.930 --> 00:55:06.370
you know, Nitesh, who's one of our, uh, who who fits this. He's an AI engineer, and he he fits sort of forward deployed,

00:55:06.530 --> 00:55:16.255
um, category, he's on our team. He spends most of his time actually talking to one of our agents in Slack. We have an agent internally called Claudia, which runs our whole consulting practice.

00:55:16.655 --> 00:55:17.215
And

00:55:18.335 --> 00:55:21.055
and he spent a lot of time in Slack. Like, there's there is code,

00:55:21.740 --> 00:55:28.220
and he is using Cloud Code and other things like that. But a lot of it is just talking to it and being like, why did you do this dumb thing? Like, let's

00:55:28.860 --> 00:55:40.565
let's fix that. You know? And so there are certain kinds of engineers that I think love that and love having their hands on the latest thing and also love making this, like, being that's, like, in in the works in a workspace,

00:55:40.645 --> 00:55:42.485
and it looks a bit different than

00:55:42.965 --> 00:55:45.365
more traditional building more traditional software.

00:55:45.685 --> 00:55:47.525
And your sense there is we're not gonna

00:55:48.040 --> 00:55:59.160
we're not near a place where these agents don't need a human. You've said that so many times now that agents need a human, and there's kind of, like, the setup part, and then there's the maintaining it forever part. And it feels like both are important

00:55:59.925 --> 00:56:08.005
Is what I'm hearing, like, this is gonna be a job for a long time. AI is not gonna get smart enough to just automate its be fully automated for a while? Yes. I'm simultaneously

00:56:08.005 --> 00:56:10.565
extremely AI pilled, extremely,

00:56:11.045 --> 00:56:17.460
and very bullish on humans and the role of humans in making sure that AI is working well. Interesting.

00:56:17.940 --> 00:56:49.940
Okay. So the two kind of buckets here that you're talking about, one is, um, like, the way I think I hear what you described earlier is just the pace of shipping software and everything is just increasing, which also means, uh, there's so much more work reviewing all this sloppy output. Uh, was just talking to a data science friend, and he was saying how his team is just his data science team is just their job used to be do analysis, answer questions, see if this experiment was a good was a was positive. Now it's just everyone's doing that, and they're sharing their results, and they're and they're like, no. This is not correct.

00:56:50.180 --> 00:56:53.460
And most of their job is now reviewing bad data science work.

00:56:53.620 --> 00:56:58.295
Which is a problem, and it means that and the same thing are is happening with engineers.

00:56:58.455 --> 00:57:03.175
And it means that you need more like, you actually need that engineers

00:57:03.175 --> 00:57:05.895
for this, and you need data scientists. And

00:57:06.215 --> 00:57:15.960
it means that you haven't set up the appropriate systems or agents to help you with this. So, like, the way that it works inside of the big model companies, for example, like, at least one of them has

00:57:16.280 --> 00:57:20.520
literally a data science bot that every single person in the org can query

00:57:20.680 --> 00:57:23.960
that, um, is hooked up to their data warehouse that

00:57:24.415 --> 00:57:29.935
knows who's who so that it knows at the warehouse level, like, who has permission to access what.

00:57:30.335 --> 00:57:35.375
And so all of the basic questions because there there's a team that sets up this bot.

00:57:35.775 --> 00:57:40.240
All of the basic questions that people might wanna ask that it sometimes gets that might get wrong,

00:57:40.560 --> 00:57:52.240
that they're constantly making sure it's getting it right. And so the data science team doesn't have to answer all the, like, bullshit questions because there's another team building an agent that that that is set up to do that really well.

00:57:53.005 --> 00:57:56.285
But if the team didn't exist, the data scientists would hate their lives.

00:57:56.605 --> 00:58:08.925
Yeah. It does, though, make the job maybe less fun because you're just sitting there, you know, gardening people's sloppy work where Well, that's what think is, like critical. It it can actually make the job better because

00:58:08.550 --> 00:58:15.750
for the data scientists, you are now not dealing with all the silly requests. You're dealing with, um, the deep the deeper questions

00:58:15.750 --> 00:58:20.630
that are harder for the the team who's dealing with all the basic requests and building agent to do that.

00:58:22.115 --> 00:58:33.795
It's it's, like, filtering all that stuff out so you can focus. Here's a question I've been thinking about. I was not planning to talk about this, but it's something that, uh, I've been thinking about. The question is which product tech role is the least

00:58:34.180 --> 00:58:35.060
changed

00:58:35.380 --> 00:58:37.380
now? So, like, engineers,

00:58:37.620 --> 00:58:41.060
100% of code AI now. It's like a completely different job.

00:58:41.220 --> 00:58:42.660
Uh, product management,

00:58:42.980 --> 00:58:58.045
a lot of the, you know, PRDs are you don't have to write as much. You can ship code. You don't have to wait for people. Design the whole design process, uh, dead according to, um, recent guests. Just like there's no time to do the whole design process. Very different role. Data science, very different work now.

00:58:58.525 --> 00:59:00.365
Um, there's marketing. There's sales.

00:59:00.990 --> 00:59:04.910
So here's the question. What do think is the least fundamentally changed

00:59:04.910 --> 00:59:07.630
role so far? Well, one interesting thing

00:59:07.870 --> 00:59:08.510
is,

00:59:09.150 --> 00:59:13.470
you know, I don't know if this counts, but, like, CEOs and investors,

00:59:13.470 --> 00:59:18.495
it seems still very, very optional whether or not they use this stuff. Mhmm.

00:59:18.735 --> 00:59:26.975
It seems that way. I I I think the opposite is actually true. Like, my experience and we do a lot of this with senior executives and senior leadership teams.

00:59:27.215 --> 00:59:46.025
My experience is that your company is only gonna go as far as your CEO goes in AI, and it's not something you can delegate. You have to have your hands in it, uh, because you don't otherwise, you don't have an intuition for it. But for a long time, it has seemed like, yeah, that's something that the people who are doing the work have to do, but, like, I don't have to do that. Like, I'll just tell them what to do.

00:59:46.425 --> 00:59:47.145
And

00:59:48.265 --> 00:59:52.825
and so I think if you're a CEO, you kinda can get away with your day looking very similar.

00:59:53.850 --> 00:59:58.890
I I think that will change rapidly at some point where it'll be like, oh, no. I'm, like, way behind. But for now,

00:59:59.210 --> 01:00:04.810
because or maybe even middle managers, like, those kinds of people, I think, are are it's fairly similar.

01:00:05.885 --> 01:00:07.005
I think, like,

01:00:07.725 --> 01:00:10.285
maybe sales because That's exactly

01:00:10.445 --> 01:00:23.150
what person. That that's yeah. That's my vote. You know? There it's sort of creeping up in the kind of BDR. Like, we can deal with a lot of, you know, BDR type type queries. You're only talking to, like, people who actually want it.

01:00:23.550 --> 01:00:27.070
And you can do for sales, it's, like, re it's so useful to

01:00:28.190 --> 01:00:28.990
to,

01:00:29.950 --> 01:00:32.525
like, do research. Like, my favorite

01:00:32.525 --> 01:00:35.565
codex like, one of my favorite codex experiences is

01:00:36.285 --> 01:00:37.325
we're hiring

01:00:37.485 --> 01:00:39.085
a head of l and d.

01:00:39.485 --> 01:00:40.125
And

01:00:40.685 --> 01:00:41.405
I

01:00:41.645 --> 01:00:54.380
you know, we always put out a job post, whatever, but I was like, I feel like there's this company called General Assembly in New York, and they do like, they've done really good technology education for a long time. And so I was like, I feel like someone who is into

01:00:54.940 --> 01:01:03.995
who who who worked at General Assembly and is now into AI would be really good. And I just, like, literally typed it into Codex and then, like, went off and was doing something else, and I came back.

01:01:04.475 --> 01:01:10.235
And it found, like, this the perfect guy. It was, like, worked at general assembly, was an instructor,

01:01:12.475 --> 01:01:16.200
like, is super AI pilled and follows me on Twitter.

01:01:17.080 --> 01:01:22.920
So I just DM'd him, and then I had dinner with him. And it's like, that's crazy. You know? That would have taken so long before

01:01:23.320 --> 01:01:38.815
and, uh, super valuable for sales, for recruiting, all that kind of stuff. Yeah. Sales is where my mind went. Like, the top of funnel AI is helping a lot with sourcing and qualifying things like that. It feels like the the work of a salesperson is not fundamentally different. Yeah. And

01:01:39.135 --> 01:01:44.575
and customer support's fundamentally changed. So that's interesting. Sales. So far, so good for the for those folks. Yep.

01:01:45.270 --> 01:01:45.910
Okay.

01:01:46.150 --> 01:02:04.225
So maybe just summarizing some of the predictions in this bucket of just, the shape of the work, how it's gonna change. What I'm hearing so far is there's gonna be a lot more reviewing of other people's output as a part of the work. And then two, there's gonna be a lot of, like, almost babysitting of AI agents to make them do the thing you want them to do for deploying and then just

01:02:04.625 --> 01:02:07.345
gardening them along the way, make sure they continue to do their work.

01:02:08.705 --> 01:02:13.185
Anything else before we get into our third bucket? I would sort of split it into

01:02:14.350 --> 01:02:17.230
less babysitting agents and more

01:02:17.710 --> 01:02:26.350
your forward deployed team is trying to build a whole system that makes it so that people who have less knowledge can use that system without,

01:02:26.670 --> 01:02:27.950
like, doing something dumb.

01:02:28.475 --> 01:02:30.555
And that's, like, a really interesting

01:02:30.555 --> 01:02:31.355
engineering

01:02:31.355 --> 01:02:31.995
challenge.

01:02:32.555 --> 01:02:35.035
I think babysitting kinda makes it feel like it's

01:02:35.515 --> 01:03:03.155
yeah. You're just kinda, like, you know, waiting for it to fuck up and then fixing it or whatever. And you you that can be the case, but I think a lot of it is this extremely interesting engineering challenge of building a system for to enable everybody else in your organization to do what used to be a technical job. And then if you're not one of those people, like, you're the data scientist or whatever, you can go a lot deeper with AI into, like, really important questions that eventually probably filter into the work that the, you know, the forward deployed engineering team is doing,

01:03:03.475 --> 01:03:08.675
but is, like, more generative and more new and and and you're you're dealing with harder questions.

01:03:09.075 --> 01:03:11.475
One other one last thing that I think is really interesting

01:03:11.830 --> 01:03:18.070
is I think that we will be reading way more AI generated writing in documents and emails,

01:03:18.390 --> 01:03:22.870
and we will like it. And I think we're we will we are already doing this

01:03:23.430 --> 01:03:25.750
in coding where we read plan documents.

01:03:26.695 --> 01:03:33.015
Like, I don't want an engineer to handwrite a plan document. That would be very silly. It would be it would be obviously silly.

01:03:34.775 --> 01:03:38.055
And I think the same is true. You know, when we did our our,

01:03:38.455 --> 01:03:50.960
uh, quarterly planning for every at the end of twenty twenty five, We did it all with Notion agents, and we just had a bunch of Notion agents. And we're we had really one Notion agent, and then we had a top level company strategy.

01:03:51.600 --> 01:03:52.720
And then we had

01:03:53.520 --> 01:03:55.600
everybody in the company just,

01:03:55.760 --> 01:03:56.480
um,

01:03:56.880 --> 01:04:00.565
talked to an agent, and it asked them about what happened last year,

01:04:00.885 --> 01:04:12.405
how did it go, what were your goals, what what do you wanna do this year, what are your metrics? They pushed back, and then it was like, how does it how does this relate to the overall company idea? Like, all that kind of stuff. And then I got I got these, like, incredibly good

01:04:12.840 --> 01:04:14.120
AI generated,

01:04:14.520 --> 01:04:27.560
like, strategy reports or or plan like, quarterly plans for for each part of each team. And then I could go in and be like, okay. Who needs to who's, like, who needs to talk to each other? Like, which teams need to talk to each other that, like, don't know they need to talk to each other?

01:04:29.515 --> 01:04:30.155
And,

01:04:30.715 --> 01:04:38.315
you know, who's which one of these is, like, at like, actually low quality, or which one of these is high quality? Like, all that kind of stuff makes it it makes it a lot easier to process.

01:04:39.035 --> 01:04:42.830
And I see that all the time now. Like, I I I consistently

01:04:42.830 --> 01:04:50.830
get AI generated stuff, and there is a difference between an AI generated document that's slop and not. And the slop one

01:04:51.150 --> 01:04:51.790
is

01:04:52.110 --> 01:04:55.230
it took them less time to make it than it takes me to read it,

01:04:55.985 --> 01:04:58.385
and they don't stand behind every line.

01:04:58.865 --> 01:05:03.825
So my expectation is if you send me an AI generated document, I think that's great. And

01:05:04.305 --> 01:05:09.360
if we talk about it and it's clear you have no idea what's in it, like, big no no. Not allowed to do that.

01:05:10.240 --> 01:05:12.000
And I I think we have this

01:05:12.560 --> 01:05:13.840
this aversion

01:05:13.840 --> 01:05:17.280
to AI generated stuff that will go away because

01:05:17.680 --> 01:05:21.840
the kind of strategy document that g p t five point five can write when it's directed well by

01:05:22.185 --> 01:05:27.065
someone on my team is way better than, like, them just, like, dinking and dunking, like like,

01:05:27.225 --> 01:05:31.945
their fingers on the keyboard. Right. Like, most people are really bad at writing their documents.

01:05:31.945 --> 01:05:33.705
The bar is low. Yeah.

01:05:33.945 --> 01:05:34.425
And

01:05:34.950 --> 01:05:50.345
and same thing with email. Like, I most of my email is written by g p t five point five and Codex right now. And I would I honestly would prefer it to say that it's coming from g p t five point five, and I may change it to do that. But I had this I had this experience

01:05:50.425 --> 01:05:53.465
the other day where I had this I had to send an email to

01:05:53.945 --> 01:05:55.625
to one of our investors,

01:05:55.865 --> 01:05:56.425
and

01:05:57.065 --> 01:05:57.625
I

01:05:57.945 --> 01:06:03.305
asked Codex to go do it. And use like, Codex knows to ask me, and it usually does. But this time,

01:06:03.920 --> 01:06:04.960
it didn't.

01:06:05.440 --> 01:06:07.200
And it just sent the email,

01:06:07.440 --> 01:06:22.505
and I didn't look at it at all. And I was like, fuck. And so I went to my sent and looked at it, and I was like, oh, this is exactly what I would've sent. And so it's like, it's pretty close to to that a lot of the time. It can be, like, a little over formal, and there's a couple of things that that

01:06:23.065 --> 01:06:26.905
it's just when you really think about it, most of your email is kinda

01:06:27.785 --> 01:06:28.425
it's not

01:06:29.560 --> 01:06:33.000
it's kind of rote. It's kind of prosaic. It's kind of

01:06:34.200 --> 01:06:49.905
I I definitely wanna be the one to think about what it should say, like, what what it should say, but the actual sentences don't matter that much to me, usually. Sometimes I do a lot. And this is coming from a writer. Like, I care a ton about writing. I think that human writing is incredibly important,

01:06:50.385 --> 01:07:03.480
and I expect we only publish human writing. Well, actually, we publish in the mix of human and AI writing, but we always label it. Sometimes it's nice to have an AI coauthor on certain things. Um, I absolutely think that, uh, human writing is important, and

01:07:03.960 --> 01:07:05.560
I think that the

01:07:05.880 --> 01:07:17.615
the the reaction or the aversion to AI writing is silly. It's such an interesting lens on that because when people think about AI writing, I think about social media and videos.

01:07:17.855 --> 01:07:35.510
And your point is internally, if you're just, like, working on planning and documents and email and things like that, like, that is much less scary that it's AI written in your to your point, people are already doing this. You almost prefer it a lot of times because people are really bad doing this anyway. We have this too for external stuff. Like, we publish all these guides, and the guides are often agent

01:07:35.670 --> 01:07:38.550
they're agent assisted, and the agent is a coauthor.

01:07:38.875 --> 01:07:47.275
And they're intended to be read both by humans and by agents. And that's because, like, if you're writing a huge informational thing mean, you do this all the time.

01:07:49.435 --> 01:07:50.155
In

01:07:50.235 --> 01:07:55.230
order to, like, really apply it, the best way to do that is just, like, have your agent ingest it. And

01:07:55.550 --> 01:08:02.430
remember the next time I'm, you know, doing pricing to, like, remind me of this guide, and we'll go through it together or whatever.

01:08:02.910 --> 01:08:05.150
It allows you to operationalize

01:08:06.110 --> 01:08:12.095
the ideas much better, and it allows you to go much deeper because agents can read, like, 10,000 pages in, a second.

01:08:12.495 --> 01:08:21.215
And so you you can you talk to the human about the story and the stuff that matters and the core ideas, and the agent has all the details that it can then apply for you when you need it. Awesome.

01:08:22.040 --> 01:08:26.680
Anything else in this category before we get into our final category? No.

01:08:26.920 --> 01:08:34.440
Okay. Let's do it. So the final bucket is just who will be successful in this AI future that we are approaching

01:08:34.695 --> 01:08:42.055
slash what should people be working on to be successful in this next year or two? I am super,

01:08:42.455 --> 01:08:44.295
super bullish on PMs.

01:08:45.415 --> 01:08:47.655
And I know that your audience will probably love that,

01:08:49.000 --> 01:08:50.920
but my

01:08:51.400 --> 01:08:52.040
my

01:08:52.600 --> 01:08:57.240
anecdotal case that has convinced me of this is we have this guy internally. His name is Marcus,

01:08:57.400 --> 01:09:00.200
and he runs Spiral, which is our writing app.

01:09:01.640 --> 01:09:02.680
Marcus

01:09:02.625 --> 01:09:04.145
is a PM by training.

01:09:04.625 --> 01:09:05.345
He

01:09:05.505 --> 01:09:06.705
he previously

01:09:06.705 --> 01:09:07.425
ran

01:09:07.905 --> 01:09:09.505
Axios Axios'

01:09:09.505 --> 01:09:10.385
writing product

01:09:10.625 --> 01:09:16.065
and was a was a PM and had a big team, and it got to, you know, tens of millions in of of revenue in ARR.

01:09:17.080 --> 01:09:17.720
And

01:09:18.280 --> 01:09:29.400
he took a year off that job and just got super AI built and just learned how to use cursor basically really well. Now I think he uses Cloud Code, but he was extremely cursor built for a long time.

01:09:30.465 --> 01:09:33.905
And he's I would call him, like, lightly technical.

01:09:35.825 --> 01:09:45.510
Like, knows what a database migration is. Like, if he has to look at the code, I think he can understand it, but he's like, I we never could have hired him to do this job even a year ago.

01:09:46.070 --> 01:09:53.109
But the coding models have gotten good enough that he can pair the kind of the technical knowledge that he does have

01:09:53.590 --> 01:09:58.790
with his really spiky product sense and sense for writing and sense for users,

01:09:59.525 --> 01:10:06.325
And it's, like, it's so dangerous. Like, he ships faster than almost anyone on the team, and he has such a eye for

01:10:07.445 --> 01:10:23.090
every single user, every single conversation. Like, what does it mean, and how do we collect it into a story about, like, where we wanna go next, and what are the issues we need to fix, and, like, all that kind of stuff. And I think that he feels liberated because he doesn't have to organize a whole team of people to do that. He can just do it.

01:10:23.410 --> 01:10:45.090
And it's super impressive, and it makes me very, very bullish on any PM who gets, like, really AI built. Music to my ears, Dan. Uh, you're making a lot of very happy listeners here. Uh, I've been saying this for a long time too. It's just like the skills you need to build are the things like, the building now is done for you. What do you need to be good at? Figuring out what to build, figuring out if it's great, figuring out what problems to solve.

01:10:45.890 --> 01:11:02.825
So I love that you're actually seeing this come to fruition. I I I really believe it. This could be the highest rated podcast episode of my whole podcast. They're on I love it. Hell, yeah. It's gonna be okay. SaaS is SaaS is back. PMs are back. You know? This is the most contrarian episode I've ever done.

01:11:04.265 --> 01:11:10.080
Oh my god. So okay. So the other the other people that I think are gonna be, like, super superpower

01:11:10.080 --> 01:11:14.880
people, and I again, I this is because we see this internally, is full stack designers.

01:11:15.520 --> 01:11:16.800
If you're a designer

01:11:17.120 --> 01:11:20.800
and you're in these tools all the time, you're so used to,

01:11:21.405 --> 01:11:27.245
okay, okay, I make this beautiful interaction, and the engineer, like, just doesn't wanna do it, or it doesn't, like, happen the way I think it should happen,

01:11:27.485 --> 01:11:33.645
or you know, there's all this stuff. And I see so many designers for us internally or externally where

01:11:34.690 --> 01:11:41.570
they now feel so empowered to, like, go build stuff because they're like, I have all these ideas to make things look amazing and these interesting interactions.

01:11:41.730 --> 01:11:59.105
And that's the exact thing that it's really hard to do with vibe coding because it just all looks the same. So it all looks like slop, and they can make stuff that looks so different. And now they can actually build it. And what you see when we work with them internally is now they're just, like they're just making poor poor requests. Like, they don't they

01:11:59.665 --> 01:12:16.330
don't need to hand it off as much. Sometimes they do. But, like, a lot of times, they just make poor requests, and it's like, the thing is built, and that's it. And I think it's incredible for the way that companies work, but it's also there's a huge opportunity for those people to become entrepreneurs and, like, start their own thing because they can they can make stuff now. And I think

01:12:17.165 --> 01:12:30.445
designers are such creative people, and I think AI is, like, a super tool for anyone like that. I so agree. Even though there's cloud design, there's all these AI design y tools, like, once you see it, you're like, that's definitely cloud design.

01:12:31.060 --> 01:12:36.820
They're like the creativity, to your point, is gonna it just feels like it's gonna be more and more valuable to to

01:12:37.220 --> 01:12:40.900
to stand out from all the slop that people are shipping and launching constantly.

01:12:41.220 --> 01:12:48.955
So I completely agree. It's it's interesting that designer roles I do I do research on the job market. And interestingly, designer roles have not

01:12:49.435 --> 01:12:54.795
grown in a while. So I'm waiting to see if that becomes a big trend. Just like we need more designers.

01:12:54.955 --> 01:12:57.595
That is really interesting. We'll see. Yeah.

01:12:58.075 --> 01:13:04.580
We'll see. We'll see. That that might be a way to predict this is are people hiring more designers? I don't know. That is interesting. Yeah.

01:13:05.220 --> 01:13:05.940
Alright.

01:13:06.660 --> 01:13:09.300
Uh, that's so PM designer thriving.

01:13:09.300 --> 01:13:10.900
Kilometers designer thriving.

01:13:11.220 --> 01:13:13.700
Um, I also just think generally the

01:13:13.940 --> 01:13:15.300
AI jobpocalypse

01:13:15.300 --> 01:13:20.915
is not really a thing. Absolutely. We see companies starting to reorganize, and I think that makes a lot of sense.

01:13:21.795 --> 01:13:30.995
I I think, to be honest, a lot a lot of the reorganization, you can say it's AI, but it's like we overhired and, like, the company's not doing as well and all that kind of it was, like, coming, and this is a good excuse.

01:13:31.680 --> 01:13:37.840
But the, like, mass unemployment thing, I think that, like, some AI CEOs are talking about, like, I think that's not gonna happen.

01:13:38.320 --> 01:13:42.880
The the pattern that I see so far and, again, I don't have a total crystal ball, but I

01:13:43.200 --> 01:13:49.195
I do feel like we've seen enough of the new model drops to, like, have some sense of how this is going is that

01:13:49.915 --> 01:13:53.675
what a new model drop does or what models do in general is they make

01:13:54.075 --> 01:13:57.835
yesterday's human competence cheap. So what I mean by that

01:13:58.620 --> 01:14:08.540
is they ingest all this data of what what has happened already, and they make it really cheap to deploy that in in whatever situation you want as your as your own. Right?

01:14:10.620 --> 01:14:11.980
And what happens then

01:14:12.305 --> 01:14:13.025
is

01:14:13.185 --> 01:14:18.145
every this is a new this is a new power that everyone has, so it gets adopted super rapidly.

01:14:18.385 --> 01:14:21.185
And it's and suddenly that stuff is everywhere.

01:14:21.185 --> 01:14:27.745
It's like, suddenly, anyone can make a landing page. There's new landing pages everywhere. Suddenly, everyone can write. There's, like, slop tweets everywhere.

01:14:28.250 --> 01:14:31.130
But what's interesting is because

01:14:31.690 --> 01:14:32.570
it's

01:14:33.210 --> 01:14:38.010
all from because it's all coming from these models and everyone's using basically the same models,

01:14:40.175 --> 01:14:44.095
it all looks the same if you use it in the in the most default basic

01:14:44.335 --> 01:14:44.895
way.

01:14:45.295 --> 01:14:46.735
And so that's

01:14:47.135 --> 01:14:56.730
it becomes commoditized. Like, it's not valuable anymore. And what humans do is we sort of go in there, and we're like, yeah. We we have all this, like, frozen human competence from yesterday.

01:14:57.130 --> 01:14:59.450
How do I use this, like, make something new and interesting?

01:15:00.650 --> 01:15:01.290
And

01:15:02.330 --> 01:15:05.050
I really think that structurally,

01:15:05.210 --> 01:15:10.315
because of the way the models work, because of the financial incentives of model of model companies to, like, make them,

01:15:11.995 --> 01:15:14.075
uh, compliant and aligned,

01:15:14.395 --> 01:15:15.275
structurally,

01:15:15.275 --> 01:15:27.290
they're always going to be trailing behind those people who are taking taking the models and using them to make new expertise or or make new things that haven't been done that way before for their very, very particular situation.

01:15:27.850 --> 01:15:34.410
And that stuff is gonna get incorporated into the models, but, again, it will create room for people to, um, to push further ahead.

01:15:34.810 --> 01:15:39.655
And I think that you see this in a small way in, like, pretty much all the jobs is, like,

01:15:40.215 --> 01:15:40.935
engineers.

01:15:41.095 --> 01:15:51.770
Suddenly, everyone's an engineer. That doesn't mean we fire the engineers. There's, like, way more demand for engineers because you need the engineers to, like, figure out, okay. This is all slap. How does this actually how should this actually go in our code base?

01:15:52.570 --> 01:15:56.490
And I think that's something that the benchmarks rising don't doesn't really capture,

01:15:56.650 --> 01:15:57.129
and,

01:15:57.690 --> 01:16:24.690
uh, it feels like a thing that will take a long time to change. People may be hearing in this prediction here of just, okay. The jobocalypse not gonna people are not gonna be all fired. There's gonna be human jobs remaining for quite a while. It may be almost too comforting because you may you probably have to change the way you operate to still have a job in the future. Do have any sense of just, like, here's what you need to do to not be one of these layoffs?

01:16:25.090 --> 01:16:27.570
Yes. Yeah. And I think that is actually super important.

01:16:28.815 --> 01:16:31.535
The only thing you need to do is ride the models,

01:16:32.015 --> 01:16:55.770
and that means use them for whatever it is that you do. You know, we've talked about how Codex and Cowork are becoming the sort of standard operating system for work. If you're just doing that and when new models come out, you're trying them and figuring out, okay. How can I now there are new powers? How can I use them? Instead of just being like, I'm gonna, like, try to ignore it because it, like, makes me afraid, which I think is honestly, it's rational. It's a reasonable response.

01:16:56.090 --> 01:16:56.650
And, also,

01:16:57.615 --> 01:17:02.975
if you ride on top of them, they ex extend your powers in a way that

01:17:03.215 --> 01:17:13.055
doesn't leave you behind. Like, you you're you're you're part of the future and part of the way work happens. And I think that

01:17:11.520 --> 01:17:23.440
we're going to need people doing that for a very, very long time. I like this term, ride the model. So the what's like say New Wallet comes out. What do you think someone say working at, I don't know, Salesforce?

01:17:23.440 --> 01:17:33.335
Say APM at Salesforce. What should they do to ride the model? Well, one of the things that's really interesting is a lot of companies, like, handicap their employees from even doing this because,

01:17:33.575 --> 01:17:47.169
like, I don't know what model. I don't know if you can use the latest models in Salesforce. You know? Like, a lot of times you have to wait or it's, you know, whatever. So maybe you have to do it on your in your off time. But the thing that I really like to do with new models is play.

01:17:47.810 --> 01:17:50.850
And there there there are certain things where

01:17:51.490 --> 01:17:57.815
I know it can't quite do it yet, but when a new model comes out, I, like, always turn the rock over again to be like,

01:17:58.695 --> 01:18:00.375
can I do it now? You know?

01:18:01.255 --> 01:18:08.455
So it, you know, it could not do the senior engineer benchmark last time, and I turned it over turned the rock over again, and now it's at a 60 out of a 100, which is, like, really good.

01:18:10.050 --> 01:18:10.690
So

01:18:11.570 --> 01:18:25.205
the way to ride the models is, like, not one specific thing because they're always changing, but it is to be curious and playful to apply the model, the new model to whatever it is that you care about, whether that's your job or something outside of your job,

01:18:25.605 --> 01:18:27.605
and to keep turning over rocks,

01:18:27.605 --> 01:18:28.725
uh, because

01:18:28.965 --> 01:18:30.645
it may not work now,

01:18:30.965 --> 01:18:33.045
but it may work eventually.

01:18:33.125 --> 01:18:35.205
It probably will work eventually. And

01:18:35.750 --> 01:18:39.510
the way that you use it matters. So what's

01:18:39.510 --> 01:18:41.030
really cool is that

01:18:41.750 --> 01:18:42.790
I think people

01:18:43.030 --> 01:18:45.830
think of the edge of AI as being in San Francisco,

01:18:46.310 --> 01:18:49.590
and I actually don't think that that's where it is. I think the edge of AI is

01:18:50.025 --> 01:18:53.705
wherever AI meets, like, a real human doing something.

01:18:54.185 --> 01:19:05.225
Because the people in San Francisco, they're making it, but they don't actually know a lot about how to use it. They don't know or at least they don't know everything about how to use it. They need to see how other people use it. And so you whenever a new model comes out, you get to be

01:19:05.760 --> 01:19:23.185
one of the first person one of the first people in the world to discover what it might be useful for. And that that's it's like a new discovery. And I think that's why, for example, we're in we're in Brooklyn. But I I really think of us, and I think we are, like, quite far ahead of people in San Francisco because

01:19:23.505 --> 01:19:26.545
we just use them for everything. And

01:19:27.105 --> 01:19:28.145
if people

01:19:30.145 --> 01:19:32.385
if people do that consistently,

01:19:32.465 --> 01:19:39.450
I think it's gonna be very hard to lose. That is one of the most amazing things about AI right now is

01:19:39.690 --> 01:19:48.185
no matter how much money you have or a little money you have, you have access to the most advanced AI model. Like, it's not free, so need some money.

01:19:49.625 --> 01:19:56.665
But, like and you can get it immediately when it comes out. Maybe the only people that have an advantage are the people working at OpenAI or Anthropic.

01:19:57.225 --> 01:20:10.340
But, otherwise, it's just, like, available. I know. I was at I was at their event with you their Code to Cloud event with you last week and, um, or a couple weeks ago, and they're they're, like, all using Methodos. And I'm like, god damn it. It's

01:20:10.340 --> 01:20:11.220
so annoying.

01:20:11.700 --> 01:20:19.195
But I I think that's totally true. Like, that is if IBM had invented AI, you can bet it would not be like this.

01:20:19.675 --> 01:20:23.675
And it would be, like, a bajillion dollars and only, like,

01:20:24.315 --> 01:20:27.995
top companies could use it, and they would be using it in the in the weirdest,

01:20:27.995 --> 01:20:29.275
most uninteresting ways.

01:20:29.790 --> 01:20:35.230
And I think there's it's there it's really important that AI was built

01:20:36.350 --> 01:20:37.390
in America

01:20:37.550 --> 01:20:43.230
and in the Silicon Valley culture that's like, we wanna make intelligence too cheap to meter. Like, that's not the default stance.

01:20:43.925 --> 01:20:46.325
And, um, it means that

01:20:47.285 --> 01:20:58.165
everyone has this broadly accessible tool that they can use, and I think that's amazing. That's such a good point. And interestingly, it's also created the most fastest growing companies in history, the biggest companies in history. That's true.

01:20:58.930 --> 01:21:01.650
A way to Those Silicon Valley guys, they're they're smart.

01:21:02.610 --> 01:21:12.545
If I zoom out on the conversation, it's really interesting. There's kind of these two sides to the coin. One is not a lot is actually like, so much is not changing. SaaS continues.

01:21:12.545 --> 01:21:14.145
Jobs not disappearing.

01:21:14.145 --> 01:21:17.985
We're still emailing each other. We're still working in Slack. Like, a lot of the work,

01:21:18.465 --> 01:21:21.425
not changed. On the other hand, every role transformed.

01:21:21.505 --> 01:21:23.665
Engineers don't write code. PMs

01:21:23.940 --> 01:21:43.315
don't write PRDs. Uh, design and design. You know, it's like it's so interesting how much has changed, how much has not changed. I don't know. It's interesting that people think it's gonna be this whole new world, but in many ways, it's okay. It'll continue the way it is with a lot of stuff around the edges. That's that's how I feel. Like, I'm simultaneously so excited, and it feels like everything has changed.

01:21:43.715 --> 01:21:50.930
And I'm so bullish on it and and the and the progress that we're making, all that kind of stuff. And, yeah, I just I feel like there are there are these things where

01:21:51.090 --> 01:21:57.650
they're gonna be pretty similar to how they are, and that's probably good. And I think, generally, our intuitions

01:21:57.810 --> 01:21:59.170
about the future

01:22:00.050 --> 01:22:15.145
the the model that I have of what our intuitions are about the future is the intuitions that people had in the middle ages about, like, what happened at the end of the horizon. You know? It's like, are there dragons? Like, does it drop off into nothingness or whatever? You know? Like, a

01:22:15.465 --> 01:22:19.225
lot of people have a lot of deep intuition that there's something terrible

01:22:19.440 --> 01:22:21.040
gonna happen over the horizon.

01:22:21.360 --> 01:22:36.655
And, also, that, uh, some people are like, there's something incredible. It's gonna change everything. We're gonna all gonna be happy as a utopia. And what happens is you get there, and you're like, there's some really cool things. There's some not cool things, and it's just another horizon. And I think that's

01:22:36.815 --> 01:22:40.734
that's the way to think about the future. And until you get to that place

01:22:41.055 --> 01:22:45.214
where you're starting to see it, and I think we get to see it because we get to see it internally all the time,

01:22:46.175 --> 01:22:48.790
it's important not to let your

01:22:49.270 --> 01:22:56.070
your mind get away from you and being like, this is gonna happen and this is gonna happen and whatever because you're you're gonna tell a story that sounds

01:22:57.430 --> 01:23:28.325
sounds so real in the moment, but later on, you're like, actually, it's much more complex than that in somewhere. It's sort of a both. Everything's changed and nothing has. Um, and once you get there, I think you're you're sort of start starting to see, like, oh, yeah. This is a real thing. Part of it is the AI companies are very good at scaring us about what might might happen in the future, and I think that's actually shifting. I think they've realized maybe we should not freak everybody out about the behaviors. That PR strategy just does not make any sense to me. I I do think that it's, like, genuine, but it's so ineffective,

01:23:28.805 --> 01:23:31.285
and, um, and I I think it's also wrong.

01:23:31.605 --> 01:23:32.325
How

01:23:32.885 --> 01:23:37.205
about we, um, end with maybe just, like, a few things listeners should do

01:23:37.850 --> 01:23:43.690
to be successful over the next year with the the way the world is moving? Buy the models.

01:23:44.010 --> 01:23:45.930
I would try

01:23:46.730 --> 01:23:48.970
all of your workflows

01:23:48.970 --> 01:23:49.609
in

01:23:50.010 --> 01:23:51.290
Codecs

01:23:50.905 --> 01:23:52.584
or Cowork and

01:23:53.305 --> 01:23:59.065
see how that works. And if your company doesn't let you do it on your own time, I would try out some of these

01:23:59.945 --> 01:24:01.705
agent products like

01:24:01.945 --> 01:24:04.665
OpenClaw or Hermes or

01:24:04.270 --> 01:24:11.150
for less technical, there's there's, like, Victor. We have one plus ones. I I would get comfortable with both of those ways of working

01:24:12.190 --> 01:24:16.910
and try to, like, try to have fun. I think there's too much of

01:24:17.485 --> 01:24:23.885
I'm doing this because I have FOMO. Like, it might I might lose my job or, like, I might miss out on this big thing or whatever. And

01:24:24.205 --> 01:24:25.245
the best

01:24:25.885 --> 01:24:30.765
way to actually figure out interesting useful things to do with AI is to, like, do something enjoyable.

01:24:31.220 --> 01:24:56.655
We had a, um, Nikhil Singhal was on the podcast, and the way he described it is you gotta find your moment of joy with AI once you find, like, wow. I can't believe AI did this for me. This is awesome. I'm gonna keep building stuff. Yeah. I agree. If you haven't seen that yet, then it's just like try find try solving it. The thing I hear a lot is just find a problem in your life or work and see if AI can do it. Get a lovable, get a Cloud Code, get a replete. Just try to build the thing. And often it's like, holy shit. This is so cool.

01:24:57.890 --> 01:25:04.930
Dan, is there anything else that we haven't covered? We've gone deep on so much. Is there anything else you wanted to share, anything else you want to predict

01:25:05.170 --> 01:25:31.060
or just say before we get to a very exciting lightning round? Uh, I think we covered it. We we did a lot. This is this is awesome, and I'm very excited to see how well or poorly I do, uh, in a year, and I hope that you hold me to it. And we're gonna we'll have AI score us. How about that? Well Great. Look look at the world like a dance prediction. Here it goes. Well, with that, Dan Shiffer, we've reached a very exciting lightning round. I've got five questions for you. Are you ready? I'm ready.

01:25:31.860 --> 01:25:37.140
What are two or three books that you find yourself recommending most to other people? Um, obviously,

01:25:37.140 --> 01:25:37.860
Annie Dillard.

01:25:40.045 --> 01:25:43.084
I everyone at Every has to read the writing life.

01:25:43.485 --> 01:25:49.165
Like, when you join, you get a copy and you have to read it. Uh, you only have to read the last chapter, though. I think the last chapter is

01:25:49.485 --> 01:25:49.965
incredible,

01:25:50.860 --> 01:25:52.300
and it is

01:25:52.540 --> 01:25:56.220
at the intersection of writing and technology and

01:25:56.780 --> 01:26:08.955
the future and its, like, its relationship to the future and to time. And I think that's, like it's it's everything about every, like, wrapped up into, like, a very tight chapter. It's so good. And I think Andy Dillard just generally is

01:26:09.355 --> 01:26:10.235
fantastic.

01:26:10.715 --> 01:26:15.835
What else do I recommend? I'll just I'll just tell you a couple things that I've read that I'd, like, really liked recently.

01:26:16.235 --> 01:26:21.595
And and whenever I like something, I always just, like, tell everyone about it. So I have recommended these a lot.

01:26:22.610 --> 01:26:27.810
I I've been I've been reading one of the things I I'd learned, which I didn't know, is Churchill's a really good writer.

01:26:28.370 --> 01:26:34.930
And he has a whole history of World War two that he wrote, and it's like a combination history and memoir.

01:26:35.375 --> 01:26:59.380
And I think that's so cool because he was there. You know? He did it. And there's something about what we do at everywhere. I feel some, like, sort of kinship with that of, like, we're building stuff. We're writing stuff, and it's very rare to find people that also do that. And and so Churchill's history of World War two is fantastic. I just finished the first volume. I'm on the second volume. The Nazis just invaded France. Very it's very captivating stuff. Um,

01:27:00.900 --> 01:27:05.735
that's one. I also just I I've been on, like, a little bit of, like, a quantum physics,

01:27:06.135 --> 01:27:07.975
like, kick recently.

01:27:08.215 --> 01:27:15.655
AI is very actually, very good for quantum physics if you get into it. And there's this book called the rigor of angels that I just finished, which is

01:27:17.570 --> 01:27:20.610
it's like a it's a history of ideas that relates

01:27:21.410 --> 01:27:22.450
Heisenberg,

01:27:22.850 --> 01:27:25.650
who has the his uncertainty principle,

01:27:26.690 --> 01:27:27.570
Borges,

01:27:27.570 --> 01:27:27.890
who's,

01:27:30.135 --> 01:27:31.015
like, a

01:27:31.575 --> 01:27:32.455
Argentinian

01:27:32.455 --> 01:27:41.175
fiction writer, has wrote a bunch of great short stories. They're actually starting to get, like, a lot of play now because they're very AI related and, um, and Kant.

01:27:41.735 --> 01:27:42.294
And

01:27:43.255 --> 01:27:46.740
very cool, like, super mind blowing. Lots of, like, interesting

01:27:46.900 --> 01:27:49.220
overlaps with AI stuff. And,

01:27:50.020 --> 01:27:51.459
yeah, highly recommend.

01:27:51.700 --> 01:27:57.540
I feel like we gotta have a whole podcast episode about your reading and, uh, books you recommend. I know this is a a passion of yours.

01:27:58.935 --> 01:28:05.655
My current obsession is the power broker, and I think we talked about it when I was We did. Visiting you. It's just So good. Never ends, but it's surprisingly

01:28:06.535 --> 01:28:08.855
compelling to read through the history of New York.

01:28:09.575 --> 01:28:18.090
Okay. Second question. Do you what is a recent movie or TV show you really see enjoyed if you have time for TV? So I've been watching a lot of basketball, so that's one.

01:28:18.570 --> 01:28:24.970
I'm I became a Knicks fan, like, this this year, so, uh, that's really fun. But, uh,

01:28:25.290 --> 01:28:27.665
I recently watched this

01:28:28.225 --> 01:28:33.025
I guess it's like a it's like a miniseries documentary called The Dark Wizard

01:28:33.265 --> 01:28:35.585
about this guy, Dean Potter,

01:28:35.825 --> 01:28:39.745
who he was like Alex Honnold before Alex Honnold was Alex Honnold.

01:28:40.620 --> 01:28:42.220
And, uh,

01:28:42.220 --> 01:28:50.700
he just has this, like, very extreme personality where he's, like, free soloing everything, and then he's, like, you know, base jumping in in, like, a wingsuit and stuff like that. And it's sort of exploring

01:28:51.740 --> 01:28:54.380
his psychology and what happened to him. And

01:28:55.535 --> 01:29:03.215
I I don't know. I I kinda like stuff like that. Like, there's another one called 100 foot wave where it's, like, about people who are trying to like, big wave surfers.

01:29:03.215 --> 01:29:12.910
There's something about that that sort of, I guess, just reminds me of founders or whatever, but, um, the dark wizard, highly recommend. Is there a product you recently discovered that you really love? Codex.

01:29:13.390 --> 01:29:14.109
Okay.

01:29:14.590 --> 01:29:23.905
It's like it's the it's really good. It's really good. Do you have a favorite life motto that you often come back to in work or in life? Yes. I have several.

01:29:24.385 --> 01:29:31.185
The the, like, the core one that I wrote for myself in college was, um, do things worth writing about and write things worth reading.

01:29:31.665 --> 01:29:32.305
And,

01:29:32.545 --> 01:29:37.150
uh, and then there's there's this guy, Rob Rubea, who's, like, very, um,

01:29:38.350 --> 01:29:41.950
very popular in, like, you know, a the AI meditation,

01:29:42.030 --> 01:29:43.710
like, overlap discourse,

01:29:43.710 --> 01:29:45.390
which is also a big thing,

01:29:45.950 --> 01:29:49.390
um, and who I also I really like him. He's dead, but I think he's amazing.

01:29:50.095 --> 01:29:50.655
And

01:29:51.135 --> 01:29:56.655
I've listened to, like, so many of his talks, and there's, like, this one talk that he gives

01:29:57.375 --> 01:30:02.735
where it's just, like, one sentence, but he just talks about, like, when you're dealing with stuff that's hard,

01:30:03.350 --> 01:30:05.590
what you wanna do is be able to

01:30:06.230 --> 01:30:08.709
relate to it from a

01:30:09.190 --> 01:30:11.749
position of spaciousness and strength.

01:30:13.030 --> 01:30:13.749
And

01:30:14.550 --> 01:30:25.075
there is something, I think, really interesting and important in that. Like, a lot of the meditation discourse are just generally, like, how do you deal with hard things? It's, like, a little bit more of, the David Goggins. Like, you just gotta, like,

01:30:25.635 --> 01:30:28.610
just gotta, like, go for it kind of and, like, just

01:30:30.370 --> 01:30:32.450
and sometimes that sometimes that can work.

01:30:33.090 --> 01:30:33.650
And,

01:30:33.890 --> 01:30:40.290
also, I think sometimes when you're dealing with things so, for example, when you're dealing with I'm super afraid of, like, how AI is going to,

01:30:42.865 --> 01:30:44.465
you know, change my job.

01:30:45.105 --> 01:30:45.905
It is

01:30:46.225 --> 01:30:48.945
it has been very helpful for me to be like,

01:30:49.505 --> 01:30:50.305
am I

01:30:50.705 --> 01:31:00.760
coming at this from a vantage point of spaciousness and strength? And if not, can I, like, get there? Because it will be much more productive for me to deal with it from that place.

01:31:01.160 --> 01:31:02.040
And

01:31:02.040 --> 01:31:04.040
that has been very, very helpful for me.

01:31:04.440 --> 01:31:05.160
Wow.

01:31:05.720 --> 01:31:06.840
I love that.

01:31:07.960 --> 01:31:18.325
Well, our final question, just on the on the theme of this conversation, curious if there's just, like, an AI tool that you think is still kind of underrated that you're just, like, recently

01:31:18.965 --> 01:31:28.900
I mean people have just know about. I I I I'll say codex. I hate to say this, but I have to because, like, any anyone who knows me like, we were at this this conference recently

01:31:28.980 --> 01:31:34.580
and then private conference. I'm, like, telling, like, Boris and Kat from Cloud Code, like, you have to try Codex. And,

01:31:35.140 --> 01:31:40.445
um, it's it's just really good. And the things that you can do with it are so different,

01:31:41.565 --> 01:31:42.125
especially

01:31:42.365 --> 01:31:48.845
if you're using it with the in app browser to do things like do your emails or check check analytics or, like, anything like that.

01:31:50.310 --> 01:32:00.950
It it has completely transformed the way I work, and I would be doing you a disservice if I, like, was searching for something else because it is that good. Damn. That's wild.

01:32:01.765 --> 01:32:10.245
Do you feel like Anthropic can catch up and or is this just like, well, they get No. Yes. I think I think they can. I I like like I said, I think it's gonna be a horse race, and

01:32:11.125 --> 01:32:14.965
and different people will be ahead at at different at at at different times. But

01:32:15.910 --> 01:32:38.795
I think right now, OpenAI has, like, has has gotten back the mandate of heaven a little bit. It's been it was a rough couple a couple months, like, months or so, but I think they're back. Interesting. And and you'd switch if one became I would. I would. People people it's funny. People are like, oh, are you, like, sponsored by OpenAI? And I'm like, no. I just, like, talk about what I like. I was super loud about Cloud Code when that was the thing I really liked,

01:32:39.195 --> 01:32:52.750
and I'll just say what I like when when it happens. You know? And to your point, people like, there's a lot of value in using both for different things. There is. I I switch back and forth. Like, I I truly do still use Quad a lot. Yeah. Such a big market.

01:32:53.150 --> 01:32:57.710
Well, Dan, we did it. We we went through so much. I can't wait to revisit this in a year slash,

01:32:58.215 --> 01:33:01.255
get this out so people can start planning for this next year.

01:33:02.295 --> 01:33:18.820
Two final questions. Where can folks find you and every what should people know? And then how can listeners be useful to you? You can find me on x at Dan Shipper, s h I p p e r, and you can subscribe to every. Please subscribe to everyevery.toevery.to/subscribe.

01:33:18.900 --> 01:33:20.420
How can listeners be useful?

01:33:20.740 --> 01:33:26.580
You know, have fun with AI. Like, seriously, it's it's super fun. There's, like, a lot of it's not necessarily useful to me, but, like, it's

01:33:27.165 --> 01:33:42.760
it makes it I think it makes everything better when people put their hands in it and just, like, start figuring it out together rather than, like, arguing about it. And, um, so the most useful thing you can do is, like, find ways to use it well in your life and share it. Dan, thank you so much for being here. Thank you.

01:33:43.480 --> 01:33:48.600
Thank you so much for listening. If you found this valuable, you can subscribe to the show on Apple Podcasts,

01:33:48.600 --> 01:33:51.160
Spotify, or your favorite podcast app.

01:33:51.400 --> 01:34:02.685
Also, please consider giving us a rating or leaving a review as that really helps other listeners find the podcast. You can find all past episodes or learn more about the show at lennyspodcast.com.

01:34:02.925 --> 01:34:04.045
See you in the next episode.