> ## Content Index
> Fetch the complete content index at: https://www.ceo.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Distributional CEO Scott Clark: Focusing on AI confidence
- URL: https://www.ceo.com/episodes/distributional-co-founder-ceo-scott-clark/
- Published: 2025-02-18T11:00:07.000Z
- Updated: 2026-08-28T04:21:59.000Z
- Description: Distributional raised from a16z and Two Sigma, ran term sheets through to close, and became a company of one.
- Author: Clint Betts

# Being too early is the same thing as being wrong

Scott Clark can date the moment he knew.

His company was at an onsite and had just finished the best sales call in its history. The prospect had done most of the pitching. They had the problem Distributional solved; they wanted the solution; they wanted to pilot immediately. Clark went back to his room and did the math. Best case, no enterprise drag, every milestone hit, the pilot converts in six months at the quoted rate. He had sold enterprise software before and knew how many things have to go right for that. He would need a dozen deals like it. Every other deal in the pipeline was worse.

The best case was not good enough. He wrote later, in a [post-mortem on his own site](https://scottclark.io/blog/make-all-new-mistakes-faster?ref=ceo.com), that when the best isn't good enough, you have to play a different game.

That call led to the first reduction in force and the first pivot, about two years in. Distributional moved from pre-production AI testing to post-production behavioral analytics. Then roughly the same thing happened again. This spring, Clark ran a process to find the company a buyer and saw it through to the end: diligence, term sheets, and finally an offer to acqui-hire the entire team.

He took the offer for his people and not for the company. In July, he [announced](https://distributional.com/blog/distributional-is-now-talaria?ref=ceo.com) that Distributional had sunset its product and become Talaria Scientific. Same Delaware corporation, same investors, different mission. Most of the team landed together somewhere else. Clark kept the balance sheet and the job. Talaria today is a team of one plus contractors.

The CEO.com Podcast interview with Clark was recorded before any of that, when the company was raising and hiring, and he was explaining the thesis to anyone who would listen. Listening to it now is not embarrassing for him. It is closer to the opposite. He describes the mechanics of being wrong about a market with more precision than most founders manage afterward.

## Ten years of the same question

Clark came to AI through math. He took three bachelor's degrees at Oregon State in four years, in mathematics, physics, and computational physics, then went to Cornell for a PhD in applied mathematics and a master's in computer science as a Department of Energy Computational Science Graduate Fellow, with [practicums at Los Alamos and the Joint Genome Institute](https://scottclark.io/?ref=ceo.com). The thesis became open source at Yelp and then a company. SigOpt launched in 2014, went through Y Combinator, raised $17 million, and sold to Intel in 2020\. Clark stayed to run AI and high-performance computing software in Intel's supercomputing group.

He called SigOpt "the exact wrong way to start a company" on the CEO.com Podcast: some cool math, and then maybe a product, and then maybe a market. It found one anyway. The customers were teams with models expensive enough to be worth tuning. Streaming services. Hedge funds. Credit card companies. The intelligence community.

For 10 years, after every optimization, he got the same question.

"Great, you optimized it. How did you break it by optimizing it? Is this still robust? Did you make it more brittle?"

He heard it at SigOpt. He heard it at Intel. "A little bit of a slow learner after hearing this question for a decade, I was like, okay, maybe I'm actually solving the wrong problem here."

Distributional was the answer. "The thing that's preventing enterprise value isn't performance," he said. "It's confidence."

The technical problem is that AI breaks the assumption software testing rests on. Traditional tests expect the same input to produce the same output and flag anything else as a bug. AI returns different answers to the same question, drifts as its weights change, and turns small input changes into large output changes. Clark's view was that this instability is also where the value comes from, so the job is not to clamp it down. "AI testing is more about trying to determine whether you're getting unwanted behavior," he said. "First you need to detect change, triage change, and then ultimately resolve change."

The pitch worked. An 11-person founding team came out of Bloomberg, Google, Intel, Meta, Slack, Stripe, Uber, and Yelp. Andreessen Horowitz led an [$11 million seed](https://distributional.com/blog/we-raised-11m-for-better-ai-testing?ref=ceo.com) in December 2023, with Martin Casado taking a board seat, the same investor who had backed SigOpt and joined its board as he was starting at the firm. Ten months later, Two Sigma Ventures led a [$19 million Series A](https://www.businesswire.com/news/home/20241008537423/en/Distributional-Secures-%2419-Million-in-Series-A-Funding-and-Debuts-the-Industrys-Only-Enterprise-AI-Testing-Platform?ref=ceo.com). At SigOpt, Clark said, the Series A had taken four or five years.

Asked what he had learned from doing it twice, Clark laughed and said he was "making all new mistakes but 10 times faster." He meant it as a joke about velocity. It became the title of the post he wrote about the pivot.

## The name for what went wrong

Clark's account of the failure is specific enough to be useful to other founders, which is rare.

The market call was half right. He predicted enterprises would move from prototypes to production and that reliance would create demand for confidence. "Once it becomes real, there's risk, there's downside, there's maintenance," he said on the podcast, calling 2025 the year gen AI would grow up in the enterprise. Production happened. Demand for a statistical gate in front of it did not. In [his accounting](https://distributional.com/blog/distributional-is-now-talaria?ref=ceo.com), teams do not like tests that slow them down, least of all tests they did not write and cannot easily act on, and the market decided to ship without a gate at all. That last part surprised him most, given that a decade earlier the same enterprises had balked at gradient-boosted trees as too exotic. His verdict on timing: "being too early is the same as being wrong."

The deeper error was internal. Clark had assumed his SigOpt experience would let him skip stages. Instead, he ran the old playbook against a market that no longer looked like the one it was fit to. He hired and scaled ahead of product-market fit because that timing had worked before, and the scaling then demanded a roadmap, which demanded conviction he did not have, which made contrary evidence expensive to act on. A board member's line stuck with him: every company has to figure out product, sales, and operations, and "you need to do it sequentially and in that order."

He now has [a name for the trap](https://scottclark.io/blog/make-all-new-mistakes-faster?ref=ceo.com). Vision-market fit is resonance. Senior people with real budgets agree the problem is real and coming, and interest appears wherever you push. Product-market fit is pull, meaning buyers you never called show up and put you in their architecture diagrams unasked. From inside a single good meeting, the two are indistinguishable, and the difference only shows up when you stop pushing, which is the one experiment a scaling company never runs. Distributional had the first one repeatedly. It never got the second.

There is an irony Clark states himself: he spent three years building software to catch the gap between how people assume their AI behaves and how it actually behaves, and the assumption he tested least was the one holding up his own company.

## The fourth act

Talaria Scientific is a multi-agent harness for computational science. In his [founding essay](https://talariasci.com/blog/why-im-building-talaria?ref=ceo.com), Clark describes a researcher directing a system that absorbs the work nobody went to grad school for: wrangling data, building environments, running sweeps on HPC, keeping the record. He wants to do for scientists what agentic coding did for software engineers, and he is emphatic that it augments rather than replaces, calling it "a mech suit, not an android." It is in private beta, and he has committed to open sourcing it by NeurIPS 2026.

The strongest argument for it is the one he makes about himself. He is the customer. He runs his own research on partial differential equations and Bayesian optimization through the product, so when it is wrong, he knows the same afternoon rather than the next quarter.

The open questions are the ones a $30 million raise earns. How much of it is left, and how far it goes against a problem this size. Whether investors who funded AI testing twice are funding computational science or funding Clark. And whether the discipline he says he has learned, staying small enough that hypotheses are cheap to kill, survives contact with the first real customer who wants a roadmap.

On the podcast, before any of it, Clark described the CEO he wanted to be. Not the one working 100 hours a week and becoming indispensable, which he considers a failure mode and a bad cultural norm, but the one whose team can execute without him. "Figuring out how I can get out of the way," he said, "is literally every minute of the day."

The version of that job he ended up doing this spring was finding his people somewhere to land, and then staying behind.

Asked the question that closes every CEO.com Podcast conversation, who gave him a chance, Clark named the Y Combinator partners and alumni who read his application and told him a thesis is not a product, and Casado, who backed him at SigOpt and again at Distributional.

"Every founder is naive in some way," he said. "You need to be naive to be able to take that leap."

```chapters
00:00:12 Introduction and Guest Background
00:00:37 Founding and Selling SigOpt
00:02:12 Acquisition by Intel and Role There
00:02:43 Realization of the AI Confidence Problem
00:03:36 Founding Distributional
00:04:33 Balancing Creativity and Reliability in AI
00:05:59 Mainstreaming of AI and Future Predictions
00:10:13 Fundraising Experience and Lessons Learned
00:14:30 A Typical Day as CEO
00:16:18 Recruiting and Building a Team
00:19:09 Silicon Valley and Remote Work
00:21:09 Building Company Culture
00:23:24 Thoughts on Founder Mode Essay
00:29:20 Economic Outlook and Planning for 2025
00:34:09 Acknowledging Those Who Gave a Chance
```

```vtt
WEBVTT

00:00:12.448 --> 00:00:14.502
Scott,<00:00:12.852> thank<00:00:13.014> you<00:00:13.046> so<00:00:13.257> much<00:00:13.580> for<00:00:13.661> coming<00:00:13.920> on<00:00:14.065> the<00:00:14.146> show.

00:00:14.873 --> 00:00:20.532
I'm<00:00:15.051> super<00:00:15.617> impressed<00:00:17.153> with<00:00:17.379> your<00:00:17.638> background<00:00:18.447> and<00:00:18.608> what<00:00:18.835> you're<00:00:19.093> building<00:00:19.578> with<00:00:19.740> Distributional.

00:00:22.257 --> 00:00:26.820
And,<00:00:22.580> you<00:00:22.612> know,<00:00:22.822> you<00:00:22.918> did<00:00:23.096> SigOpt<00:00:24.047> prior<00:00:24.369> to<00:00:24.531> that,<00:00:25.014> which<00:00:25.240> is<00:00:25.643> really,<00:00:25.981> really<00:00:26.223> fascinating.

00:00:26.868 --> 00:00:29.689
And<00:00:27.013> you,<00:00:27.658> you<00:00:27.771> know,<00:00:28.077> had<00:00:28.238> that<00:00:28.384> acquired<00:00:29.109> in<00:00:29.383> 2020.

00:00:31.287 --> 00:00:35.967
Uh,<00:00:31.623> tell<00:00:31.864> us<00:00:32.328> your<00:00:32.505> background<00:00:33.210> and<00:00:33.370> how<00:00:33.611> you<00:00:33.707> became<00:00:34.188> the<00:00:34.332> co-founder<00:00:34.813> and<00:00:34.909> CEO<00:00:35.310> of<00:00:35.454> Distributional.

00:00:37.057 --> 00:00:37.377
Excellent.

00:00:37.441 --> 00:00:38.964
Well,<00:00:37.698> first<00:00:37.874> of<00:00:38.018> all,<00:00:38.115> thanks<00:00:38.291> for<00:00:38.435> having<00:00:38.675> me<00:00:38.820> on.

00:00:39.076 --> 00:00:40.022
I'm<00:00:39.172> really<00:00:39.461> honored<00:00:39.637> to<00:00:39.717> be<00:00:39.878> here.

00:00:40.102 --> 00:00:41.721
I'm<00:00:40.182> really<00:00:40.358> excited<00:00:40.759> to,<00:00:41.064> to<00:00:41.144> chat<00:00:41.416> today.

00:00:41.817 --> 00:00:49.703
Uh,<00:00:42.266> yeah,<00:00:42.682> I<00:00:42.843> can<00:00:42.987> tell<00:00:43.099> you<00:00:43.227> about<00:00:43.628> how,<00:00:43.885> uh,<00:00:44.269> I<00:00:44.446> can<00:00:44.606> sum<00:00:44.910> up<00:00:45.006> the<00:00:45.087> last<00:00:45.327> 20<00:00:45.632> years<00:00:45.824> of<00:00:45.872> my<00:00:46.032> life<00:00:46.305> in<00:00:46.385> a<00:00:46.449> few<00:00:46.593> minutes<00:00:46.930> for<00:00:47.010> you<00:00:47.154> here,<00:00:47.411> but<00:00:47.795> would<00:00:48.052> love<00:00:48.228> to<00:00:48.356> dive<00:00:48.533> deeper<00:00:48.837> wherever<00:00:49.094> it's<00:00:49.254> interesting.

00:00:50.360 --> 00:00:52.828
Uh,<00:00:50.536> so<00:00:50.680> as<00:00:50.760> you<00:00:50.857> mentioned,<00:00:51.402> uh,<00:00:51.594> SigOpt<00:00:51.979> was<00:00:52.139> my<00:00:52.283> first<00:00:52.524> company.

00:00:53.485 --> 00:01:08.785
Uh,<00:00:54.078> started<00:00:54.399> that<00:00:54.767> after,<00:00:55.409> uh,<00:00:55.745> completing<00:00:56.130> my<00:00:56.306> PhD,<00:00:56.867> and<00:00:56.947> it<00:00:57.027> was<00:00:57.172> kind<00:00:57.268> of<00:00:57.364> a,<00:00:57.604> a<00:00:57.733> culmination<00:00:58.374> of<00:00:58.470> some<00:00:58.694> of<00:00:58.855> that<00:00:59.400> that<00:00:59.656> work,<00:01:00.539> um,<00:01:01.116> and<00:01:01.902> focused<00:01:02.239> mostly<00:01:02.721> on<00:01:02.865> how<00:01:03.042> do<00:01:03.106> you<00:01:03.346> optimize<00:01:04.068> AI<00:01:04.261> systems<00:01:05.031> back<00:01:05.208> when<00:01:05.368> people<00:01:05.689> were<00:01:05.769> actually<00:01:06.234> building<00:01:06.491> their<00:01:06.571> own<00:01:06.812> models<00:01:07.213> instead<00:01:07.550> of<00:01:07.630> using<00:01:07.935> foundational<00:01:08.432> models.

00:01:08.898 --> 00:01:24.204
So<00:01:09.924> using<00:01:10.181> a<00:01:10.261> combination<00:01:10.967> of<00:01:11.529> Bayesian<00:01:11.850> optimization<00:01:12.524> methods—<00:01:13.294> some<00:01:13.406> of<00:01:13.454> these<00:01:13.695> were<00:01:13.791> very<00:01:14.096> cutting<00:01:14.417> edge,<00:01:14.754> we<00:01:14.914> published<00:01:15.412> a<00:01:15.476> lot<00:01:15.716> at<00:01:15.813> NeurIPS<00:01:16.198> and<00:01:16.278> ICML,<00:01:17.064> and<00:01:17.562> was<00:01:17.722> built<00:01:17.898> upon<00:01:18.203> some<00:01:18.300> of<00:01:18.508> my<00:01:18.669> PhD<00:01:19.086> thesis<00:01:19.471> work—<00:01:20.369> to<00:01:20.449> help<00:01:20.674> people<00:01:21.332> do<00:01:21.637> hyperparameter<00:01:22.279> optimization,<00:01:23.097> neural<00:01:23.337> architecture<00:01:23.915> search.

00:01:24.300 --> 00:01:31.127
Basically,<00:01:25.006> how<00:01:25.183> do<00:01:25.263> you<00:01:25.664> tune<00:01:26.049> all<00:01:26.145> the<00:01:26.225> different<00:01:26.466> knobs<00:01:26.787> and<00:01:26.867> levers<00:01:27.268> that<00:01:27.429> go<00:01:27.669> into<00:01:28.700> how<00:01:28.895> the<00:01:28.993> model<00:01:29.368> is<00:01:29.514> actually<00:01:29.922> trained<00:01:30.508> and<00:01:30.671> learns.

00:01:33.440 --> 00:01:42.214
This<00:01:33.520> is,<00:01:33.569> of<00:01:33.699> course,<00:01:34.184> the<00:01:34.282> exact<00:01:35.010> wrong<00:01:35.253> way<00:01:35.415> to<00:01:35.480> start<00:01:35.739> a<00:01:35.804> company<00:01:36.370> where<00:01:36.694> you<00:01:37.018> have<00:01:37.358> something,<00:01:38.184> some<00:01:38.475> cool<00:01:38.718> math,<00:01:39.139> and<00:01:39.220> you're<00:01:39.317> like,<00:01:39.544> maybe<00:01:39.852> I<00:01:39.868> can<00:01:40.030> build<00:01:40.192> a<00:01:40.273> product<00:01:40.694> and<00:01:40.758> maybe<00:01:41.066> I<00:01:41.147> can<00:01:41.309> find<00:01:41.827> a<00:01:41.860> market.

00:01:43.406 --> 00:01:56.279
But<00:01:43.503> we<00:01:43.648> were<00:01:43.745> able<00:01:44.969> to<00:01:45.098> make<00:01:45.598> a<00:01:45.743> big<00:01:45.920> dent<00:01:46.806> in<00:01:46.887> the<00:01:46.967> space<00:01:47.676> and<00:01:47.870> ended<00:01:48.111> up<00:01:48.240> working<00:01:48.498> with<00:01:48.675> a<00:01:48.740> lot<00:01:48.917> of<00:01:49.046> what<00:01:49.143> I<00:01:49.223> would<00:01:49.368> call<00:01:50.577> very<00:01:50.754> sophisticated<00:01:51.640> modeling<00:01:52.026> teams,<00:01:52.526> people<00:01:52.784> who<00:01:52.928> had<00:01:53.573> big<00:01:53.895> expensive<00:01:54.394> models<00:01:54.861> that<00:01:55.006> they<00:01:55.055> needed<00:01:55.345> to<00:01:55.425> make<00:01:55.602> a<00:01:55.731> little<00:01:55.973> bit<00:01:56.150> better.

00:01:56.942 --> 00:02:10.141
Working<00:01:57.247> with<00:01:57.407> like<00:01:57.648> large<00:01:57.840> streaming<00:01:58.321> services<00:01:59.107> and<00:01:59.251> trillions<00:01:59.749> of<00:01:59.829> dollars<00:02:00.150> worth<00:02:00.310> of<00:02:00.454> hedge<00:02:00.711> funds<00:02:01.192> and<00:02:01.256> the<00:02:01.336> US<00:02:01.593> intelligence<00:02:02.138> community<00:02:02.715> and<00:02:02.876> credit<00:02:03.181> card<00:02:03.437> firms<00:02:03.838> and<00:02:04.560> basically<00:02:04.961> anyone<00:02:05.426> who<00:02:05.522> had<00:02:05.746> something<00:02:06.067> that<00:02:06.163> was<00:02:06.805> important<00:02:07.270> enough<00:02:07.671> that<00:02:07.783> it<00:02:07.847> needed<00:02:08.168> to,<00:02:08.328> to<00:02:08.569> be,<00:02:09.034> uh,<00:02:09.130> fully<00:02:09.531> optimized.

00:02:10.878 --> 00:02:21.880
Um,<00:02:12.418> I<00:02:12.498> ended<00:02:12.723> up<00:02:12.915> selling<00:02:13.204> that<00:02:13.300> company,<00:02:13.605> as<00:02:13.685> you<00:02:13.717> mentioned,<00:02:14.246> uh,<00:02:14.342> before,<00:02:15.128> uh,<00:02:15.224> to<00:02:15.385> Intel<00:02:15.722> in<00:02:15.866> 2020,<00:02:17.213> um,<00:02:17.614> where<00:02:17.710> I<00:02:17.774> ended<00:02:18.015> up<00:02:18.175> leading<00:02:18.512> the<00:02:18.736> AI<00:02:19.073> and<00:02:19.218> HPC<00:02:19.875> division<00:02:20.517> of<00:02:20.613> their<00:02:20.757> supercomputing<00:02:21.543> group.

00:02:22.104 --> 00:02:28.991
Um,<00:02:23.324> Again,<00:02:23.886> very<00:02:24.528> big<00:02:24.849> expensive<00:02:25.331> models<00:02:26.278> now<00:02:26.519> being<00:02:26.840> deployed<00:02:27.257> on<00:02:27.337> very<00:02:27.578> big<00:02:27.803> expensive<00:02:28.220> supercomputers.

00:02:30.227 --> 00:02:46.121
And<00:02:30.307> basically<00:02:30.789> over<00:02:30.966> the<00:02:31.094> course<00:02:31.431> of<00:02:31.495> those<00:02:31.832> 10<00:02:32.154> years<00:02:32.860> that<00:02:33.020> I<00:02:33.101> was<00:02:33.502> taking<00:02:34.080> that<00:02:34.257> thesis<00:02:35.043> and<00:02:35.124> helping<00:02:35.525> firms<00:02:35.990> optimize<00:02:36.552> AI<00:02:36.793> models,<00:02:37.997> I<00:02:38.094> slowly<00:02:38.655> came<00:02:38.976> to<00:02:39.137> the<00:02:39.217> realization<00:02:40.341> that<00:02:40.807> the<00:02:40.887> thing<00:02:41.064> that<00:02:41.160> was<00:02:41.449> really<00:02:41.770> preventing<00:02:42.187> people<00:02:42.492> from<00:02:42.653> getting<00:02:42.846> a<00:02:42.910> ton<00:02:43.167> of<00:02:43.231> value<00:02:43.793> out<00:02:43.873> of<00:02:44.017> AI<00:02:44.997> wasn't<00:02:45.543> performance.

00:02:46.586 --> 00:02:49.042
It<00:02:46.682> wasn't<00:02:47.020> about<00:02:47.469> how<00:02:47.565> do<00:02:47.710> I<00:02:47.886> over-optimize<00:02:48.625> something?

00:02:49.155 --> 00:02:52.205
How<00:02:49.187> do<00:02:49.331> I<00:02:49.556> overfit<00:02:50.070> some<00:02:50.375> eval<00:02:50.616> metric<00:02:51.081> by<00:02:51.178> another<00:02:51.418> half<00:02:51.675> a<00:02:51.739> percent?

00:02:52.798 --> 00:02:57.959
It<00:02:52.862> was<00:02:53.023> really<00:02:53.343> fear<00:02:54.080> of<00:02:54.225> the<00:02:54.305> downside,<00:02:55.266> not<00:02:55.587> trying<00:02:55.827> to<00:02:55.891> squeeze<00:02:56.388> that<00:02:56.628> last<00:02:56.949> little<00:02:57.109> bit<00:02:57.286> of<00:02:57.430> the<00:02:57.574> upside.

00:02:58.696 --> 00:03:03.199
Um,<00:02:59.193> and<00:02:59.257> so<00:02:59.433> time<00:02:59.770> and<00:02:59.834> time<00:03:00.090> again,<00:03:00.475> we'd<00:03:00.635> get<00:03:00.939> questions<00:03:01.292> about<00:03:01.581> like,<00:03:01.917> great,<00:03:02.238> you<00:03:02.462> optimized<00:03:03.039> it.

00:03:03.904 --> 00:03:05.747
How<00:03:04.000> did<00:03:04.145> you<00:03:04.321> break<00:03:04.577> it<00:03:04.802> by<00:03:05.026> optimizing<00:03:05.683> it?

00:03:06.180 --> 00:03:07.334
Is<00:03:06.308> this<00:03:06.469> still<00:03:06.709> robust?

00:03:07.526 --> 00:03:08.744
Like,<00:03:07.703> did<00:03:07.847> you<00:03:07.927> make<00:03:08.103> it<00:03:08.231> more<00:03:08.408> brittle?

00:03:08.953 --> 00:03:09.818
All<00:03:09.049> this<00:03:09.129> sorts<00:03:09.449> of<00:03:09.529> things.

00:03:09.930 --> 00:03:12.222
And<00:03:10.010> I<00:03:10.090> saw<00:03:10.331> this<00:03:11.389> in<00:03:11.613> the<00:03:11.693> SIGOP<00:03:12.014> days.

00:03:12.270 --> 00:03:13.328
I<00:03:12.318> saw<00:03:12.478> it<00:03:12.574> in<00:03:12.639> the<00:03:12.735> Intel<00:03:13.055> days.

00:03:13.376 --> 00:03:14.994
I<00:03:13.392> saw<00:03:13.937> it<00:03:14.097> just<00:03:14.257> time<00:03:14.498> and<00:03:14.546> time<00:03:14.754> again.

00:03:15.058 --> 00:03:19.482
And<00:03:15.203> so<00:03:15.459> a<00:03:16.020> little<00:03:16.180> bit<00:03:16.341> of<00:03:16.421> a<00:03:16.437> slow<00:03:16.757> learner<00:03:17.142> after<00:03:17.687> hearing<00:03:18.023> this<00:03:18.184> question<00:03:18.584> for<00:03:18.841> a<00:03:18.985> decade.

00:03:19.770 --> 00:03:22.831
Um,<00:03:19.946> I<00:03:19.995> was<00:03:20.107> like,<00:03:20.347> okay,<00:03:20.652> maybe<00:03:20.844> I'm<00:03:21.052> actually<00:03:21.325> solving<00:03:21.789> the<00:03:21.870> wrong<00:03:22.110> problem<00:03:22.591> here.

00:03:23.393 --> 00:03:27.698
And<00:03:23.473> that's<00:03:23.698> really<00:03:24.180> where<00:03:24.999> the<00:03:25.079> idea<00:03:25.401> for<00:03:25.561> distributional<00:03:26.686> came<00:03:26.927> into<00:03:27.489> being.

00:03:27.906 --> 00:03:32.195
It<00:03:27.955> was<00:03:28.774> the<00:03:28.870> thing<00:03:29.175> that's<00:03:29.336> preventing<00:03:29.738> enterprise<00:03:30.220> value<00:03:30.605> isn't<00:03:30.782> performance,<00:03:31.440> it's<00:03:31.585> confidence.

00:03:32.934 --> 00:03:36.500
And<00:03:33.577> confidence<00:03:34.059> is<00:03:34.155> about<00:03:34.476> consistent<00:03:35.279> and<00:03:35.360> reliable<00:03:35.922> behavior.

00:03:36.789 --> 00:03:38.283
And<00:03:36.869> performance<00:03:37.367> is<00:03:37.448> an<00:03:37.592> aspect<00:03:37.994> of<00:03:38.154> that.

00:03:38.636 --> 00:03:42.395
You<00:03:38.733> want<00:03:38.974> consistent<00:03:39.440> performance,<00:03:40.018> you<00:03:40.098> want<00:03:40.259> reliable<00:03:40.821> performance,<00:03:41.383> you<00:03:41.463> want<00:03:41.688> good<00:03:41.865> performance.

00:03:43.471 --> 00:03:46.363
But<00:03:44.017> behavior<00:03:44.515> itself<00:03:45.174> is<00:03:45.318> more<00:03:45.559> than<00:03:45.656> just<00:03:45.800> performance.

00:03:46.539 --> 00:03:48.386
It's<00:03:46.700> not<00:03:46.909> just<00:03:47.407> the<00:03:47.712> end<00:03:47.969> result.

00:03:48.692 --> 00:03:49.628
But<00:03:48.837> it's<00:03:48.999> how<00:03:49.241> you<00:03:49.338> get<00:03:49.499> there.

00:03:49.725 --> 00:03:50.516
It's<00:03:49.967> what<00:03:50.145> you<00:03:50.371> do.

00:03:51.178 --> 00:03:58.501
And<00:03:51.275> sometimes<00:03:51.904> behavior<00:03:52.486> can<00:03:52.647> give<00:03:52.792> you<00:03:52.905> information<00:03:53.519> about<00:03:53.842> how<00:03:53.939> performance<00:03:54.568> might<00:03:54.746> be<00:03:54.827> shifting<00:03:55.408> or<00:03:55.537> changing<00:03:56.425> well<00:03:56.667> ahead<00:03:56.941> of<00:03:57.071> that<00:03:57.232> performance<00:03:57.653> actually<00:03:57.985> degrading.

00:03:59.980 --> 00:04:12.786
And<00:04:00.044> so<00:04:00.560> the<00:04:00.656> idea<00:04:00.979> was,<00:04:01.204> can<00:04:01.301> we<00:04:01.462> take<00:04:01.849> a<00:04:01.929> lot<00:04:02.171> of<00:04:02.413> what<00:04:02.574> we<00:04:02.735> learned<00:04:03.073> basically<00:04:03.718> by<00:04:04.024> helping<00:04:04.363> people<00:04:04.749> build<00:04:05.571> evals,<00:04:06.038> helping<00:04:06.296> people<00:04:06.538> do<00:04:06.764> optimization,<00:04:07.505> helping<00:04:07.827> people<00:04:08.133> build<00:04:08.311> more<00:04:08.440> performant<00:04:08.874> models<00:04:10.001> to<00:04:10.098> now<00:04:10.404> help<00:04:10.629> them<00:04:10.726> build<00:04:10.967> more<00:04:11.434> confidence<00:04:12.175> in<00:04:12.255> their<00:04:12.416> models.

00:04:13.769 --> 00:04:18.568
And<00:04:14.735> that's<00:04:14.992> kind<00:04:15.153> of<00:04:15.185> the<00:04:15.298> crux<00:04:15.862> of<00:04:16.200> Distributional<00:04:16.844> is<00:04:17.230> doing<00:04:17.472> that<00:04:17.713> through<00:04:18.100> an<00:04:18.136> enterprise<00:04:18.316> testing<00:04:18.424> platform.

00:04:20.062 --> 00:04:32.068
Using<00:04:20.382> some<00:04:20.703> of<00:04:20.767> the<00:04:20.799> same<00:04:21.248> tricks<00:04:21.585> and<00:04:21.665> techniques<00:04:22.210> that<00:04:22.370> you<00:04:22.466> would<00:04:22.627> use<00:04:22.931> to<00:04:23.027> gain<00:04:23.252> confidence<00:04:23.749> with<00:04:23.909> traditional<00:04:24.326> software,<00:04:25.432> but<00:04:25.576> applying<00:04:26.137> it<00:04:26.313> in<00:04:26.394> this<00:04:26.618> more<00:04:27.436> non-deterministic,<00:04:28.461> non-stationary,<00:04:29.984> kind<00:04:30.097> of<00:04:30.305> extremely<00:04:30.866> complex<00:04:31.363> world<00:04:31.619> of<00:04:31.748> AI.

00:04:33.206 --> 00:04:33.383
Yeah.

00:04:33.431 --> 00:04:35.002
How<00:04:33.527> do<00:04:33.671> you<00:04:33.767> do<00:04:33.944> it<00:04:34.072> in<00:04:34.168> the<00:04:34.312> world<00:04:34.569> of<00:04:34.713> AI?

00:04:35.290 --> 00:04:42.039
So<00:04:35.531> where,<00:04:35.932> like,<00:04:36.108> how<00:04:36.188> do<00:04:36.268> you<00:04:36.332> find<00:04:36.557> that<00:04:36.733> balance,<00:04:37.198> I<00:04:37.278> guess,<00:04:38.320> is<00:04:38.416> my<00:04:38.577> question,<00:04:38.977> because<00:04:39.699> part<00:04:39.843> of<00:04:40.003> AI<00:04:40.324> is<00:04:40.404> like,<00:04:40.837> it's<00:04:41.125> meant<00:04:41.382> to<00:04:41.526> be<00:04:41.686> creative.

00:04:42.119 --> 00:04:44.395
It's<00:04:42.328> meant<00:04:42.504> to<00:04:42.648> kind<00:04:42.824> of<00:04:42.905> come<00:04:43.081> up<00:04:43.289> with<00:04:43.402> stuff<00:04:43.770> on<00:04:43.866> its<00:04:44.091> own.

00:04:44.908 --> 00:04:46.351
And<00:04:45.053> so<00:04:45.213> how<00:04:45.309> do<00:04:45.453> you<00:04:45.630> find<00:04:46.031> that<00:04:46.079> balance?

00:04:46.560 --> 00:04:47.858
Balance<00:04:46.992> between<00:04:47.457> the<00:04:47.553> two?

00:04:48.419 --> 00:04:56.915
Yeah,<00:04:48.675> that's<00:04:48.852> a<00:04:48.916> great<00:04:49.140> question,<00:04:49.942> uh,<00:04:50.198> because<00:04:51.705> you<00:04:51.737> don't<00:04:51.946> want<00:04:52.090> to<00:04:52.122> necessarily<00:04:52.987> put<00:04:53.148> it<00:04:53.244> in<00:04:53.308> a<00:04:53.404> box<00:04:53.805> that's<00:04:54.045> so<00:04:54.270> constrained<00:04:54.847> that<00:04:54.943> it<00:04:55.007> basically<00:04:55.488> looks<00:04:55.729> like<00:04:55.985> traditional<00:04:56.386> software.

00:04:57.476 --> 00:05:06.388
Um,<00:04:58.117> but<00:04:58.213> it's<00:04:58.357> a<00:04:58.374> double-edged<00:04:58.758> sword,<00:04:59.399> uh,<00:04:59.896> the—<00:05:00.137> that<00:05:00.778> non-determinism,<00:05:01.579> getting<00:05:01.804> a<00:05:01.884> different<00:05:02.205> answer<00:05:02.926> to<00:05:03.006> the<00:05:03.086> same<00:05:03.343> question,<00:05:03.984> could<00:05:04.208> actually<00:05:04.545> lead<00:05:04.769> to<00:05:04.866> really<00:05:05.266> interesting<00:05:05.587> and<00:05:05.667> emergent<00:05:05.972> behavior.

00:05:07.270 --> 00:05:13.169
The<00:05:07.350> non-stationarity,<00:05:08.296> the<00:05:08.376> fact<00:05:08.617> that<00:05:08.713> these<00:05:08.873> models<00:05:09.274> are<00:05:09.354> evolving<00:05:09.915> and<00:05:09.979> changing<00:05:10.332> underneath<00:05:10.716> you,<00:05:10.861> actually<00:05:11.742> sometimes<00:05:12.303> helps<00:05:12.464> them<00:05:12.640> get<00:05:12.881> better.

00:05:14.083 --> 00:05:28.376
And<00:05:14.227> so<00:05:14.387> I<00:05:14.467> think<00:05:14.723> the<00:05:14.804> trade-off<00:05:15.364> is<00:05:16.486> being<00:05:16.727> able<00:05:16.967> to<00:05:17.432> understand<00:05:18.409> what<00:05:18.553> that<00:05:19.210> behavior<00:05:19.755> is,<00:05:20.172> what<00:05:20.508> it<00:05:20.733> looks<00:05:20.957> like<00:05:21.149> in<00:05:21.277> a<00:05:21.293> holistic<00:05:21.918> way,<00:05:22.479> being<00:05:22.671> able<00:05:22.800> to<00:05:22.896> detect<00:05:23.377> when<00:05:23.537> that<00:05:23.697> behavior<00:05:24.178> changes,<00:05:25.380> and<00:05:25.460> then<00:05:25.540> being<00:05:25.780> able<00:05:25.892> to<00:05:26.085> quickly<00:05:26.421> triage,<00:05:26.966> like,<00:05:27.847> was<00:05:27.943> this<00:05:28.104> good?

00:05:28.745 --> 00:05:29.514
Was<00:05:28.841> this<00:05:29.065> bad?

00:05:29.866 --> 00:05:31.276
Do<00:05:30.011> I<00:05:30.107> care<00:05:30.331> about<00:05:30.587> this<00:05:30.908> or<00:05:30.940> not?

00:05:31.693 --> 00:05:40.683
But<00:05:31.789> it<00:05:31.853> ends<00:05:32.014> up<00:05:32.094> being<00:05:32.270> fundamentally<00:05:32.991> different<00:05:33.472> than<00:05:33.856> traditional<00:05:34.369> software<00:05:35.555> where<00:05:36.340> you<00:05:36.901> want<00:05:37.013> it<00:05:37.141> to<00:05:37.221> do<00:05:37.318> the<00:05:37.462> exact<00:05:37.782> same<00:05:38.023> thing,<00:05:38.439> and<00:05:38.583> whenever<00:05:38.840> it<00:05:39.000> doesn't<00:05:39.305> do<00:05:39.465> that,<00:05:40.026> you<00:05:40.106> found<00:05:40.298> a<00:05:40.442> bug.

00:05:41.725 --> 00:05:47.479
With<00:05:42.029> AI<00:05:42.286> testing,<00:05:42.847> it's<00:05:42.991> more<00:05:43.151> about<00:05:43.808> you're<00:05:43.969> trying<00:05:44.434> to<00:05:45.011> determine<00:05:45.636> whether<00:05:45.876> you're<00:05:46.053> getting<00:05:46.437> unwanted<00:05:46.918> behavior.

00:05:47.896 --> 00:05:52.400
So<00:05:48.056> first<00:05:48.361> you<00:05:48.393> need<00:05:48.617> to<00:05:48.697> detect<00:05:49.098> change,<00:05:49.964> triage<00:05:50.444> change,<00:05:50.941> and<00:05:51.021> then<00:05:51.182> ultimately<00:05:51.582> resolve<00:05:51.983> change.

00:05:52.785 --> 00:05:58.122
But<00:05:52.945> it<00:05:53.089> is<00:05:53.249> a<00:05:53.265> little<00:05:53.506> bit<00:05:53.826> more<00:05:54.067> of<00:05:54.227> an<00:05:54.612> unsupervised<00:05:56.551> sense<00:05:56.872> of<00:05:57.513> doing<00:05:57.754> testing.

00:05:59.372 --> 00:06:04.037
OpenAI's<00:05:59.982> really<00:06:00.222> kind<00:06:00.462> of<00:06:00.799> made<00:06:02.226> AI<00:06:03.219> mainstream.

00:06:04.101 --> 00:06:11.954
Everybody's<00:06:04.470> talked<00:06:04.710> about<00:06:04.822> it<00:06:04.870> ever<00:06:04.967> since<00:06:05.207> ChatGPT,<00:06:06.008> and<00:06:06.169> they're<00:06:06.249> not<00:06:06.425> the<00:06:06.553> only<00:06:06.730> ones,<00:06:07.034> obviously,<00:06:07.451> but<00:06:08.610> ChatGPT<00:06:08.968> really<00:06:09.278> kind<00:06:09.442> of<00:06:09.474> put<00:06:09.686> this<00:06:09.866> on<00:06:10.094> everyone's<00:06:10.584> radar<00:06:10.910> and<00:06:10.991> had<00:06:11.252> everybody<00:06:11.579> freaking<00:06:11.889> out.

00:06:12.133 --> 00:06:20.075
Obviously<00:06:12.541> AI<00:06:12.623> has<00:06:12.884> been<00:06:13.031> around<00:06:13.292> since<00:06:13.520> like<00:06:13.863> the<00:06:13.895> '50s,<00:06:15.720> but<00:06:15.988> you<00:06:16.089> know,<00:06:18.140> now<00:06:18.414> it's<00:06:18.575> like,<00:06:18.897> hey,<00:06:19.059> what<00:06:19.204> are<00:06:19.236> we<00:06:19.365> going<00:06:19.446> to<00:06:19.542> do<00:06:19.639> about<00:06:19.849> this<00:06:20.010> thing?

00:06:20.171 --> 00:06:25.767
So,<00:06:20.832> and<00:06:20.897> I<00:06:20.929> know<00:06:21.090> that<00:06:21.316> this<00:06:21.542> question<00:06:22.026> is<00:06:22.767> probably<00:06:23.251> unanswerable,<00:06:23.977> but<00:06:24.138> where<00:06:24.299> do<00:06:24.364> you<00:06:24.525> think<00:06:24.702> we're<00:06:24.864> going<00:06:25.412> with<00:06:25.589> AI?

00:06:25.831 --> 00:06:27.041
Like,<00:06:26.073> where<00:06:26.218> does<00:06:26.379> this<00:06:26.621> all<00:06:26.783> go?

00:06:27.520 --> 00:06:31.298
Like<00:06:27.618> over<00:06:27.716> the<00:06:27.798> next<00:06:28.043> 5<00:06:28.305> to<00:06:28.452> 10<00:06:28.844> years,<00:06:29.499> what<00:06:30.104> are<00:06:30.398> you<00:06:30.644> planning<00:06:31.053> for?

00:06:31.396 --> 00:06:32.018
What<00:06:31.560> are<00:06:31.625> you<00:06:31.805> expecting?

00:06:33.463 --> 00:06:39.972
I'm<00:06:33.609> excited<00:06:34.093> about<00:06:34.820> just<00:06:35.951> the<00:06:36.047> surface<00:06:36.516> area<00:06:36.774> of<00:06:36.855> types<00:06:37.194> of<00:06:37.259> problems<00:06:37.727> that<00:06:37.824> it<00:06:37.905> can<00:06:38.050> solve<00:06:38.406> just<00:06:38.793> continuing<00:06:39.423> to<00:06:39.536> increase.

00:06:40.085 --> 00:06:48.574
And<00:06:40.166> so<00:06:40.715> you<00:06:40.796> go<00:06:40.974> back<00:06:41.377> to<00:06:42.023> the<00:06:42.184> '50s<00:06:42.668> and<00:06:42.749> that<00:06:43.395> original<00:06:45.992> Summer<00:06:47.041> Symposium<00:06:47.686> or<00:06:47.815> whatever<00:06:48.074> it<00:06:48.154> may<00:06:48.316> be.

00:06:48.864 --> 00:06:49.219
Yeah.

00:06:50.171 --> 00:06:52.964
And<00:06:50.316> I'd<00:06:50.478> say<00:06:51.058> they<00:06:51.220> had<00:06:51.462> a<00:06:51.543> grand<00:06:52.011> vision<00:06:52.496> back<00:06:52.754> then.

00:06:53.384 --> 00:06:59.620
And<00:06:53.901> it's<00:06:54.111> really<00:06:54.289> interesting<00:06:54.596> to<00:06:54.677> go<00:06:54.854> back<00:06:55.113> and<00:06:55.242> look<00:06:55.355> at<00:06:55.517> some<00:06:55.759> of<00:06:55.824> the<00:06:55.985> things<00:06:56.292> where<00:06:56.405> it<00:06:56.454> was<00:06:56.615> like,<00:06:57.181> man,<00:06:57.359> if<00:06:57.439> we<00:06:57.601> can<00:06:57.762> do<00:06:57.924> spell<00:06:58.392> check,<00:06:58.893> like,<00:06:59.136> that's<00:06:59.443> AI.

00:07:00.333 --> 00:07:03.320
And<00:07:00.413> like,<00:07:00.590> that<00:07:00.798> was<00:07:01.120> largely<00:07:01.473> solved<00:07:01.939> in<00:07:02.003> the<00:07:02.035> '90s<00:07:02.565> and<00:07:02.661> things<00:07:02.886> like<00:07:03.063> that.

00:07:03.384 --> 00:07:03.722
And<00:07:03.529> it—.

00:07:03.866 --> 00:07:03.946
The—.

00:07:04.653 --> 00:07:08.058
What<00:07:04.894> AI<00:07:05.312> is,<00:07:05.553> is<00:07:05.697> always<00:07:06.018> like,<00:07:06.259> what,<00:07:06.436> what<00:07:06.580> can<00:07:06.693> we<00:07:06.773> not<00:07:07.143> do<00:07:07.255> super<00:07:07.641> well<00:07:07.817> today?

00:07:08.604 --> 00:07:13.872
But<00:07:08.701> one<00:07:08.909> thing<00:07:09.070> that<00:07:09.166> has<00:07:09.488> been<00:07:09.648> completely<00:07:10.130> true,<00:07:10.371> I<00:07:10.451> think,<00:07:10.708> over<00:07:10.869> the<00:07:10.997> years<00:07:11.270> is<00:07:11.415> the<00:07:11.495> surface<00:07:12.041> area<00:07:12.218> of<00:07:12.298> the<00:07:12.379> types<00:07:12.780> of<00:07:12.861> problems<00:07:13.262> that<00:07:13.407> it<00:07:13.439> can<00:07:13.599> solve.

00:07:14.467 --> 00:07:20.795
So<00:07:15.430> can<00:07:15.607> it<00:07:15.671> do<00:07:15.816> spell<00:07:16.073> check<00:07:16.330> is,<00:07:16.651> is<00:07:16.715> an<00:07:16.860> example,<00:07:17.342> but<00:07:17.502> like,<00:07:17.840> can<00:07:18.016> it<00:07:18.145> make<00:07:18.321> a<00:07:18.386> prediction,<00:07:18.803> a<00:07:18.884> binary<00:07:19.430> prediction<00:07:19.912> of<00:07:20.008> just<00:07:20.233> true/false?

00:07:20.907 --> 00:07:25.581
Okay,<00:07:21.213> can<00:07:21.389> it,<00:07:21.919> can<00:07:22.016> it<00:07:22.176> do<00:07:22.497> more,<00:07:23.365> um,<00:07:24.505> continuous<00:07:25.051> predictions?

00:07:25.629 --> 00:07:27.927
Can<00:07:25.854> it<00:07:26.834> be<00:07:26.995> able<00:07:27.107> to<00:07:27.172> detect<00:07:27.702> things.

00:07:27.975 --> 00:07:43.767
And<00:07:29.148> as<00:07:29.308> we<00:07:29.341> start<00:07:29.646> to<00:07:29.806> kind<00:07:29.983> of<00:07:30.208> layer<00:07:30.594> on<00:07:30.835> these<00:07:30.947> systems,<00:07:31.413> and<00:07:31.574> this<00:07:31.734> is<00:07:31.815> where<00:07:31.975> I<00:07:32.072> think<00:07:32.361> some—<00:07:32.682> what's<00:07:32.939> now<00:07:33.742> being<00:07:34.144> coined<00:07:34.385> as<00:07:34.465> like<00:07:34.690> agentic<00:07:35.124> systems<00:07:35.606> of<00:07:35.751> like<00:07:36.072> compositions<00:07:36.715> of<00:07:36.811> systems<00:07:37.839> and<00:07:37.984> things<00:07:38.241> like<00:07:38.482> that,<00:07:38.884> like<00:07:39.124> the<00:07:39.189> surface<00:07:39.606> area<00:07:39.847> is<00:07:40.313> only<00:07:40.554> expanding<00:07:41.293> because<00:07:42.000> the<00:07:42.402> foundation<00:07:42.980> is<00:07:43.125> getting<00:07:43.318> stronger.

00:07:43.848 --> 00:07:45.872
We're<00:07:43.944> having<00:07:44.266> more<00:07:44.506> of<00:07:44.587> these<00:07:44.747> individual<00:07:45.229> components.

00:07:46.258 --> 00:07:47.334
They<00:07:46.418> can<00:07:46.579> do<00:07:46.675> more<00:07:46.916> things.

00:07:47.398 --> 00:07:49.776
They're<00:07:47.543> getting<00:07:47.720> more<00:07:47.880> performant<00:07:48.507> at<00:07:48.603> each<00:07:48.780> of<00:07:48.844> the<00:07:48.925> things<00:07:49.182> that<00:07:49.326> they're<00:07:49.503> doing.

00:07:50.611 --> 00:08:14.094
And<00:07:50.676> then<00:07:50.756> when<00:07:50.836> you<00:07:50.933> combine<00:07:51.399> all<00:07:51.559> of<00:07:51.640> those<00:07:51.816> things<00:07:52.057> together,<00:07:52.604> you<00:07:52.700> get<00:07:53.326> self-driving<00:07:53.903> cars,<00:07:54.688> you<00:07:54.784> get<00:07:56.707> personal<00:07:57.044> assistants,<00:07:57.669> you<00:07:57.813> get<00:07:57.973> all<00:07:58.165> of<00:07:58.229> these<00:07:59.992> things<00:08:00.232> that<00:08:01.739> are<00:08:01.995> built<00:08:02.251> upon<00:08:03.117> technology<00:08:03.582> that's<00:08:03.838> been<00:08:03.998> around<00:08:04.383> a<00:08:04.463> long<00:08:04.703> time,<00:08:04.960> but<00:08:05.120> combined<00:08:05.601> in<00:08:05.841> interesting<00:08:06.386> ways<00:08:06.947> and<00:08:07.267> made<00:08:07.524> better<00:08:08.485> through<00:08:08.805> some<00:08:08.966> of<00:08:09.046> the<00:08:09.190> leapfrogs<00:08:10.248> in<00:08:10.328> technology,<00:08:11.033> both<00:08:11.834> software<00:08:12.251> and<00:08:12.315> hardware<00:08:12.876> that<00:08:12.988> have<00:08:13.437> happened<00:08:13.693> recently.

00:08:14.670 --> 00:08:18.997
So<00:08:14.799> I<00:08:14.895> just<00:08:15.055> think<00:08:15.920> we'll<00:08:16.016> see<00:08:16.193> it<00:08:16.257> be<00:08:16.417> more<00:08:16.577> pervasive<00:08:17.218> and<00:08:17.266> we'll<00:08:17.475> see<00:08:17.683> it<00:08:17.779> be<00:08:17.859> more<00:08:18.084> performant.

00:08:19.767 --> 00:08:23.233
But<00:08:19.863> that<00:08:20.023> only<00:08:20.505> increases<00:08:21.066> the<00:08:21.163> need<00:08:21.468> for<00:08:21.933> us<00:08:22.013> to<00:08:22.110> have<00:08:22.270> more<00:08:22.414> confidence<00:08:22.992> in<00:08:23.072> it.

00:08:24.180 --> 00:08:24.372
Yeah.

00:08:24.581 --> 00:08:24.837
Yeah.

00:08:24.918 --> 00:08:41.574
The<00:08:25.078> problem<00:08:25.399> you're<00:08:25.560> trying<00:08:25.720> to<00:08:25.816> solve<00:08:26.057> is<00:08:26.266> really<00:08:26.426> interesting<00:08:26.843> around<00:08:27.148> the<00:08:27.244> confidence<00:08:27.790> and<00:08:27.870> like,<00:08:28.528> can<00:08:28.753> you<00:08:28.849> trust<00:08:29.330> what's<00:08:29.555> being<00:08:29.732> spit<00:08:30.036> out<00:08:30.357> and<00:08:30.438> all<00:08:30.598> of<00:08:30.694> that<00:08:30.855> type<00:08:31.112> of<00:08:31.160> stuff<00:08:31.962> while<00:08:32.203> also<00:08:32.395> letting<00:08:32.716> it<00:08:32.861> be<00:08:33.005> creative<00:08:33.406> again,<00:08:33.663> like<00:08:33.904> that<00:08:34.225> you<00:08:34.321> have<00:08:34.465> like<00:08:34.626> this<00:08:34.850> really<00:08:35.107> interesting,<00:08:36.455> uh,<00:08:36.792> like<00:08:36.953> you<00:08:37.033> said,<00:08:37.273> double-edged<00:08:37.675> sword<00:08:38.076> there<00:08:38.541> to<00:08:38.637> make<00:08:38.878> it<00:08:38.958> kind<00:08:39.119> of<00:08:39.199> like<00:08:39.456> exactly<00:08:40.162> right<00:08:40.386> down<00:08:40.563> the<00:08:40.643> middle<00:08:40.948> is<00:08:41.044> interesting.

00:08:42.312 --> 00:08:50.079
Um,<00:08:42.890> what<00:08:43.050> are<00:08:43.130> some<00:08:43.355> interesting<00:08:43.917> use<00:08:44.157> cases<00:08:44.559> that<00:08:44.655> you're<00:08:44.815> seeing<00:08:45.281> beyond<00:08:45.602> what<00:08:46.003> you<00:08:46.099> guys<00:08:46.243> are<00:08:46.324> working<00:08:46.580> on<00:08:46.805> for<00:08:47.126> AI<00:08:47.623> and<00:08:47.768> like<00:08:47.944> interesting<00:08:48.426> companies<00:08:49.003> or<00:08:49.228> things<00:08:49.453> that<00:08:49.613> you're<00:08:49.710> seeing?

00:08:51.153 --> 00:08:53.367
That<00:08:51.330> are<00:08:51.474> getting<00:08:51.779> funded<00:08:52.276> that<00:08:52.517> get<00:08:52.821> you<00:08:52.934> excited?

00:08:54.842 --> 00:09:00.263
Yeah,<00:08:55.644> I<00:08:55.660> mean,<00:08:57.168> there's<00:08:57.729> probably—<00:08:58.130> we<00:08:58.290> don't<00:08:58.435> have<00:08:58.531> enough<00:08:58.691> time<00:08:58.916> to<00:08:58.996> go<00:08:59.092> through<00:08:59.253> all<00:08:59.397> of<00:08:59.509> it,<00:08:59.574> to<00:08:59.718> be<00:08:59.798> completely<00:09:00.135> honest.

00:09:00.311 --> 00:09:08.202
It<00:09:00.391> seems<00:09:00.616> like<00:09:00.857> every<00:09:01.097> day<00:09:01.274> someone<00:09:01.594> is<00:09:01.739> coming<00:09:01.995> up<00:09:02.124> with<00:09:02.220> some,<00:09:02.476> some<00:09:02.701> new—<00:09:04.305> again,<00:09:04.610> yeah,<00:09:04.962> approach,<00:09:05.524> some<00:09:05.748> new<00:09:07.272> thing<00:09:07.432> to<00:09:07.577> attack<00:09:07.914> with<00:09:08.138> it.

00:09:09.197 --> 00:09:29.463
I'd<00:09:09.357> say<00:09:09.902> one<00:09:10.063> of<00:09:10.095> the<00:09:10.175> things<00:09:10.399> that<00:09:10.480> we're<00:09:10.656> seeing<00:09:10.961> more<00:09:11.185> and<00:09:11.282> more<00:09:11.522> is<00:09:11.827> as<00:09:12.003> people<00:09:12.308> start<00:09:12.821> to<00:09:13.190> feel<00:09:13.367> more<00:09:13.607> comfortable<00:09:14.249> with<00:09:14.409> this<00:09:14.569> next<00:09:14.794> wave<00:09:15.067> of<00:09:15.195> Gen<00:09:15.436> AI,<00:09:16.093> they're<00:09:16.254> getting<00:09:16.574> beyond<00:09:17.216> kind<00:09:17.376> of<00:09:17.601> those<00:09:17.841> science<00:09:18.322> projects<00:09:18.900> and<00:09:18.964> those<00:09:19.221> like<00:09:19.766> POCs,<00:09:20.184> I<00:09:20.328> wonder<00:09:20.505> if<00:09:20.665> we<00:09:20.745> can<00:09:20.890> just<00:09:21.050> like<00:09:22.030> have<00:09:22.271> it<00:09:22.672> do<00:09:22.897> QA<00:09:23.459> on<00:09:23.539> our<00:09:23.651> internal<00:09:24.133> documentation,<00:09:25.080> or<00:09:25.160> I<00:09:25.305> wonder<00:09:25.562> if<00:09:25.706> we<00:09:25.803> can<00:09:25.947> like<00:09:26.188> help<00:09:26.445> it<00:09:26.830> answer<00:09:27.231> HR<00:09:27.633> questions<00:09:28.114> just<00:09:28.275> for<00:09:28.387> our<00:09:28.516> internal<00:09:28.917> employees.

00:09:29.977 --> 00:09:38.775
Once<00:09:30.202> they're<00:09:30.282> kind<00:09:30.394> of<00:09:30.539> getting<00:09:31.020> that<00:09:31.325> feel<00:09:31.502> for<00:09:31.807> it<00:09:32.224> and<00:09:32.305> starting<00:09:32.626> to<00:09:32.706> build<00:09:32.931> confidence,<00:09:33.653> in<00:09:33.734> all<00:09:33.910> honesty,<00:09:35.034> now<00:09:35.339> they're<00:09:35.435> starting<00:09:35.821> to<00:09:35.901> say,<00:09:36.158> okay,<00:09:36.399> like,<00:09:36.704> how<00:09:36.800> do<00:09:36.864> we<00:09:37.025> use<00:09:37.185> this<00:09:37.298> to<00:09:37.506> actually<00:09:37.844> transform<00:09:38.405> the<00:09:38.486> business?

00:09:38.887 --> 00:09:45.855
What<00:09:38.983> if<00:09:39.128> we<00:09:39.353> use<00:09:39.529> it<00:09:39.610> to<00:09:39.770> like<00:09:40.412> not<00:09:40.573> just<00:09:40.798> take<00:09:41.135> half<00:09:41.472> a<00:09:41.520> step,<00:09:42.564> take<00:09:42.724> a<00:09:42.804> full<00:09:42.981> step,<00:09:43.302> take<00:09:43.607> 2<00:09:43.768> steps,<00:09:44.105> take<00:09:44.426> 3<00:09:44.667> steps,<00:09:44.972> whatever<00:09:45.293> it<00:09:45.389> may<00:09:45.550> be.

00:09:46.176 --> 00:09:56.095
We're<00:09:46.448> seeing<00:09:46.833> more<00:09:47.137> and<00:09:47.233> more<00:09:47.954> Fortune<00:09:48.355> 500,<00:09:49.236> Global<00:09:49.733> 2000<00:09:50.278> enterprises<00:09:50.855> start<00:09:51.143> to<00:09:51.239> embrace<00:09:51.704> this<00:09:51.945> idea<00:09:52.185> of<00:09:53.018> it's<00:09:53.162> not<00:09:53.339> enough<00:09:53.707> just<00:09:53.883> to<00:09:54.108> do<00:09:54.284> it<00:09:54.524> and<00:09:54.605> check<00:09:54.909> the<00:09:55.005> box<00:09:55.326> saying<00:09:55.646> we<00:09:55.806> did<00:09:55.967> it.

00:09:57.553 --> 00:09:59.396
It's<00:09:57.729> how<00:09:57.890> do<00:09:57.970> we<00:09:58.274> actually<00:09:58.675> create<00:09:58.995> value?

00:09:59.572 --> 00:10:05.229
And<00:09:59.636> I<00:09:59.652> think<00:09:59.813> that's<00:10:00.037> always<00:10:00.357> been<00:10:00.470> the<00:10:00.614> issue<00:10:00.854> even<00:10:01.095> through<00:10:01.559> the<00:10:01.655> last<00:10:02.040> wave<00:10:02.232> of<00:10:02.441> AI<00:10:02.857> and<00:10:03.402> machine<00:10:03.739> learning<00:10:04.059> and<00:10:04.139> data<00:10:04.364> science<00:10:04.700> before<00:10:05.181> it.

00:10:06.303 --> 00:10:13.002
It's<00:10:06.623> a<00:10:06.703> really<00:10:06.928> cool<00:10:07.184> technology<00:10:08.306> and<00:10:08.386> it's<00:10:08.546> novel,<00:10:09.011> but<00:10:09.123> it<00:10:09.572> always<00:10:09.908> comes<00:10:10.148> back<00:10:10.405> to<00:10:10.485> like,<00:10:10.773> how<00:10:10.854> does<00:10:11.014> this<00:10:11.190> actually<00:10:12.120> move<00:10:12.376> the<00:10:12.456> ball<00:10:12.649> forward.

00:10:13.901 --> 00:10:14.093
Yeah.

00:10:14.799 --> 00:10:15.104
Yeah.

00:10:15.345 --> 00:10:19.132
And<00:10:15.521> then,<00:10:15.987> um,<00:10:17.030> as<00:10:17.206> you,<00:10:17.527> like,<00:10:17.752> you<00:10:17.832> just<00:10:18.009> raised<00:10:18.265> a<00:10:18.330> Series<00:10:18.715> A,<00:10:18.875> right?

00:10:19.437 --> 00:10:26.417
Uh,<00:10:19.934> with,<00:10:20.336> with<00:10:20.496> Andreessen<00:10:21.363> as,<00:10:21.764> as<00:10:21.924> the<00:10:22.020> lead,<00:10:22.325> which<00:10:22.502> is<00:10:22.646> a<00:10:22.823> huge,<00:10:23.850> you<00:10:23.882> know,<00:10:24.026> so<00:10:24.203> actually<00:10:24.411> it<00:10:24.492> was<00:10:24.588> Two<00:10:24.764> Sigma<00:10:25.134> Ventures<00:10:25.775> for<00:10:25.872> the<00:10:25.936> Series<00:10:26.337> A.

00:10:26.514 --> 00:10:27.525
Andreessen<00:10:26.851> led<00:10:27.011> our<00:10:27.155> seed.

00:10:27.922 --> 00:10:28.776
Uh,<00:10:27.942> Andreessen<00:10:28.231> led<00:10:28.503> the<00:10:28.536> seed.

00:10:28.905 --> 00:10:29.177
Yeah.

00:10:29.290 --> 00:10:29.338
Yeah.

00:10:29.482 --> 00:10:31.841
And<00:10:29.563> participated<00:10:30.204> in<00:10:30.349> the<00:10:30.429> A.<00:10:31.713> Yeah.

00:10:31.889 --> 00:10:33.863
Tell<00:10:32.066> me,<00:10:32.355> like,<00:10:32.676> what's<00:10:32.852> it<00:10:32.997> been<00:10:33.093> like<00:10:33.317> raising<00:10:33.654> money?

00:10:33.975 --> 00:10:38.051
And<00:10:34.056> maybe<00:10:34.377> like<00:10:34.633> a<00:10:34.778> broader<00:10:35.163> question<00:10:35.500> there<00:10:35.965> is,<00:10:36.463> I<00:10:36.527> mean,<00:10:36.703> you've<00:10:36.848> done<00:10:37.024> this<00:10:37.185> once<00:10:37.425> before,<00:10:37.827> right?

00:10:38.132 --> 00:10:39.239
You've<00:10:38.292> gone<00:10:38.549> through<00:10:38.645> this<00:10:38.854> process.

00:10:39.672 --> 00:10:45.918
What<00:10:39.769> have<00:10:39.897> you<00:10:40.138> learned<00:10:40.491> and<00:10:40.555> what<00:10:40.684> have<00:10:40.796> you<00:10:40.957> taken<00:10:41.358> there<00:10:41.776> to<00:10:42.145> kind<00:10:42.386> of,<00:10:42.707> uh,<00:10:43.269> you<00:10:43.301> know,<00:10:43.991> maybe<00:10:44.232> avoid<00:10:44.569> some<00:10:44.794> pitfalls<00:10:45.276> from<00:10:45.388> the<00:10:45.453> first<00:10:45.709> time?

00:10:46.737 --> 00:10:50.767
Yeah,<00:10:47.203> I'd<00:10:47.363> say,<00:10:47.925> uh,<00:10:48.423> I<00:10:48.487> made<00:10:48.744> a<00:10:48.760> lot<00:10:49.146> of<00:10:49.178> mistakes<00:10:49.868> the<00:10:50.013> first<00:10:50.189> time<00:10:50.430> around.

00:10:51.056 --> 00:10:56.837
Um,<00:10:51.554> and<00:10:51.650> so<00:10:52.517> the<00:10:52.598> joy<00:10:52.919> of<00:10:52.999> doing<00:10:53.256> it<00:10:53.320> the<00:10:53.401> second<00:10:53.722> time<00:10:53.979> is<00:10:54.043> you<00:10:54.123> can<00:10:54.284> learn<00:10:54.525> from<00:10:54.765> those<00:10:55.006> mistakes<00:10:55.552> and<00:10:55.632> make<00:10:56.114> all<00:10:56.227> new<00:10:56.371> mistakes.

00:10:57.720 --> 00:11:04.769
Um,<00:10:58.603> but<00:10:59.181> one<00:10:59.277> of<00:10:59.342> the<00:10:59.422> nice<00:10:59.663> things<00:10:59.839> about<00:11:00.080> fundraising,<00:11:00.626> especially<00:11:01.011> the<00:11:01.124> second<00:11:01.429> time<00:11:01.589> around,<00:11:01.911> is<00:11:01.991> we'd<00:11:02.232> been<00:11:02.392> able<00:11:02.553> to<00:11:02.697> build<00:11:02.938> up<00:11:03.099> some<00:11:03.179> of<00:11:03.211> these<00:11:03.436> relationships<00:11:04.159> over<00:11:04.319> many<00:11:04.704> years.

00:11:05.266 --> 00:11:09.192
And<00:11:05.923> Andreessen<00:11:06.323> Horowitz<00:11:06.884> was<00:11:07.429> one<00:11:07.541> of<00:11:07.605> the<00:11:07.750> lead<00:11:07.846> investors<00:11:08.327> in<00:11:08.663> SigOpt.

00:11:09.993 --> 00:11:15.986
And<00:11:10.249> I<00:11:10.330> got<00:11:10.490> to<00:11:10.570> work<00:11:10.730> with<00:11:10.890> Martin<00:11:11.275> Casado,<00:11:11.916> like,<00:11:12.156> right<00:11:12.349> as<00:11:12.493> he<00:11:12.589> was<00:11:13.134> joining<00:11:13.438> Andreessen<00:11:13.935> Horowitz,<00:11:14.480> and<00:11:14.576> he<00:11:14.656> was<00:11:14.816> on<00:11:14.880> our<00:11:15.057> board<00:11:15.361> for<00:11:15.441> many<00:11:15.698> years.

00:11:16.098 --> 00:11:26.242
And<00:11:16.675> so<00:11:17.140> he<00:11:17.236> got<00:11:17.412> to<00:11:17.460> see<00:11:18.021> what<00:11:18.165> things<00:11:18.406> looked<00:11:18.646> like<00:11:18.902> when<00:11:19.047> they<00:11:19.143> were<00:11:19.223> going<00:11:19.463> well,<00:11:20.088> they<00:11:20.184> were<00:11:20.329> going<00:11:20.585> poorly,<00:11:21.546> we<00:11:21.627> were<00:11:22.524> selling<00:11:22.973> the<00:11:23.053> company<00:11:24.190> through<00:11:24.351> the<00:11:24.495> pandemic,<00:11:25.152> like,<00:11:25.376> all<00:11:25.456> of<00:11:25.552> these<00:11:25.649> sorts<00:11:25.937> of<00:11:26.033> things.

00:11:26.370 --> 00:11:32.677
So<00:11:26.995> it<00:11:27.075> made<00:11:27.235> it<00:11:27.395> really<00:11:27.860> easy<00:11:28.437> when<00:11:28.581> I<00:11:28.677> was<00:11:29.078> going<00:11:29.318> to<00:11:29.639> start<00:11:29.879> the<00:11:29.959> next<00:11:30.183> company<00:11:30.520> to<00:11:30.664> say,<00:11:30.985> hey,<00:11:31.885> Do<00:11:31.917> you<00:11:31.950> want<00:11:32.047> to<00:11:32.111> run<00:11:32.289> it<00:11:32.370> back?

00:11:32.806 --> 00:11:43.238
Do<00:11:32.935> you<00:11:33.016> want<00:11:33.145> to<00:11:33.178> like<00:11:34.244> try<00:11:34.373> it<00:11:34.486> again<00:11:34.874> now<00:11:35.116> that<00:11:35.278> we've<00:11:35.520> learned<00:11:35.908> all<00:11:35.989> of<00:11:36.086> these<00:11:36.263> different<00:11:36.490> things<00:11:36.797> together<00:11:37.217> and<00:11:37.378> built<00:11:37.637> up<00:11:37.782> all<00:11:37.847> of<00:11:38.008> this<00:11:38.170> trust<00:11:38.428> and<00:11:38.574> confidence<00:11:39.075> in<00:11:39.220> each<00:11:39.398> other<00:11:41.157> and<00:11:41.657> attack<00:11:42.141> this<00:11:42.448> adjacent<00:11:42.867> problem?

00:11:43.287 --> 00:11:51.116
It's<00:11:43.432> not<00:11:43.674> a<00:11:43.754> completely<00:11:44.158> dissimilar<00:11:44.642> problem,<00:11:45.029> but<00:11:45.126> it's<00:11:45.287> like<00:11:46.013> really<00:11:46.400> taking<00:11:46.804> everything<00:11:47.142> that<00:11:47.239> we<00:11:47.449> learned<00:11:47.772> from<00:11:47.885> the<00:11:47.965> last<00:11:48.272> decade<00:11:48.917> and<00:11:49.078> seeing<00:11:49.401> if<00:11:49.450> we<00:11:49.564> can<00:11:50.299> go<00:11:50.479> even<00:11:50.773> further.

00:11:52.590 --> 00:12:02.203
And<00:11:52.654> similarly<00:11:53.092> with<00:11:53.190> Two<00:11:53.336> Sigma,<00:11:53.595> they<00:11:53.741> were<00:11:53.903> a<00:11:54.001> big<00:11:54.163> customer<00:11:54.568> of<00:11:54.747> SigOpt<00:11:55.298> and<00:11:55.704> investor<00:11:56.174> in<00:11:56.271> SigOpt<00:11:56.985> and<00:11:57.715> one<00:11:57.828> of<00:11:57.893> these<00:11:58.120> firms<00:11:58.445> that<00:11:58.710> just<00:11:58.887> knows<00:11:59.611> how<00:11:59.933> we<00:12:00.030> do<00:12:00.191> business,<00:12:00.657> what<00:12:00.818> we<00:12:00.996> build,<00:12:01.559> et<00:12:01.656> cetera.

00:12:03.571 --> 00:12:09.126
And<00:12:03.652> so<00:12:03.797> I<00:12:03.877> think<00:12:04.038> a<00:12:04.054> lot<00:12:04.231> of<00:12:04.312> it<00:12:04.441> is<00:12:04.602> that<00:12:04.698> we<00:12:04.779> get<00:12:04.924> to<00:12:05.810> stand<00:12:06.154> on<00:12:06.236> our<00:12:06.811> own<00:12:06.992> shoulders<00:12:08.223> to<00:12:08.371> jump<00:12:08.551> forward<00:12:08.945> here.

00:12:11.610 --> 00:12:12.850
How<00:12:11.658> much<00:12:11.787> did<00:12:11.964> you<00:12:12.028> raise<00:12:12.173> in<00:12:12.286> the<00:12:12.367> Series<00:12:12.769> A?

00:12:12.930 --> 00:12:15.169
Like<00:12:13.494> the<00:12:13.591> seed<00:12:13.961> was<00:12:14.138> $19<00:12:14.557> million,<00:12:14.702> is<00:12:14.782> that<00:12:14.927> right?

00:12:15.765 --> 00:12:17.730
Seed<00:12:16.071> was<00:12:16.232> $11<00:12:16.570> million,<00:12:16.715> Series<00:12:17.021> A<00:12:17.118> was<00:12:17.295> $19<00:12:17.585> million.

00:12:17.778 --> 00:12:18.712
So<00:12:18.020> $30<00:12:18.213> million<00:12:18.390> total.

00:12:20.594 --> 00:12:26.763
And<00:12:20.691> one<00:12:21.325> big<00:12:21.503> difference<00:12:21.828> between<00:12:22.153> this<00:12:22.477> and<00:12:22.607> the<00:12:22.640> last<00:12:23.451> company<00:12:23.857> was<00:12:24.020> I<00:12:24.101> think<00:12:24.263> it<00:12:24.344> took<00:12:24.523> us<00:12:24.669> like<00:12:24.848> 4<00:12:24.994> or<00:12:25.221> 5<00:12:25.497> years<00:12:25.806> to<00:12:25.968> raise<00:12:26.211> a<00:12:26.228> Series<00:12:26.698> A.

00:12:28.088 --> 00:12:30.487
After<00:12:28.330> going<00:12:28.507> through<00:12:28.652> Y<00:12:28.829> Combinator<00:12:29.634> and<00:12:29.714> the<00:12:29.795> SIGOP<00:12:30.197> days.

00:12:30.825 --> 00:12:38.635
And<00:12:31.083> here<00:12:31.405> we<00:12:31.807> raised<00:12:32.129> the<00:12:32.226> seed<00:12:32.677> right<00:12:32.838> out<00:12:32.950> of<00:12:33.015> the<00:12:33.160> gate,<00:12:33.594> started<00:12:33.981> with<00:12:34.142> a<00:12:34.222> team<00:12:34.544> of<00:12:34.689> 11,<00:12:36.557> and<00:12:36.637> then<00:12:36.718> we<00:12:36.863> raised<00:12:37.120> the<00:12:37.330> Series<00:12:37.587> A<00:12:37.668> within<00:12:37.910> the<00:12:38.023> first<00:12:38.394> year.

00:12:38.974 --> 00:12:43.746
And<00:12:39.071> so<00:12:40.892> again,<00:12:41.553> making<00:12:41.940> all<00:12:42.053> new<00:12:42.198> mistakes,<00:12:42.746> but<00:12:42.843> 10<00:12:43.085> times<00:12:43.391> faster.

00:12:45.370 --> 00:12:54.559
What<00:12:45.434> do<00:12:45.466> you<00:12:45.498> make<00:12:45.821> of<00:12:45.966> like<00:12:46.433> those<00:12:46.997> numbers<00:12:47.497> being<00:12:47.819> seed<00:12:48.222> and<00:12:48.367> Series<00:12:48.851> A<00:12:48.867> rounds<00:12:49.351> that<00:12:49.592> those<00:12:49.915> used<00:12:50.076> to<00:12:50.221> be<00:12:50.318> like<00:12:50.656> Series<00:12:51.285> D<00:12:51.688> rounds,<00:12:52.171> Series<00:12:52.574> C,<00:12:53.042> you<00:12:53.074> know,<00:12:53.219> like<00:12:53.461> the<00:12:53.982> that<00:12:54.142> type<00:12:54.318> of<00:12:54.383> thing.

00:12:54.943 --> 00:12:57.716
How<00:12:55.040> do<00:12:55.168> you<00:12:55.264> think<00:12:55.569> that's<00:12:55.905> changing<00:12:56.402> and<00:12:56.530> why<00:12:56.771> do<00:12:56.803> you<00:12:56.947> think<00:12:57.107> that's<00:12:57.364> changing?

00:12:58.373 --> 00:12:59.415
Yeah,<00:12:58.710> that's<00:12:58.870> a<00:12:58.934> great<00:12:59.111> question.

00:12:59.591 --> 00:13:11.419
I<00:12:59.736> remember<00:13:00.216> when<00:13:00.713> we<00:13:00.938> were<00:13:02.396> demo<00:13:02.781> day<00:13:03.262> of<00:13:03.438> Winter<00:13:03.903> '15<00:13:04.704> Y<00:13:04.880> Combinator,<00:13:05.505> we<00:13:05.842> raised<00:13:06.098> a<00:13:06.162> $2.2<00:13:06.804> million<00:13:07.365> seed<00:13:07.925> and<00:13:08.006> that<00:13:08.102> was<00:13:08.470> considered<00:13:08.887> a<00:13:09.208> mango<00:13:09.929> seed,<00:13:10.410> like<00:13:10.570> a<00:13:10.650> large<00:13:11.051> seed.

00:13:12.012 --> 00:13:15.330
And<00:13:12.093> now<00:13:12.637> that<00:13:12.798> would<00:13:12.958> be<00:13:13.054> considered<00:13:13.936> an<00:13:14.160> extremely<00:13:14.753> small<00:13:15.058> seed.

00:13:15.378 --> 00:13:16.997
So<00:13:15.458> I<00:13:15.538> think<00:13:16.035> a<00:13:16.099> few<00:13:16.340> things<00:13:16.500> have<00:13:16.644> happened.

00:13:17.526 --> 00:13:23.087
One<00:13:17.686> is,<00:13:17.846> I<00:13:17.878> mean,<00:13:18.087> some<00:13:18.263> businesses<00:13:18.808> have<00:13:19.225> pretty<00:13:19.609> large<00:13:19.946> capital<00:13:21.148> constraints<00:13:21.869> and<00:13:21.965> so<00:13:22.126> they<00:13:22.206> need<00:13:22.350> to<00:13:22.446> be<00:13:22.574> able<00:13:22.751> to<00:13:22.831> do<00:13:22.895> that.

00:13:23.280 --> 00:13:25.801
And<00:13:23.456> those<00:13:23.633> companies<00:13:24.018> are<00:13:24.099> raising<00:13:24.356> like<00:13:24.516> $100<00:13:24.837> million<00:13:25.319> seats.

00:13:26.427 --> 00:13:29.639
Um,<00:13:27.246> but<00:13:27.327> I'd<00:13:27.471> say<00:13:27.567> like<00:13:27.808> expectations<00:13:28.499> have<00:13:28.611> continued<00:13:29.013> to<00:13:29.093> ratchet<00:13:29.479> up.

00:13:29.816 --> 00:13:31.631
Like,<00:13:30.523> what<00:13:30.699> can<00:13:30.860> you<00:13:31.085> do<00:13:31.406> when?

00:13:31.823 --> 00:13:33.526
Like,<00:13:32.128> when<00:13:32.225> do<00:13:32.289> you<00:13:32.385> need<00:13:32.546> to<00:13:32.610> start<00:13:32.948> making<00:13:33.349> money?

00:13:33.654 --> 00:13:35.244
When<00:13:33.751> do<00:13:33.799> you<00:13:33.895> need<00:13:34.056> to<00:13:34.088> start<00:13:34.393> being<00:13:34.634> profitable?

00:13:35.597 --> 00:13:40.544
Like,<00:13:36.577> how<00:13:37.107> large<00:13:37.380> of<00:13:37.460> a<00:13:37.589> team<00:13:37.910> do<00:13:37.990> you<00:13:38.022> need<00:13:38.231> to<00:13:38.263> do<00:13:38.569> to<00:13:38.729> really<00:13:38.890> make<00:13:39.115> a<00:13:39.147> splash<00:13:39.693> in<00:13:39.773> this<00:13:39.934> environment?

00:13:41.058 --> 00:13:46.968
And<00:13:41.138> so<00:13:41.315> as<00:13:41.443> those<00:13:41.620> expectations<00:13:42.310> have<00:13:42.423> increased,<00:13:43.049> like,<00:13:43.451> you<00:13:43.483> need<00:13:43.692> to<00:13:43.788> be<00:13:43.965> able<00:13:44.109> to<00:13:44.254> put<00:13:44.414> capital<00:13:44.912> to<00:13:45.073> work<00:13:45.394> to<00:13:45.538> be<00:13:45.635> able<00:13:45.747> to<00:13:45.812> meet<00:13:46.036> those<00:13:46.277> expectations.

00:13:47.883 --> 00:13:56.844
But<00:13:47.964> I<00:13:48.124> do<00:13:48.590> look<00:13:48.767> back<00:13:49.329> and<00:13:49.489> think,<00:13:50.437> when<00:13:50.597> I<00:13:50.629> was<00:13:50.854> just<00:13:51.095> a<00:13:51.256> dumb<00:13:51.480> grad<00:13:51.737> student<00:13:52.123> trying<00:13:52.299> to<00:13:52.380> figure<00:13:52.604> out<00:13:52.685> how<00:13:52.781> to<00:13:52.845> start<00:13:53.038> a<00:13:53.102> company<00:13:53.423> for<00:13:53.568> the<00:13:53.745> first<00:13:53.905> time,<00:13:54.226> like<00:13:54.612> I<00:13:54.708> wouldn't<00:13:54.901> have<00:13:55.045> known<00:13:55.350> what<00:13:55.511> to<00:13:55.655> do<00:13:55.832> with<00:13:55.993> $30<00:13:56.169> million.

00:13:57.036 --> 00:14:10.171
And<00:13:57.133> so<00:13:57.277> I<00:13:57.357> really<00:13:57.518> feel<00:13:57.903> for<00:13:58.016> some<00:13:58.160> of<00:13:58.241> these<00:13:58.401> first-time<00:13:58.803> founders<00:13:59.284> that<00:13:59.445> now<00:13:59.991> have<00:14:00.264> these<00:14:00.489> raised<00:14:00.810> expectations,<00:14:02.078> have<00:14:02.801> the<00:14:02.897> capital<00:14:03.379> because<00:14:03.684> that's<00:14:03.861> become<00:14:04.166> the<00:14:04.198> standard,<00:14:05.386> but<00:14:05.466> then<00:14:05.627> maybe<00:14:05.932> haven't<00:14:06.446> deployed<00:14:06.831> that<00:14:06.992> before<00:14:07.955> or<00:14:08.116> been<00:14:08.341> able<00:14:08.598> to<00:14:09.079> build<00:14:09.304> a<00:14:09.384> team<00:14:09.722> or<00:14:09.754> scale<00:14:10.123> it.

00:14:10.364 --> 00:14:14.651
And<00:14:10.508> it's,<00:14:11.151> it<00:14:11.183> just<00:14:11.584> exacerbates<00:14:12.612> all<00:14:12.756> of<00:14:12.853> the<00:14:12.997> different<00:14:13.254> things<00:14:13.656> that<00:14:13.800> you<00:14:13.896> need<00:14:14.137> to<00:14:14.298> get<00:14:14.458> right.

00:14:14.924 --> 00:14:19.092
Like<00:14:15.036> a<00:14:15.101> startup<00:14:15.518> is<00:14:15.663> about<00:14:15.984> getting<00:14:16.193> 100<00:14:16.706> things<00:14:17.028> right,<00:14:17.445> or<00:14:17.509> at<00:14:17.654> least<00:14:18.152> not<00:14:18.460> getting<00:14:18.703> them<00:14:18.865> wrong.

00:14:19.838 --> 00:14:25.462
Um,<00:14:20.648> and<00:14:21.394> the<00:14:21.539> idea<00:14:21.880> is<00:14:22.188> part<00:14:22.382> of<00:14:22.512> it,<00:14:22.609> the<00:14:22.690> go-to-market<00:14:23.258> is<00:14:23.403> part<00:14:23.582> of<00:14:23.663> it,<00:14:23.825> but<00:14:23.906> there's<00:14:24.084> a<00:14:24.149> million<00:14:24.716> other<00:14:24.943> ways<00:14:25.284> to<00:14:25.365> die.

00:14:27.104 --> 00:14:28.520
What<00:14:27.200> does<00:14:27.345> a<00:14:27.426> typical<00:14:27.812> day<00:14:27.973> look<00:14:28.150> like<00:14:28.391> for<00:14:28.472> you?

00:14:28.617 --> 00:14:31.176
Oh,<00:14:28.858> that's<00:14:30.339> a,<00:14:30.484> that's<00:14:30.661> a<00:14:30.709> good<00:14:30.886> question.

00:14:31.434 --> 00:14:33.236
Uh,<00:14:32.013> every<00:14:32.254> day<00:14:32.399> is<00:14:32.512> a<00:14:32.576> little<00:14:32.770> bit<00:14:32.898> different.

00:14:34.570 --> 00:14:44.426
Um,<00:14:34.618> but,<00:14:35.102> uh,<00:14:35.602> when<00:14:35.715> I'm<00:14:35.844> at<00:14:35.989> home,<00:14:36.634> uh,<00:14:36.796> and<00:14:36.892> not<00:14:37.054> on<00:14:37.118> the<00:14:37.280> road,<00:14:37.844> uh,<00:14:38.409> I,<00:14:38.731> I<00:14:38.748> have<00:14:39.054> two<00:14:39.215> young<00:14:39.473> sons,<00:14:40.022> so<00:14:40.199> it's<00:14:40.748> helping<00:14:41.151> get<00:14:41.329> my,<00:14:41.716> uh,<00:14:41.974> 2-year-old<00:14:42.458> and<00:14:42.554> 6-year-old<00:14:43.329> out<00:14:43.426> the<00:14:43.506> door<00:14:43.813> and<00:14:43.910> ready<00:14:44.076> for<00:14:44.168> school.

00:14:46.004 --> 00:15:03.451
And<00:14:46.068> then<00:14:46.180> a<00:14:46.229> massive<00:14:46.645> context<00:14:47.191> switch<00:14:47.688> into,<00:14:48.474> uh,<00:14:49.275> trying<00:14:49.436> to<00:14:49.516> figure<00:14:49.837> out,<00:14:50.253> uh,<00:14:50.414> what<00:14:50.526> I'm<00:14:50.654> doing<00:14:51.135> that<00:14:51.376> day,<00:14:51.697> whether<00:14:51.953> it's<00:14:53.220> podcast<00:14:53.701> interviews<00:14:54.102> like<00:14:54.326> this,<00:14:54.663> whether<00:14:54.888> it's<00:14:54.984> customer<00:14:55.465> calls,<00:14:56.026> whether<00:14:56.267> it's<00:14:56.587> diving<00:14:56.988> deep<00:14:57.309> on<00:14:57.870> go-to-market,<00:14:58.977> uh,<00:14:59.153> product,<00:15:00.099> um,<00:15:00.837> operations,<00:15:01.623> all<00:15:01.655> the<00:15:01.799> various,<00:15:02.344> uh,<00:15:02.681> parts<00:15:02.921> of<00:15:02.986> the<00:15:03.066> company.

00:15:04.108 --> 00:15:05.840
Thankfully,<00:15:04.669> I<00:15:04.749> have<00:15:04.846> an<00:15:04.926> incredible<00:15:05.567> team.

00:15:05.888 --> 00:15:11.789
This<00:15:06.064> is<00:15:06.128> one<00:15:06.289> of<00:15:06.353> the<00:15:06.385> things<00:15:06.738> that<00:15:06.834> we<00:15:06.914> were<00:15:07.010> able<00:15:07.235> to<00:15:07.331> build<00:15:07.508> upon,<00:15:08.197> um,<00:15:08.389> along<00:15:08.694> with<00:15:08.838> the<00:15:08.919> investors<00:15:09.656> and<00:15:09.720> some<00:15:09.897> of<00:15:09.961> our<00:15:10.137> customers<00:15:10.683> from<00:15:10.843> the<00:15:10.923> previous<00:15:11.324> companies.

00:15:12.509 --> 00:15:20.899
Being<00:15:12.687> able<00:15:12.833> to<00:15:12.995> bring<00:15:13.254> in<00:15:13.902> a<00:15:13.983> lot<00:15:14.161> of<00:15:14.226> people<00:15:14.550> who<00:15:14.615> had<00:15:14.777> worked<00:15:14.955> with<00:15:15.052> us<00:15:15.198> before,<00:15:16.089> and<00:15:16.170> then<00:15:16.656> their<00:15:16.980> network,<00:15:17.855> and<00:15:18.357> again,<00:15:19.248> having<00:15:20.042> a<00:15:20.058> strong<00:15:20.447> culture.

00:15:22.646 --> 00:15:27.192
But<00:15:22.792> just<00:15:23.035> figuring<00:15:23.520> out<00:15:23.682> how<00:15:23.860> I<00:15:24.005> can<00:15:24.102> get<00:15:24.329> out<00:15:24.410> of<00:15:24.442> the<00:15:24.571> way<00:15:24.911> is<00:15:25.461> literally<00:15:26.270> every<00:15:26.448> minute<00:15:26.755> of<00:15:26.852> the<00:15:26.933> day.

00:15:27.661 --> 00:15:34.429
We<00:15:27.806> have<00:15:27.920> a<00:15:27.984> lot<00:15:28.211> of<00:15:28.308> really<00:15:28.470> smart<00:15:28.793> people,<00:15:29.117> and<00:15:29.197> I<00:15:29.295> just<00:15:29.505> need<00:15:29.634> to<00:15:30.801> figure<00:15:31.060> out<00:15:31.206> how<00:15:31.238> to<00:15:31.367> get<00:15:31.853> out<00:15:31.951> of<00:15:32.032> the<00:15:32.096> way<00:15:32.258> as<00:15:32.356> much<00:15:32.534> as<00:15:32.663> I<00:15:32.744> can,<00:15:33.246> and<00:15:33.392> then<00:15:33.797> tell<00:15:33.975> our<00:15:34.056> story.

00:15:34.866 --> 00:15:37.669
And<00:15:36.065> every<00:15:36.308> day<00:15:36.486> I<00:15:36.567> try<00:15:36.745> to<00:15:36.891> do<00:15:36.988> that<00:15:37.150> a<00:15:37.215> little<00:15:37.458> bit<00:15:37.604> better.

00:15:38.633 --> 00:15:43.609
Every<00:15:38.954> little,<00:15:39.597> every<00:15:39.902> 1%<00:15:40.560> gain<00:15:41.202> towards<00:15:41.603> that<00:15:42.068> end<00:15:43.272> helps.

00:15:43.770 --> 00:15:49.324
And<00:15:43.914> if<00:15:43.995> you<00:15:44.075> can<00:15:44.171> get<00:15:44.492> 71%<00:15:45.552> gains<00:15:46.820> cumulatively,<00:15:47.542> that's<00:15:47.719> a<00:15:47.847> doubling<00:15:48.264> in<00:15:48.746> productivity.

00:15:49.404 --> 00:15:52.245
So<00:15:49.613> I<00:15:49.709> just,<00:15:50.351> if<00:15:50.447> I<00:15:50.576> can<00:15:50.752> pull<00:15:50.993> one<00:15:51.154> of<00:15:51.186> those<00:15:51.378> out<00:15:51.539> a<00:15:51.555> week,<00:15:51.748> I'm<00:15:51.940> happy.

00:15:53.706 --> 00:15:55.038
How<00:15:53.802> have<00:15:53.979> you<00:15:54.107> thought<00:15:54.300> about<00:15:54.525> recruiting?

00:15:55.231 --> 00:16:00.319
I<00:15:55.247> mean,<00:15:55.407> you<00:15:55.488> mentioned<00:15:55.793> like<00:15:55.969> a<00:15:55.985> lot<00:15:56.226> of<00:15:56.387> the<00:15:56.451> folks<00:15:56.836> who<00:15:57.013> are<00:15:57.157> coming<00:15:57.414> in<00:15:57.558> here<00:15:57.735> now<00:15:58.441> are,<00:15:58.778> you<00:15:58.875> know,<00:15:59.019> you've<00:15:59.244> worked<00:15:59.420> with<00:15:59.597> in<00:15:59.661> the<00:15:59.822> past.

00:16:00.865 --> 00:16:14.046
How<00:16:00.945> do<00:16:01.170> you<00:16:01.571> develop<00:16:02.454> and<00:16:02.952> find<00:16:03.433> talent<00:16:03.995> in<00:16:04.156> this<00:16:04.557> new<00:16:04.862> kind<00:16:05.039> of<00:16:05.183> world<00:16:05.504> where<00:16:06.146> You<00:16:06.226> know,<00:16:06.467> you,<00:16:06.660> some,<00:16:07.110> I<00:16:07.126> don't<00:16:07.286> know<00:16:07.382> where<00:16:07.591> you're<00:16:07.752> at<00:16:08.009> and<00:16:08.266> this<00:16:08.635> may<00:16:08.731> be<00:16:08.892> part<00:16:09.052> of<00:16:09.101> the<00:16:09.197> question<00:16:09.454> is<00:16:09.534> like<00:16:09.775> if<00:16:09.919> you're<00:16:10.321> all<00:16:10.417> in<00:16:10.562> office,<00:16:11.268> hybrid,<00:16:12.488> uh,<00:16:12.890> all<00:16:13.050> remote,<00:16:13.404> things<00:16:13.709> like<00:16:13.869> that.

00:16:14.110 --> 00:16:17.723
Like<00:16:14.271> how,<00:16:14.576> how<00:16:14.688> do<00:16:14.817> you<00:16:14.977> go<00:16:15.074> about,<00:16:15.395> it<00:16:15.475> must<00:16:15.716> be<00:16:15.876> different<00:16:16.198> recruiting<00:16:16.679> than<00:16:16.824> it<00:16:16.920> was<00:16:17.241> the<00:16:17.322> first<00:16:17.562> time.

00:16:18.269 --> 00:16:19.007
Yeah,<00:16:18.622> definitely.

00:16:19.329 --> 00:16:28.738
And<00:16:19.409> I<00:16:19.489> think<00:16:20.035> again,<00:16:20.276> it's,<00:16:20.693> there's<00:16:21.015> a<00:16:21.095> handful<00:16:21.560> of<00:16:21.657> things<00:16:21.882> that<00:16:21.978> have<00:16:22.122> changed,<00:16:22.460> so<00:16:22.556> it's<00:16:22.700> hard<00:16:22.861> to<00:16:23.022> kind<00:16:23.166> of<00:16:23.262> isolate<00:16:23.760> it<00:16:23.889> to<00:16:24.306> one<00:16:24.531> specific<00:16:25.013> change,<00:16:25.510> but<00:16:25.832> one<00:16:26.056> is<00:16:26.233> like<00:16:26.875> we<00:16:26.955> have<00:16:27.277> more<00:16:27.437> experience<00:16:28.385> with<00:16:28.625> it.

00:16:29.348 --> 00:16:30.761
Um,<00:16:30.472> and<00:16:30.568> we.

00:16:31.612 --> 00:16:36.039
We're<00:16:31.997> able<00:16:32.253> to<00:16:32.398> kind<00:16:32.574> of<00:16:32.671> again<00:16:33.858> build<00:16:34.114> upon<00:16:34.820> the<00:16:34.900> success<00:16:35.302> that<00:16:35.462> we<00:16:35.606> had<00:16:35.783> before.

00:16:36.280 --> 00:16:43.098
So<00:16:36.826> it's<00:16:37.323> not<00:16:37.708> again<00:16:38.029> that<00:16:38.173> I'm<00:16:38.831> a<00:16:38.911> dumb<00:16:39.152> grad<00:16:39.456> student<00:16:39.858> who's<00:16:40.018> like<00:16:40.259> figuring<00:16:40.660> out<00:16:40.836> how<00:16:40.997> to<00:16:41.879> build<00:16:42.120> a<00:16:42.184> team<00:16:42.360> for<00:16:42.505> the<00:16:42.601> first<00:16:42.777> time.

00:16:43.146 --> 00:16:51.681
It's<00:16:43.307> like<00:16:43.483> we<00:16:43.627> have<00:16:43.964> better<00:16:44.141> systems<00:16:44.526> in<00:16:44.686> place<00:16:45.087> so<00:16:45.248> we<00:16:45.328> can<00:16:45.472> attract,<00:16:46.050> I'd<00:16:46.194> say,<00:16:46.852> more<00:16:46.996> experienced<00:16:47.734> talent<00:16:48.937> who's<00:16:49.242> looking<00:16:49.579> to<00:16:49.659> just<00:16:49.900> run<00:16:50.157> fast<00:16:50.606> as<00:16:50.702> opposed<00:16:51.023> to<00:16:51.167> figure<00:16:51.408> it<00:16:51.504> out.

00:16:52.210 --> 00:16:57.729
Figuring<00:16:52.531> it<00:16:52.611> out<00:16:52.723> is<00:16:52.868> very<00:16:53.108> valuable<00:16:53.510> and<00:16:53.606> a<00:16:53.670> startup<00:16:54.055> is<00:16:54.135> a<00:16:54.231> great<00:16:54.616> way<00:16:54.713> to<00:16:54.953> learn<00:16:55.595> really<00:16:55.755> fast<00:16:56.028> because<00:16:56.237> it's<00:16:56.381> like<00:16:56.558> drinking<00:16:56.943> out<00:16:56.975> of<00:16:57.119> a<00:16:57.199> fire<00:16:57.440> hose.

00:16:58.306 --> 00:17:08.841
Um,<00:16:59.204> but<00:16:59.605> at<00:16:59.685> this<00:16:59.845> point<00:17:00.102> it<00:17:00.230> feels<00:17:00.407> like<00:17:00.567> we're<00:17:00.647> trying<00:17:00.888> to<00:17:01.834> like<00:17:02.010> assemble<00:17:02.395> the<00:17:02.555> actual<00:17:03.133> fire<00:17:03.630> team,<00:17:04.175> like,<00:17:04.961> and<00:17:05.153> so,<00:17:05.843> um,<00:17:07.366> that's,<00:17:07.606> that's<00:17:08.104> been<00:17:08.200> a<00:17:08.264> big<00:17:08.488> difference.

00:17:08.889 --> 00:17:22.391
I'd<00:17:09.050> also<00:17:09.210> say<00:17:09.386> like<00:17:09.563> the<00:17:09.643> market<00:17:10.092> has<00:17:10.252> shifted<00:17:10.813> pretty<00:17:10.990> dramatically,<00:17:11.856> um,<00:17:12.754> whereas<00:17:14.197> 5,<00:17:14.502> 6<00:17:14.758> years<00:17:15.079> ago<00:17:15.480> we<00:17:15.640> would,<00:17:16.746> we'd<00:17:16.987> unfortunately<00:17:17.564> get<00:17:17.741> into<00:17:17.885> like<00:17:18.125> bidding<00:17:18.430> wars<00:17:19.312> with<00:17:19.408> like<00:17:19.569> Google<00:17:19.969> and<00:17:20.130> Facebook<00:17:20.787> or<00:17:20.916> hedge<00:17:21.172> funds<00:17:21.493> and<00:17:21.637> things<00:17:21.878> like<00:17:22.118> that.

00:17:22.519 --> 00:17:35.644
And<00:17:23.257> we<00:17:23.337> had<00:17:23.481> to<00:17:23.577> really<00:17:23.882> pitch<00:17:24.139> people<00:17:24.443> on<00:17:24.539> the<00:17:24.684> vision<00:17:25.004> of<00:17:25.101> like<00:17:25.646> Well,<00:17:25.822> you<00:17:25.902> get<00:17:26.046> to<00:17:26.207> work<00:17:26.447> with<00:17:26.559> these<00:17:26.767> teams<00:17:27.104> because<00:17:27.408> they're<00:17:27.568> our<00:17:27.969> customers,<00:17:29.091> but<00:17:30.052> at<00:17:30.132> the<00:17:30.228> end<00:17:30.324> of<00:17:30.453> the<00:17:30.533> day,<00:17:31.190> like,<00:17:31.975> we<00:17:32.071> can't<00:17:32.391> match<00:17:32.712> them<00:17:33.112> in<00:17:33.273> certain<00:17:33.513> ways<00:17:33.801> because<00:17:34.298> the<00:17:34.394> bidding<00:17:34.635> wars<00:17:34.875> were<00:17:34.955> getting<00:17:35.195> insane.

00:17:36.381 --> 00:17:58.322
And<00:17:36.477> now<00:17:36.798> that<00:17:37.118> some<00:17:37.422> of<00:17:37.519> that<00:17:37.679> is<00:17:37.823> retracted,<00:17:38.656> some<00:17:38.897> of<00:17:39.041> this<00:17:39.265> idea<00:17:39.842> of,<00:17:40.323> well,<00:17:40.451> as<00:17:40.563> long<00:17:40.739> as<00:17:40.819> you<00:17:40.947> go<00:17:41.044> to<00:17:41.124> a<00:17:41.140> FAANG<00:17:41.524> company,<00:17:42.005> like,<00:17:42.229> you're<00:17:42.325> set<00:17:42.582> for<00:17:42.726> life,<00:17:43.127> they'll<00:17:43.223> never<00:17:43.527> do<00:17:43.687> layoffs,<00:17:44.248> the<00:17:44.889> price<00:17:45.193> will<00:17:45.514> always<00:17:45.754> go<00:17:45.995> up,<00:17:46.571> like,<00:17:46.796> it's<00:17:46.956> created<00:17:47.373> a<00:17:47.389> little<00:17:47.549> bit<00:17:47.757> more<00:17:47.853> liquidity<00:17:48.414> in<00:17:48.494> the<00:17:48.574> market,<00:17:49.119> which<00:17:49.295> I<00:17:49.359> think<00:17:49.536> is<00:17:49.680> good<00:17:49.824> for<00:17:50.000> everyone,<00:17:51.218> but<00:17:51.362> allows<00:17:51.763> us<00:17:51.939> to<00:17:52.500> attract<00:17:52.884> talent<00:17:53.301> that<00:17:54.263> wants<00:17:54.584> to<00:17:54.664> be<00:17:54.825> able<00:17:54.985> to<00:17:55.145> do<00:17:55.370> really<00:17:55.546> good<00:17:55.787> things,<00:17:56.172> wants<00:17:56.349> to<00:17:56.509> learn<00:17:56.734> a<00:17:56.814> lot,<00:17:57.215> wants<00:17:57.375> to<00:17:57.471> move<00:17:57.696> really<00:17:57.937> quickly.

00:17:59.060 --> 00:18:06.663
Um,<00:18:00.118> but<00:18:00.263> then<00:18:00.423> we<00:18:00.519> can<00:18:00.664> also<00:18:01.065> be<00:18:01.466> a<00:18:01.546> lot<00:18:01.674> more<00:18:01.883> competitive<00:18:02.765> because,<00:18:03.487> um,<00:18:04.353> some<00:18:04.674> of<00:18:04.835> the,<00:18:05.011> the<00:18:05.091> markets<00:18:05.476> retracted<00:18:05.974> a<00:18:06.038> bit<00:18:06.150> in<00:18:06.278> other<00:18:06.439> ways.

00:18:07.321 --> 00:18:13.433
Uh,<00:18:07.738> but<00:18:07.883> we<00:18:08.043> do<00:18:08.220> rely<00:18:08.621> very<00:18:08.941> heavily<00:18:09.567> on<00:18:10.449> the<00:18:10.530> network<00:18:10.963> of<00:18:11.091> the<00:18:11.187> people<00:18:11.989> who<00:18:12.471> are<00:18:12.551> already<00:18:12.936> in<00:18:13.016> the<00:18:13.096> company.

00:18:14.139 --> 00:18:23.753
Um,<00:18:14.716> a<00:18:14.781> lot<00:18:14.957> of<00:18:15.021> them<00:18:15.182> came<00:18:15.438> from<00:18:15.583> SigOpt,<00:18:16.160> but,<00:18:16.802> uh,<00:18:17.363> at<00:18:17.444> this<00:18:17.604> point<00:18:17.781> I<00:18:17.845> think<00:18:18.085> less<00:18:18.310> than<00:18:18.470> half<00:18:18.663> of<00:18:18.727> the<00:18:18.807> team<00:18:19.064> is<00:18:19.192> from<00:18:19.385> SigOpt,<00:18:19.866> and<00:18:19.946> so<00:18:20.813> Everybody<00:18:21.198> knows<00:18:21.391> someone,<00:18:21.937> everybody<00:18:22.403> has<00:18:22.580> gone<00:18:22.885> on<00:18:23.142> and<00:18:23.207> done<00:18:23.399> something.

00:18:23.865 --> 00:18:36.412
Some<00:18:24.042> of<00:18:24.090> the<00:18:24.187> people<00:18:24.588> who<00:18:24.733> were<00:18:24.910> at<00:18:25.231> SigOpt<00:18:25.713> and<00:18:25.793> rejoined<00:18:26.114> this<00:18:26.259> company<00:18:26.741> had<00:18:27.544> gone<00:18:27.785> on<00:18:27.898> to<00:18:28.685> form<00:18:28.845> their<00:18:28.990> own<00:18:29.649> YC<00:18:29.729> company<00:18:30.291> and<00:18:30.452> then<00:18:30.613> now<00:18:30.870> are<00:18:31.014> back<00:18:31.255> with<00:18:31.480> us<00:18:31.801> or<00:18:32.139> gone<00:18:32.380> on<00:18:32.605> to<00:18:32.765> work<00:18:33.006> at<00:18:33.183> Google<00:18:34.211> or<00:18:34.693> Facebook<00:18:35.111> or<00:18:35.175> whatever<00:18:35.432> and<00:18:35.593> now<00:18:35.641> have<00:18:35.898> come<00:18:36.075> back.

00:18:36.476 --> 00:18:45.039
And<00:18:36.573> so<00:18:37.681> that<00:18:37.842> cross-pollination<00:18:38.709> helps<00:18:39.031> with<00:18:39.127> just<00:18:39.352> like<00:18:39.529> better<00:18:39.770> ideas,<00:18:40.412> but<00:18:40.509> it<00:18:40.637> also<00:18:41.119> brings<00:18:41.457> in<00:18:41.537> a<00:18:41.617> much,<00:18:42.420> much<00:18:42.661> broader<00:18:43.063> network<00:18:43.545> that<00:18:44.091> we're<00:18:44.268> able<00:18:44.541> to<00:18:44.589> leverage.

00:18:45.874 --> 00:18:47.481
And<00:18:46.196> investors<00:18:46.758> as<00:18:46.903> well<00:18:47.224> can<00:18:47.320> help.

00:18:47.706 --> 00:18:48.622
Help<00:18:48.268> with<00:18:48.365> that.

00:18:48.686 --> 00:18:57.971
We<00:18:48.766> get<00:18:50.051> a<00:18:50.132> lot<00:18:50.308> of<00:18:50.437> referrals<00:18:51.015> to<00:18:51.112> people<00:18:51.417> who<00:18:51.497> are<00:18:51.577> like<00:18:52.204> excited<00:18:52.686> about<00:18:52.927> the<00:18:53.023> idea<00:18:53.248> of<00:18:53.345> not<00:18:53.505> only<00:18:53.762> the<00:18:53.891> quality<00:18:54.212> of<00:18:54.244> our<00:18:54.389> investors,<00:18:54.951> but<00:18:55.658> second-time<00:18:56.236> founders<00:18:57.425> doing<00:18:57.602> it<00:18:57.746> again.

00:18:58.967 --> 00:19:00.429
What's<00:18:59.208> it<00:18:59.288> like<00:18:59.513> in<00:18:59.610> Silicon<00:19:00.011> Valley<00:19:00.333> now?

00:19:00.493 --> 00:19:08.413
Like,<00:19:00.959> I<00:19:01.152> know<00:19:01.377> there<00:19:01.473> was<00:19:01.682> like<00:19:01.843> this<00:19:02.084> exodus<00:19:02.983> or,<00:19:03.610> you<00:19:03.642> know,<00:19:03.851> supposed<00:19:04.413> exodus<00:19:04.815> to<00:19:04.847> like<00:19:05.056> places<00:19:05.457> like<00:19:05.554> Miami<00:19:06.437> or<00:19:06.582> Austin<00:19:07.305> and<00:19:07.642> these<00:19:07.803> types<00:19:08.044> of<00:19:08.108> places.

00:19:09.007 --> 00:19:14.582
It<00:19:09.088> still<00:19:09.313> seems<00:19:09.554> like<00:19:09.714> there's<00:19:09.891> no<00:19:10.052> better<00:19:10.293> place<00:19:10.598> in<00:19:10.678> the<00:19:10.775> entire<00:19:11.240> world<00:19:11.562> to<00:19:11.722> have<00:19:11.915> it<00:19:12.686> run<00:19:12.863> a<00:19:12.879> startup,<00:19:13.506> particularly<00:19:13.971> if<00:19:14.052> you're<00:19:14.164> an<00:19:14.293> AI.

00:19:15.032 --> 00:19:16.362
Got<00:19:15.160> to<00:19:15.192> be<00:19:15.352> the<00:19:15.497> epicenter<00:19:15.913> of<00:19:16.074> that.

00:19:17.020 --> 00:19:22.327
Uh,<00:19:17.421> but<00:19:17.821> how<00:19:17.998> have<00:19:18.158> you<00:19:18.559> kind<00:19:18.799> of<00:19:18.960> managed<00:19:19.441> the<00:19:19.601> ups<00:19:19.842> and<00:19:20.002> downs<00:19:20.724> of<00:19:21.221> Silicon<00:19:21.686> Valley<00:19:21.862> itself?

00:19:23.129 --> 00:19:23.337
Yeah.

00:19:23.594 --> 00:19:32.155
Um,<00:19:24.091> I,<00:19:24.315> I<00:19:24.395> guess<00:19:24.716> I<00:19:24.796> was<00:19:24.908> part<00:19:25.117> of<00:19:25.197> that<00:19:25.357> exodus<00:19:26.031> to<00:19:26.175> a<00:19:26.191> certain<00:19:26.480> extent<00:19:26.880> after<00:19:27.073> selling<00:19:27.618> the<00:19:27.762> company,<00:19:28.404> uh,<00:19:28.885> the<00:19:29.446> first<00:19:29.606> company,<00:19:29.863> SigOps,<00:19:30.408> to,<00:19:30.889> uh,<00:19:30.985> Intel<00:19:31.386> in<00:19:31.466> 2020.

00:19:32.633 --> 00:19:35.314
Uh,<00:19:32.653> I<00:19:32.829> moved<00:19:33.069> up<00:19:33.150> to<00:19:33.470> Portland,<00:19:33.855> Oregon,<00:19:34.224> which<00:19:34.448> is<00:19:34.577> where<00:19:34.673> I<00:19:34.833> grew<00:19:35.138> up.

00:19:35.939 --> 00:19:38.296
Um,<00:19:36.821> and<00:19:36.901> it<00:19:37.062> was<00:19:37.398> middle<00:19:37.575> of<00:19:37.639> the<00:19:37.783> pandemic.

00:19:38.377 --> 00:19:43.299
So<00:19:38.537> it<00:19:38.585> was<00:19:38.681> like,<00:19:38.906> I've<00:19:39.002> got<00:19:39.387> two<00:19:39.483> young<00:19:39.723> kids,<00:19:40.589> might<00:19:40.749> as<00:19:40.846> well,<00:19:41.407> uh,<00:19:41.567> be<00:19:41.712> close<00:19:41.968> to<00:19:42.112> family<00:19:42.513> and<00:19:42.674> that<00:19:42.786> sort<00:19:43.010> of<00:19:43.171> thing.

00:19:43.956 --> 00:19:46.313
But<00:19:44.004> like<00:19:44.197> you<00:19:44.293> said,<00:19:44.854> Silicon<00:19:45.174> Valley<00:19:45.720> is<00:19:45.816> the<00:19:45.960> epicenter.

00:19:46.393 --> 00:19:49.791
So<00:19:46.713> no<00:19:46.794> matter<00:19:47.002> how<00:19:47.483> much<00:19:47.659> you<00:19:47.739> try<00:19:47.900> to<00:19:48.044> get<00:19:48.156> away,<00:19:48.781> it<00:19:49.022> pulls<00:19:49.262> you<00:19:49.342> back<00:19:49.647> in.

00:19:50.304 --> 00:19:53.318
And<00:19:50.384> so<00:19:50.705> moved<00:19:50.929> back<00:19:51.122> down<00:19:51.282> to<00:19:51.410> Palo<00:19:51.667> Alto<00:19:51.923> about<00:19:52.084> 4<00:19:52.388> weeks<00:19:52.629> ago,<00:19:52.885> actually.

00:19:53.955 --> 00:19:54.440
Uh,<00:19:53.975> oh<00:19:54.232> really?

00:19:54.857 --> 00:19:54.985
Yeah.

00:19:55.354 --> 00:19:58.320
So<00:19:55.530> it's,<00:19:55.675> it's<00:19:56.059> been<00:19:56.172> an<00:19:56.300> intense,<00:19:56.861> uh,<00:19:57.678> couple<00:19:57.919> of<00:19:57.999> months.

00:19:58.945 --> 00:20:11.583
Um,<00:20:00.307> but<00:20:00.404> I'd<00:20:00.564> say<00:20:01.045> one<00:20:01.430> interesting<00:20:01.766> shift<00:20:02.087> that<00:20:02.231> we've<00:20:02.391> seen<00:20:02.872> is<00:20:03.770> it,<00:20:04.411> while<00:20:04.812> it<00:20:04.972> is<00:20:05.277> still<00:20:05.694> very<00:20:05.950> much,<00:20:06.255> I'd<00:20:06.399> say<00:20:06.511> the<00:20:06.720> epicenter<00:20:07.120> and<00:20:07.217> there's<00:20:07.377> a<00:20:07.521> lot<00:20:07.714> of<00:20:07.778> really<00:20:08.082> great<00:20:08.259> stuff<00:20:08.563> happening,<00:20:09.060> there's<00:20:09.220> a<00:20:09.237> lot<00:20:09.461> of<00:20:09.782> this<00:20:09.960> is<00:20:10.171> obviously<00:20:10.512> where<00:20:10.593> the<00:20:10.739> investors<00:20:11.307> are.

00:20:11.794 --> 00:20:13.968
There's<00:20:11.956> a<00:20:11.973> lot<00:20:12.135> of<00:20:12.232> events,<00:20:12.768> there's<00:20:12.946> a<00:20:12.962> lot<00:20:13.125> of<00:20:13.189> things<00:20:13.433> going<00:20:13.741> on.

00:20:15.610 --> 00:20:17.007
Talent<00:20:15.867> is<00:20:16.268> more<00:20:16.381> dispersed<00:20:16.847> now.

00:20:17.168 --> 00:20:20.076
And<00:20:17.313> so<00:20:17.473> we<00:20:17.634> are<00:20:19.000> mostly<00:20:19.402> remote<00:20:19.659> company.

00:20:20.141 --> 00:20:26.316
We<00:20:20.607> do<00:20:20.767> have<00:20:20.944> hubs<00:20:21.346> and<00:20:21.426> small<00:20:21.828> offices<00:20:23.017> both<00:20:23.194> in<00:20:23.756> the<00:20:23.852> Bay<00:20:24.061> Area,<00:20:24.559> but<00:20:24.784> also<00:20:25.122> New<00:20:25.267> York<00:20:25.622> and<00:20:25.751> Toronto.

00:20:27.785 --> 00:20:32.433
And<00:20:27.882> it's<00:20:28.091> really<00:20:28.350> about<00:20:29.963> having<00:20:30.512> the<00:20:30.609> right<00:20:31.255> people<00:20:31.739> around<00:20:31.981> the<00:20:32.078> table.

00:20:32.626 --> 00:20:34.963
And<00:20:32.707> those<00:20:33.014> people<00:20:33.579> have<00:20:33.982> a<00:20:33.998> lot<00:20:34.240> more<00:20:34.471> options.

00:20:35.082 --> 00:20:36.252
They<00:20:35.234> have<00:20:35.421> a<00:20:35.438> lot<00:20:35.760> more<00:20:35.930> mobility.

00:20:37.116 --> 00:20:44.630
Um,<00:20:37.613> and<00:20:37.693> a<00:20:37.773> lot<00:20:37.933> of<00:20:37.997> the<00:20:38.093> technology<00:20:38.814> and<00:20:38.894> the<00:20:39.039> way<00:20:39.135> that<00:20:39.295> people<00:20:39.615> are<00:20:39.696> just<00:20:39.936> used<00:20:40.176> to<00:20:40.320> working<00:20:40.721> has,<00:20:41.362> um,<00:20:42.099> has<00:20:42.259> become<00:20:42.980> more<00:20:43.461> feasible<00:20:43.861> in<00:20:44.006> a<00:20:44.022> remote<00:20:44.358> setting.

00:20:44.743 --> 00:20:50.479
I<00:20:44.759> mean,<00:20:44.967> we're<00:20:45.127> having<00:20:45.448> this<00:20:45.624> interview<00:20:46.169> right<00:20:46.425> now<00:20:46.665> remote,<00:20:47.386> whereas<00:20:47.691> 10<00:20:47.867> years<00:20:48.107> ago<00:20:48.508> I<00:20:48.524> would<00:20:48.652> have<00:20:48.748> probably<00:20:49.133> had<00:20:49.213> to<00:20:49.309> fly<00:20:49.549> to<00:20:49.613> your<00:20:49.854> office<00:20:50.126> or<00:20:50.190> whatever.

00:20:50.671 --> 00:20:52.209
Yeah,<00:20:50.975> we<00:20:51.136> couldn't<00:20:51.392> have<00:20:51.536> done<00:20:51.648> it<00:20:51.808> 10<00:20:52.033> years.

00:20:52.337 --> 00:20:55.654
It's<00:20:52.497> actually<00:20:52.754> incredible,<00:20:54.020> uh,<00:20:54.276> that<00:20:54.420> we<00:20:54.516> can<00:20:54.676> do<00:20:54.773> something<00:20:55.141> like<00:20:55.381> this<00:20:55.622> now.

00:20:56.439 --> 00:20:58.137
What<00:20:56.519> do<00:20:56.663> you<00:20:56.743> think,<00:20:57.064> how<00:20:57.160> do<00:20:57.224> you<00:20:57.320> build<00:20:57.496> a<00:20:57.625> culture?

00:20:58.666 --> 00:21:05.187
Uh,<00:20:59.483> and<00:20:59.884> maybe<00:21:00.204> talk<00:21:00.445> about<00:21:00.685> a<00:21:00.701> little<00:21:00.941> bit<00:21:01.149> how<00:21:01.310> you<00:21:01.406> build<00:21:01.566> it<00:21:01.662> the<00:21:01.790> first<00:21:02.031> time<00:21:02.207> and<00:21:02.303> how<00:21:02.511> you've<00:21:02.928> maybe<00:21:03.313> think,<00:21:03.553> thought<00:21:03.665> about<00:21:03.953> building<00:21:04.274> it<00:21:04.514> this<00:21:04.690> time.

00:21:05.876 --> 00:21:08.763
And<00:21:06.037> how<00:21:06.133> do<00:21:06.197> you<00:21:06.357> do<00:21:06.534> it<00:21:06.678> in<00:21:06.774> the,<00:21:07.255> you<00:21:07.336> know,<00:21:07.432> mostly<00:21:07.897> remote<00:21:08.218> environment?

00:21:09.725 --> 00:21:09.965
Yeah.

00:21:10.286 --> 00:21:17.727
Um,<00:21:10.719> so<00:21:11.409> one<00:21:11.585> of<00:21:11.649> the<00:21:11.745> things,<00:21:12.371> uh,<00:21:12.467> that<00:21:12.691> I'm<00:21:12.868> a<00:21:12.964> strong<00:21:13.333> proponent<00:21:13.750> of<00:21:13.926> is<00:21:14.054> like<00:21:14.391> culture<00:21:15.129> needs<00:21:15.514> to<00:21:15.674> be<00:21:16.075> one<00:21:16.235> of<00:21:16.716> the<00:21:16.797> pillars<00:21:17.278> of<00:21:17.358> the<00:21:17.438> company.

00:21:17.903 --> 00:21:24.381
I<00:21:17.935> think<00:21:18.641> you<00:21:18.705> have<00:21:18.961> strategy,<00:21:20.004> execution,<00:21:20.805> and<00:21:20.966> culture,<00:21:21.607> but<00:21:21.751> you<00:21:21.848> need<00:21:22.008> to<00:21:22.088> treat<00:21:22.473> it<00:21:22.794> as<00:21:23.596> a<00:21:23.676> peer<00:21:24.012> to<00:21:24.173> those.

00:21:25.119 --> 00:21:31.934
Where<00:21:25.359> you're<00:21:25.520> going,<00:21:26.241> how<00:21:26.418> you're<00:21:26.578> going<00:21:26.674> to<00:21:26.803> get<00:21:26.979> there,<00:21:27.380> and<00:21:27.604> who<00:21:27.701> you<00:21:27.925> are,<00:21:28.807> because<00:21:29.705> the<00:21:29.849> culture<00:21:30.266> ends<00:21:30.491> up<00:21:30.587> permeating<00:21:31.453> all<00:21:31.565> of<00:21:31.709> that.

00:21:32.046 --> 00:21:37.984
It's<00:21:32.896> how<00:21:33.057> big<00:21:33.169> of<00:21:33.249> a<00:21:33.314> swing<00:21:33.635> you're<00:21:33.795> going<00:21:33.972> to<00:21:34.020> take,<00:21:34.357> how<00:21:34.533> effective<00:21:35.079> you<00:21:35.304> are<00:21:36.363> in<00:21:36.524> your<00:21:36.668> ability<00:21:37.085> to<00:21:37.230> actually<00:21:37.647> get<00:21:37.872> there.

00:21:38.690 --> 00:21:43.794
And<00:21:38.770> so<00:21:38.995> I,<00:21:39.653> I<00:21:39.733> think<00:21:39.910> it's<00:21:40.054> something<00:21:40.359> that<00:21:40.472> needs<00:21:40.680> to<00:21:40.761> be<00:21:40.857> extremely<00:21:41.322> important,<00:21:41.804> but<00:21:41.900> then<00:21:42.061> it<00:21:42.205> also<00:21:42.398> needs<00:21:42.622> to<00:21:42.703> be<00:21:42.847> extremely<00:21:43.489> active.

00:21:44.163 --> 00:21:47.887
It's<00:21:44.789> who<00:21:44.933> you<00:21:45.030> recruit,<00:21:45.832> who<00:21:45.913> you<00:21:46.137> reward,<00:21:46.699> and<00:21:46.859> who<00:21:46.940> you<00:21:47.036> end<00:21:47.132> up<00:21:47.293> releasing.

00:21:48.705 --> 00:21:53.151
Almost<00:21:50.711> to<00:21:50.888> use<00:21:51.209> another<00:21:51.514> analogy,<00:21:51.915> I<00:21:52.011> think<00:21:52.172> of<00:21:52.252> it<00:21:52.332> almost<00:21:52.573> like<00:21:52.734> a<00:21:52.750> tree.

00:21:53.311 --> 00:21:59.218
It'll<00:21:53.841> grow<00:21:54.419> on<00:21:54.499> its<00:21:55.606> own<00:21:56.008> with<00:21:57.003> sunlight<00:21:57.468> and<00:21:57.629> rainwater,<00:21:58.351> like<00:21:58.447> it'll<00:21:58.672> do<00:21:58.832> whatever.

00:21:59.314 --> 00:22:03.021
But<00:21:59.410> if<00:21:59.555> you<00:22:00.020> actively<00:22:00.502> tend<00:22:00.758> to<00:22:00.983> it,<00:22:01.224> you<00:22:01.256> maybe<00:22:01.770> make<00:22:01.898> it<00:22:02.026> grow<00:22:02.123> a<00:22:02.139> little<00:22:02.379> bit<00:22:02.524> faster.

00:22:03.085 --> 00:22:08.509
If<00:22:03.246> you<00:22:03.406> prune<00:22:03.791> it,<00:22:03.872> you<00:22:03.904> might<00:22:04.112> be<00:22:04.160> able<00:22:04.289> to<00:22:04.385> make<00:22:04.594> it<00:22:04.690> grow<00:22:04.867> a<00:22:04.931> specific<00:22:05.412> way<00:22:05.797> if<00:22:05.878> you're<00:22:05.974> trying<00:22:06.214> to<00:22:06.295> make<00:22:06.535> it<00:22:07.338> fruit<00:22:07.739> or<00:22:07.819> whatever<00:22:08.140> it<00:22:08.285> is.

00:22:08.782 --> 00:22:17.094
There's<00:22:08.975> specific<00:22:09.344> things<00:22:09.649> that<00:22:09.745> you<00:22:09.889> can<00:22:10.066> do<00:22:10.387> to<00:22:10.547> encourage<00:22:10.948> specific<00:22:11.815> aspects<00:22:12.232> of<00:22:12.377> it,<00:22:12.858> but<00:22:12.954> that<00:22:13.195> needs<00:22:13.436> to<00:22:13.596> be<00:22:14.479> a<00:22:14.784> constant<00:22:16.629> process.

00:22:17.913 --> 00:22:28.889
And<00:22:18.009> so<00:22:18.795> we<00:22:18.876> thought<00:22:19.068> about<00:22:19.213> this<00:22:20.079> very<00:22:20.897> carefully<00:22:21.684> as<00:22:22.021> we<00:22:22.181> built<00:22:22.358> up<00:22:22.502> the<00:22:22.582> team<00:22:23.048> in<00:22:23.304> SigOpt,<00:22:25.070> trying<00:22:25.375> to<00:22:25.471> make<00:22:25.647> sure<00:22:25.936> that<00:22:26.097> we<00:22:26.273> grew<00:22:26.819> organically<00:22:27.717> and<00:22:28.119> intentionally.

00:22:30.710 --> 00:22:39.562
And<00:22:30.855> then<00:22:31.517> like<00:22:31.759> everybody<00:22:32.163> else<00:22:32.405> in<00:22:32.469> the<00:22:32.550> world<00:22:32.873> had<00:22:33.535> that<00:22:34.180> whiplash<00:22:34.745> of,<00:22:34.988> okay,<00:22:35.875> we're<00:22:36.021> going<00:22:36.101> to<00:22:36.198> go<00:22:36.360> into<00:22:36.505> lockdown<00:22:37.005> for<00:22:37.167> 2<00:22:37.328> or<00:22:37.409> 3<00:22:37.716> weeks<00:22:38.055> and<00:22:38.200> we'll<00:22:38.297> see<00:22:38.523> everybody<00:22:39.187> again.

00:22:39.611 --> 00:22:43.320
And<00:22:39.741> then<00:22:40.164> 2<00:22:40.326> years<00:22:40.652> later<00:22:41.124> we<00:22:41.205> had<00:22:41.368> sold<00:22:41.710> to<00:22:42.588> Intel.

00:22:46.071 --> 00:22:51.251
But<00:22:46.216> I<00:22:46.232> think<00:22:46.474> that<00:22:47.120> strong<00:22:47.523> nucleus<00:22:48.088> allowed<00:22:48.556> us<00:22:48.734> to<00:22:48.895> then<00:22:49.540> be<00:22:49.621> able<00:22:50.041> to<00:22:50.186> withstand<00:22:50.606> something<00:22:50.977> like<00:22:51.154> that.

00:22:51.316 --> 00:23:02.845
We<00:22:51.477> actually<00:22:51.703> saw<00:22:51.945> an<00:22:52.042> increase<00:22:52.526> in<00:22:52.671> productivity<00:22:53.962> when<00:22:54.124> we<00:22:54.220> went<00:22:54.350> remote<00:22:54.835> because<00:22:55.470> Everybody<00:22:55.822> knew<00:22:56.383> the<00:22:56.608> role,<00:22:56.929> the<00:22:57.025> responsibility,<00:22:57.987> how<00:22:58.147> they<00:22:58.292> were<00:22:58.388> going<00:22:58.468> to<00:22:58.612> execute,<00:22:59.109> how<00:22:59.270> they<00:22:59.414> could<00:22:59.574> communicate,<00:23:00.296> how<00:23:00.392> they<00:23:00.472> could<00:23:00.665> disagree<00:23:01.274> and<00:23:02.076> collaborate.

00:23:04.160 --> 00:23:14.871
And<00:23:04.256> so<00:23:04.417> it<00:23:04.465> was<00:23:04.641> really<00:23:04.946> handy<00:23:05.266> the<00:23:05.363> second<00:23:05.699> time<00:23:05.940> around<00:23:06.309> to<00:23:06.389> start<00:23:06.870> with<00:23:07.046> a<00:23:07.062> somewhat<00:23:07.447> similar<00:23:07.848> nucleus,<00:23:09.115> obviously<00:23:09.531> allowing<00:23:10.077> it<00:23:10.173> to<00:23:10.333> grow<00:23:10.750> so<00:23:10.894> that<00:23:11.039> we<00:23:11.135> didn't<00:23:11.375> have<00:23:11.856> a<00:23:12.033> separate<00:23:12.418> culture<00:23:12.818> for<00:23:12.979> the<00:23:13.540> old<00:23:13.700> SIGOPT<00:23:14.021> people<00:23:14.262> or<00:23:14.342> whatever<00:23:14.646> it<00:23:14.727> is.

00:23:14.983 --> 00:23:19.521
But<00:23:15.208> it's<00:23:16.250> way<00:23:16.506> easier<00:23:16.827> to<00:23:16.971> grow<00:23:17.164> a<00:23:17.228> tree<00:23:17.629> from<00:23:17.869> a<00:23:17.885> sapling<00:23:18.511> than<00:23:18.671> it<00:23:18.751> is<00:23:18.911> from<00:23:19.088> a<00:23:19.136> seed.

00:23:19.858 --> 00:23:23.401
And<00:23:20.675> that's<00:23:21.220> really<00:23:21.621> allowed<00:23:21.878> us<00:23:22.038> to<00:23:22.615> do<00:23:22.696> that<00:23:22.984> well,<00:23:23.161> I<00:23:23.241> think.

00:23:24.428 --> 00:23:27.620
What<00:23:24.508> did<00:23:24.668> you<00:23:24.748> think<00:23:24.909> of<00:23:24.989> Paul<00:23:25.230> Graham's<00:23:25.567> essay,<00:23:26.272> uh,<00:23:26.529> the<00:23:26.609> Founder<00:23:27.026> Mode<00:23:27.395> one?

00:23:28.294 --> 00:23:43.148
Because<00:23:28.935> previously,<00:23:29.802> a<00:23:29.978> little<00:23:30.139> bit,<00:23:30.684> uh,<00:23:31.342> before<00:23:31.887> you<00:23:31.967> were<00:23:32.047> saying<00:23:32.288> like,<00:23:32.449> hey,<00:23:32.625> my<00:23:32.785> job<00:23:33.042> is<00:23:33.170> to<00:23:33.267> kind<00:23:33.427> of<00:23:33.507> get<00:23:33.652> out<00:23:33.748> of<00:23:33.812> the<00:23:33.892> way<00:23:34.390> and<00:23:35.176> kind<00:23:35.272> of<00:23:35.352> let<00:23:35.577> that—<00:23:35.753> and<00:23:35.833> then<00:23:36.234> like<00:23:36.475> Paul<00:23:36.796> was,<00:23:37.421> and<00:23:37.839> you<00:23:37.903> know,<00:23:38.079> Brian<00:23:38.560> over<00:23:38.881> at<00:23:39.042> Airbnb,<00:23:39.587> it<00:23:39.667> seems<00:23:39.924> like<00:23:40.084> they're<00:23:40.261> saying<00:23:40.550> like,<00:23:41.207> you<00:23:41.304> need<00:23:41.480> to<00:23:41.528> be<00:23:41.689> in<00:23:41.769> the<00:23:41.833> middle<00:23:42.090> of<00:23:42.234> everything<00:23:42.651> if<00:23:42.892> you<00:23:42.972> want.

00:23:43.229 --> 00:23:44.704
So<00:23:43.437> how,<00:23:43.678> how<00:23:43.774> do<00:23:43.854> you<00:23:43.999> think<00:23:44.255> about<00:23:44.560> that?

00:23:45.474 --> 00:23:54.121
Yeah,<00:23:45.940> I<00:23:46.004> think<00:23:46.902> Paul<00:23:47.223> is<00:23:47.399> a<00:23:47.464> great<00:23:47.768> way<00:23:47.945> of<00:23:48.121> kind<00:23:48.282> of<00:23:48.426> dialing<00:23:48.811> up<00:23:48.907> the<00:23:48.988> contrast<00:23:49.709> in,<00:23:50.014> in<00:23:50.191> things,<00:23:50.736> um,<00:23:50.913> which<00:23:51.073> is<00:23:51.137> great<00:23:51.378> because<00:23:51.618> obviously<00:23:51.955> we're<00:23:52.052> still<00:23:52.276> talking<00:23:52.677> about<00:23:52.982> it,<00:23:53.223> uh,<00:23:53.495> months<00:23:53.800> later.

00:23:55.685 --> 00:24:02.903
I<00:23:55.701> think<00:23:55.865> it's<00:23:56.981> like<00:23:57.637> all<00:23:57.801> things,<00:23:58.080> like<00:23:58.294> all<00:23:58.458> values,<00:24:01.309> there's<00:24:01.953> no<00:24:02.356> truisms.

00:24:03.338 --> 00:24:13.848
Like<00:24:03.563> you<00:24:03.595> need<00:24:03.740> to<00:24:03.821> be<00:24:03.965> able<00:24:04.062> to<00:24:04.143> do<00:24:04.304> one<00:24:04.465> thing<00:24:04.786> or<00:24:04.883> the<00:24:05.092> other<00:24:05.495> and<00:24:05.897> either<00:24:06.155> one<00:24:06.396> can<00:24:06.557> be<00:24:06.718> successful,<00:24:07.443> but<00:24:07.539> it's<00:24:07.684> just<00:24:07.845> about<00:24:08.087> making<00:24:08.328> sure<00:24:08.570> that<00:24:08.715> you<00:24:08.892> do<00:24:09.536> what<00:24:09.681> you<00:24:09.858> plan<00:24:10.180> to<00:24:10.260> do<00:24:10.470> well<00:24:10.743> instead<00:24:11.081> of<00:24:11.210> trying<00:24:11.467> to<00:24:11.547> like<00:24:11.869> cut<00:24:12.110> to<00:24:12.207> the<00:24:12.336> middle<00:24:12.915> and<00:24:12.979> then<00:24:13.076> failing<00:24:13.478> at<00:24:13.639> both.

00:24:15.006 --> 00:24:24.289
And<00:24:15.103> so<00:24:15.392> I'm<00:24:15.505> a<00:24:15.633> big<00:24:15.810> proponent<00:24:16.599> of<00:24:18.770> a<00:24:18.931> good<00:24:19.189> CEO<00:24:20.282> is<00:24:20.378> someone<00:24:21.067> who<00:24:21.869> isn't<00:24:22.189> required<00:24:22.750> for<00:24:22.846> everything.

00:24:24.674 --> 00:24:36.889
And<00:24:25.171> this<00:24:25.572> might<00:24:25.732> be<00:24:26.533> more<00:24:26.694> controversial,<00:24:27.239> but<00:24:27.351> it's<00:24:27.559> like,<00:24:27.896> I<00:24:27.960> remember<00:24:29.643> when<00:24:29.755> I<00:24:29.803> was<00:24:29.852> starting<00:24:30.284> out<00:24:30.381> in<00:24:30.445> my<00:24:30.621> career,<00:24:31.342> I'd<00:24:31.887> think<00:24:32.064> of<00:24:32.304> like,<00:24:32.609> there's<00:24:33.106> this<00:24:33.571> caricature<00:24:34.148> of<00:24:34.308> the<00:24:34.372> CEO<00:24:34.933> who's<00:24:35.109> on<00:24:35.190> the<00:24:35.334> golf<00:24:35.590> course<00:24:36.007> or<00:24:36.071> something<00:24:36.392> like<00:24:36.632> that.

00:24:36.953 --> 00:24:39.662
And<00:24:37.274> everybody<00:24:37.674> else<00:24:37.851> is<00:24:37.995> doing<00:24:38.155> the<00:24:38.299> hard<00:24:38.556> work<00:24:38.812> or<00:24:38.877> something<00:24:39.261> like<00:24:39.438> that.

00:24:39.742 --> 00:24:45.112
But<00:24:39.838> it's<00:24:39.999> like,<00:24:40.800> in<00:24:40.944> my<00:24:41.185> mind,<00:24:41.762> that's<00:24:41.986> actually<00:24:42.323> the<00:24:42.483> best<00:24:42.724> CEO<00:24:43.205> because<00:24:43.589> the<00:24:43.686> team<00:24:44.070> can<00:24:44.231> execute<00:24:44.631> without<00:24:44.968> them.

00:24:45.769 --> 00:24:47.902
The<00:24:45.850> team<00:24:46.074> doesn't<00:24:46.411> need<00:24:46.571> them<00:24:46.795> for<00:24:46.892> every<00:24:47.228> single<00:24:47.533> decision.

00:24:48.014 --> 00:24:54.428
They<00:24:48.158> can<00:24:48.254> be<00:24:48.479> on<00:24:48.575> the<00:24:48.655> golf<00:24:48.976> course,<00:24:49.296> which<00:24:49.440> means<00:24:49.697> they<00:24:49.777> can<00:24:49.921> obviously<00:24:50.402> not<00:24:50.579> be<00:24:50.659> on<00:24:50.739> the<00:24:50.819> golf<00:24:51.059> course<00:24:51.300> and<00:24:51.380> dive<00:24:51.701> in<00:24:51.781> when<00:24:51.925> they<00:24:52.150> absolutely<00:24:52.502> need<00:24:52.663> to,<00:24:53.144> but<00:24:53.224> they're<00:24:53.337> not<00:24:53.690> needed<00:24:54.026> day<00:24:54.187> to<00:24:54.347> day.

00:24:55.133 --> 00:25:04.181
The<00:24:55.214> CEO<00:24:55.695> that's<00:24:55.935> working<00:24:56.497> 80,<00:24:56.802> 100<00:24:57.219> hours<00:24:57.411> a<00:24:57.476> week<00:24:57.716> and<00:24:57.796> burning<00:24:58.101> themselves<00:24:58.663> out<00:24:59.545> not<00:24:59.721> only<00:24:59.962> is<00:25:00.106> setting<00:25:00.363> like<00:25:00.524> a<00:25:00.604> bad<00:25:00.909> cultural<00:25:01.342> norm,<00:25:02.192> but<00:25:02.240> they've<00:25:02.513> also<00:25:02.770> made<00:25:03.010> themselves<00:25:03.572> indispensable.

00:25:04.999 --> 00:25:06.748
And<00:25:05.080> that's<00:25:05.240> actually<00:25:05.513> a<00:25:05.657> bad<00:25:05.978> thing<00:25:06.219> in<00:25:06.299> my<00:25:06.379> mind.

00:25:07.582 --> 00:25:11.047
If<00:25:07.743> the<00:25:07.823> company<00:25:08.208> can<00:25:08.304> be<00:25:08.448> successful<00:25:09.010> without<00:25:09.347> you,<00:25:09.652> then<00:25:09.812> you've<00:25:09.908> set<00:25:10.374> up<00:25:10.550> a<00:25:10.630> good<00:25:10.775> company.

00:25:11.095 --> 00:25:12.940
You've<00:25:11.336> set<00:25:11.513> up<00:25:11.609> a<00:25:11.673> self-sustaining<00:25:12.491> system.

00:25:13.518 --> 00:25:18.539
And<00:25:13.582> that's<00:25:13.823> always<00:25:14.079> what<00:25:14.176> I'm<00:25:14.336> striving<00:25:14.801> to<00:25:14.946> do<00:25:15.138> is<00:25:15.587> how<00:25:15.700> do<00:25:15.844> I<00:25:16.646> make<00:25:16.951> this<00:25:17.192> more<00:25:17.368> and<00:25:17.512> more<00:25:17.705> self-sustaining?

00:25:18.813 --> 00:25:23.892
Which<00:25:18.960> then<00:25:19.123> allows<00:25:19.546> me<00:25:19.644> to<00:25:19.790> push<00:25:20.050> it<00:25:20.197> maybe<00:25:20.376> a<00:25:20.506> little<00:25:20.685> bit<00:25:20.783> further<00:25:21.239> or<00:25:21.320> whatever<00:25:21.662> it<00:25:21.743> is,<00:25:22.541> but<00:25:22.622> I'm<00:25:22.736> not<00:25:22.964> needed<00:25:23.306> for<00:25:23.452> the<00:25:23.534> day-to-day.

00:25:24.600 --> 00:25:25.020
That's<00:25:24.713> my<00:25:24.874> goal.

00:25:25.747 --> 00:25:26.716
Yeah,<00:25:25.925> I<00:25:26.086> agree<00:25:26.393> with<00:25:26.490> that.

00:25:26.797 --> 00:25:29.997
I<00:25:26.894> think<00:25:27.120> that's<00:25:27.848> healthier<00:25:28.429> and<00:25:28.591> more<00:25:28.753> sustainable<00:25:29.641> for<00:25:29.738> sure.

00:25:31.112 --> 00:25:31.723
What<00:25:31.225> do<00:25:31.273> you<00:25:31.498> read?

00:25:31.900 --> 00:25:33.652
What<00:25:32.077> reading<00:25:32.318> recommendations<00:25:33.025> would<00:25:33.122> you<00:25:33.202> have<00:25:33.363> for<00:25:33.523> us?

00:25:34.552 --> 00:25:42.156
Yeah,<00:25:35.035> I<00:25:35.292> am<00:25:35.597> a<00:25:35.678> huge<00:25:36.658> sci-fi<00:25:38.909> nerd<00:25:39.311> and<00:25:39.440> reader<00:25:41.786> right<00:25:42.027> now.

00:25:42.381 --> 00:25:47.428
So<00:25:42.750> I've<00:25:43.088> started<00:25:43.490> to<00:25:43.570> try<00:25:43.731> to<00:25:44.052> dovetail<00:25:45.258> a<00:25:45.274> nonfiction<00:25:45.901> and<00:25:45.997> a<00:25:46.061> fiction<00:25:46.463> book<00:25:46.608> simultaneously.

00:25:47.524 --> 00:25:58.147
So<00:25:47.733> I'm<00:25:47.910> reading<00:25:49.173> The<00:25:49.238> Ministry<00:25:49.671> of<00:25:49.816> the<00:25:49.960> Future<00:25:51.100> by<00:25:51.565> Kim<00:25:51.742> Stanley<00:25:52.544> Robinson,<00:25:54.150> which<00:25:54.374> is<00:25:55.032> maybe<00:25:55.354> a<00:25:55.370> little<00:25:56.140> not<00:25:57.280> far<00:25:57.456> enough<00:25:57.922> away.

00:25:58.243 --> 00:26:02.978
I<00:25:58.307> usually<00:25:59.126> read<00:25:59.351> like<00:25:59.463> space<00:25:59.832> operas<00:26:00.249> or<00:26:00.282> like<00:26:00.570> stuff<00:26:00.827> about<00:26:01.052> like<00:26:01.357> alien<00:26:01.598> encounters,<00:26:02.160> which<00:26:02.272> is<00:26:02.304> a<00:26:02.320> little<00:26:02.497> bit<00:26:02.641> further<00:26:02.818> away.

00:26:03.059 --> 00:26:13.701
This<00:26:03.219> one's—<00:26:03.476> I<00:26:04.182> think<00:26:04.343> one<00:26:04.487> of<00:26:04.905> the<00:26:05.322> taglines<00:26:05.868> on<00:26:07.393> the<00:26:08.356> title<00:26:08.677> page<00:26:08.998> is<00:26:09.640> this<00:26:09.800> is<00:26:09.881> the<00:26:10.025> best<00:26:10.266> science<00:26:10.667> nonfiction<00:26:11.245> I've<00:26:11.454> ever<00:26:11.646> read<00:26:11.968> because<00:26:12.224> it's<00:26:12.690> literally<00:26:13.011> happening<00:26:13.412> around<00:26:13.653> us.

00:26:13.749 --> 00:26:14.664
It's<00:26:13.894> about<00:26:14.054> climate<00:26:14.440> change.

00:26:14.712 --> 00:26:18.051
So<00:26:14.857> it's<00:26:15.258> a<00:26:15.322> little<00:26:15.499> bit<00:26:17.088> disconcerting<00:26:17.666> in<00:26:17.730> that<00:26:17.907> way.

00:26:18.147 --> 00:26:25.289
But<00:26:19.094> the<00:26:19.191> way<00:26:19.431> that—<00:26:19.816> the<00:26:19.897> thing<00:26:20.057> I<00:26:20.137> love<00:26:20.314> about<00:26:20.555> sci-fi<00:26:21.261> is<00:26:21.903> it<00:26:21.983> presents<00:26:22.320> these<00:26:22.481> really<00:26:22.721> interesting<00:26:23.123> problems<00:26:23.604> in<00:26:23.668> this<00:26:23.829> kind<00:26:23.941> of<00:26:24.070> like<00:26:24.246> thought<00:26:24.423> experiment<00:26:25.033> way.

00:26:25.915 --> 00:26:31.067
And<00:26:26.734> fundamentally,<00:26:27.392> most<00:26:27.681> sci-fi<00:26:28.114> is<00:26:28.339> optimistic<00:26:28.900> at<00:26:28.965> some<00:26:29.173> level<00:26:29.542> of<00:26:29.703> like,<00:26:30.168> we'll<00:26:30.345> figure<00:26:30.746> it<00:26:30.891> out.

00:26:31.773 --> 00:26:39.092
It<00:26:31.870> might<00:26:32.030> be<00:26:32.175> hard,<00:26:32.752> it<00:26:32.913> might<00:26:33.699> look<00:26:34.261> bad<00:26:34.662> or<00:26:34.743> a<00:26:34.759> little<00:26:34.983> bit<00:26:35.031> dystopian,<00:26:35.722> but<00:26:35.882> like<00:26:36.444> the<00:26:36.524> human<00:26:36.845> spirit<00:26:37.326> and<00:26:37.471> ingenuity<00:26:38.113> will<00:26:38.193> like<00:26:38.434> figure<00:26:38.707> it<00:26:38.915> out.

00:26:39.188 --> 00:26:40.825
And<00:26:39.333> that<00:26:39.718> to<00:26:39.814> me<00:26:39.975> is<00:26:40.119> very<00:26:40.360> motivating.

00:26:42.319 --> 00:26:47.442
So<00:26:42.465> reading<00:26:42.708> that<00:26:42.935> right<00:26:43.114> now,<00:26:43.357> even<00:26:43.519> if<00:26:43.600> it's<00:26:43.762> a<00:26:43.827> little<00:26:44.021> bit,<00:26:44.994> a<00:26:45.124> little<00:26:45.302> bit<00:26:46.113> too<00:26:46.275> close<00:26:46.518> to<00:26:46.664> home<00:26:46.940> sometimes.

00:26:48.227 --> 00:26:54.225
And<00:26:48.485> then<00:26:49.035> I'm<00:26:49.455> also<00:26:49.763> reading<00:26:50.733> Gun<00:26:50.911> Germs<00:26:51.234> and<00:26:51.331> Steel,<00:26:52.043> which<00:26:52.511> obviously<00:26:52.835> came<00:26:53.077> out<00:26:53.433> a<00:26:53.482> long<00:26:53.659> time<00:26:53.902> ago.

00:26:54.533 --> 00:27:02.894
I<00:26:55.018> recently<00:26:55.357> finished<00:26:55.681> Sapiens,<00:26:56.570> which<00:26:56.813> feels<00:26:57.120> like<00:26:57.298> it's<00:26:57.443> kind<00:26:57.556> of<00:26:57.670> like<00:27:01.826> a<00:27:01.907> continuation<00:27:02.619> of<00:27:02.700> that.

00:27:02.958 --> 00:27:04.835
And<00:27:03.039> I<00:27:03.120> realized<00:27:03.524> that<00:27:03.686> I<00:27:03.832> hadn't<00:27:04.010> read<00:27:04.236> the<00:27:04.317> original<00:27:04.721> book.

00:27:04.915 --> 00:27:09.152
So<00:27:05.061> going<00:27:05.304> back<00:27:05.627> and<00:27:06.290> I<00:27:06.420> always<00:27:06.678> find<00:27:06.905> it<00:27:06.970> really<00:27:07.149> interesting<00:27:07.898> reading<00:27:08.224> nonfiction<00:27:08.875> books.

00:27:10.492 --> 00:27:17.230
Just<00:27:11.551> seeing<00:27:13.765> how<00:27:13.797> many<00:27:13.957> mistakes<00:27:14.423> are<00:27:14.503> repeated,<00:27:15.289> how<00:27:15.385> many<00:27:16.107> patterns<00:27:16.669> exist.

00:27:18.112 --> 00:27:24.097
I'm<00:27:18.834> very<00:27:19.011> much<00:27:19.235> a<00:27:19.300> proponent<00:27:19.717> of<00:27:20.054> the<00:27:20.262> old<00:27:20.439> saying<00:27:20.856> of,<00:27:21.064> you<00:27:22.107> don't<00:27:22.284> understand<00:27:22.765> history,<00:27:23.150> you're<00:27:23.310> doomed<00:27:23.487> to<00:27:23.647> repeat<00:27:24.032> it.

00:27:24.353 --> 00:27:27.754
But<00:27:25.958> I<00:27:25.974> find<00:27:26.214> it<00:27:26.278> very<00:27:26.519> motivating<00:27:27.161> to<00:27:27.257> go<00:27:27.482> back.

00:27:27.802 --> 00:27:33.562
And<00:27:28.380> before<00:27:28.701> that,<00:27:29.423> the<00:27:29.888> previous<00:27:30.225> nonfiction<00:27:30.706> I<00:27:30.787> read<00:27:31.027> was<00:27:32.054> Ben<00:27:32.230> Franklin's<00:27:33.033> biography.

00:27:33.738 --> 00:27:48.312
And<00:27:33.915> before<00:27:34.236> that,<00:27:34.460> it<00:27:34.541> was<00:27:34.942> Chicken<00:27:35.263> Hawk<00:27:35.519> about<00:27:36.370> pilot<00:27:37.123> in<00:27:37.187> the<00:27:37.267> Vietnam<00:27:37.748> War<00:27:38.069> and<00:27:38.566> just<00:27:38.710> how<00:27:38.950> a<00:27:38.966> lot<00:27:39.287> of<00:27:40.634> these<00:27:40.794> management<00:27:41.355> issues,<00:27:41.932> these<00:27:42.253> organizational<00:27:43.038> issues,<00:27:43.519> like<00:27:43.695> they've<00:27:43.920> existed<00:27:44.641> over<00:27:44.801> and<00:27:44.882> over<00:27:45.058> again<00:27:45.314> and<00:27:45.459> some<00:27:45.699> of<00:27:45.763> them<00:27:45.923> can<00:27:46.020> be<00:27:46.565> done<00:27:46.741> really<00:27:46.965> well,<00:27:47.318> some<00:27:47.543> really<00:27:47.767> poorly.

00:27:49.691 --> 00:27:52.881
But<00:27:50.588> a<00:27:50.668> lot<00:27:50.845> of<00:27:50.925> it<00:27:51.005> is<00:27:51.149> stuff<00:27:51.566> that<00:27:51.694> we<00:27:51.791> can<00:27:51.951> continue<00:27:52.287> to<00:27:52.368> learn<00:27:52.656> from.

00:27:52.929 --> 00:27:58.812
So<00:27:53.377> a<00:27:53.458> little<00:27:53.618> bit<00:27:53.730> of<00:27:53.874> like<00:27:54.115> mistakes<00:27:54.532> of<00:27:54.660> the<00:27:54.756> past<00:27:55.141> to<00:27:55.237> avoid<00:27:55.558> them<00:27:55.798> and<00:27:55.878> a<00:27:56.022> little<00:27:56.199> bit<00:27:56.295> of<00:27:56.503> optimism<00:27:56.984> for<00:27:57.064> the<00:27:57.161> future<00:27:57.642> to<00:27:58.122> get<00:27:58.267> up<00:27:58.363> in<00:27:58.427> the<00:27:58.507> morning.

00:27:59.245 --> 00:28:00.319
And<00:27:59.309> that's<00:27:59.485> how<00:27:59.629> I<00:27:59.661> split<00:27:59.950> my<00:28:00.046> reading.

00:28:01.649 --> 00:28:08.189
What<00:28:01.809> are<00:28:01.954> your<00:28:02.771> like<00:28:03.092> tools,<00:28:03.957> apps,<00:28:04.534> things<00:28:04.839> like<00:28:05.015> that<00:28:05.320> that<00:28:05.496> you<00:28:05.961> could<00:28:06.218> not<00:28:06.458> live<00:28:06.683> without<00:28:07.019> that<00:28:07.244> you<00:28:07.404> use<00:28:07.660> on<00:28:07.757> a<00:28:07.885> daily<00:28:08.029> basis?

00:28:09.589 --> 00:28:12.389
I<00:28:10.091> need<00:28:10.415> to<00:28:10.900> do<00:28:11.046> more<00:28:11.240> of<00:28:11.370> this<00:28:11.548> to<00:28:11.629> be<00:28:11.774> completely<00:28:12.179> honest.

00:28:12.438 --> 00:28:18.135
So<00:28:12.584> I'm<00:28:13.571> not<00:28:15.189> a<00:28:15.206> huge<00:28:16.322> productivity<00:28:17.035> app<00:28:17.213> kind<00:28:17.552> of<00:28:18.054> guy.

00:28:18.830 --> 00:28:32.643
It's<00:28:18.927> interesting,<00:28:19.719> Martine,<00:28:20.155> our<00:28:20.285> board<00:28:20.527> member,<00:28:20.689> I<00:28:20.769> think<00:28:20.996> recently<00:28:21.578> did<00:28:22.046> a<00:28:22.143> podcast<00:28:22.709> or<00:28:22.951> wrote<00:28:23.016> an<00:28:23.194> article<00:28:23.598> about<00:28:23.921> this<00:28:24.180> of<00:28:25.861> you<00:28:25.941> have<00:28:26.022> to<00:28:26.119> be<00:28:26.249> careful<00:28:26.523> not<00:28:26.733> to<00:28:26.766> get<00:28:27.235> caught<00:28:27.703> up<00:28:27.816> in<00:28:27.930> the<00:28:27.962> cult<00:28:28.464> of<00:28:28.594> optimization<00:28:29.274> or<00:28:29.339> productivity<00:28:30.148> because<00:28:30.456> then<00:28:30.553> it<00:28:30.634> can<00:28:30.780> be<00:28:30.877> productivity<00:28:31.444> for<00:28:31.590> productivity's<00:28:32.319> sake.

00:28:34.154 --> 00:28:36.906
But<00:28:34.299> I<00:28:34.315> would<00:28:34.460> say,<00:28:34.782> I<00:28:34.814> mean,<00:28:35.200> I<00:28:35.603> use<00:28:36.021> email<00:28:36.407> all<00:28:36.472> the<00:28:36.568> time.

00:28:36.971 --> 00:28:37.776
I<00:28:37.035> use<00:28:37.228> Slack.

00:28:37.856 --> 00:28:41.156
I<00:28:37.937> have<00:28:38.420> an<00:28:38.516> executive<00:28:38.919> assistant<00:28:39.466> that<00:28:39.563> I<00:28:39.708> love<00:28:39.949> and<00:28:40.030> makes<00:28:40.271> my<00:28:40.416> day<00:28:40.738> way<00:28:40.915> better.

00:28:41.237 --> 00:28:44.090
So<00:28:41.382> I'm<00:28:41.720> a<00:28:41.800> little<00:28:42.042> bit<00:28:42.267> old<00:28:42.444> school<00:28:42.766> in<00:28:42.911> some<00:28:43.024> of<00:28:43.104> the<00:28:43.169> ways<00:28:43.410> that<00:28:43.555> I<00:28:43.620> do<00:28:43.759> things.

00:28:46.060 --> 00:28:54.723
But<00:28:46.999> I<00:28:47.015> mean,<00:28:47.873> I,<00:28:48.618> something,<00:28:49.169> things<00:28:49.428> that<00:28:49.573> I<00:28:49.719> didn't<00:28:50.043> use<00:28:50.448> every<00:28:50.707> day<00:28:51.355> 5<00:28:51.614> years<00:28:51.905> ago<00:28:52.164> are<00:28:52.407> like<00:28:52.715> obviously<00:28:53.622> video<00:28:53.945> conferencing.

00:28:55.937 --> 00:28:58.712
I<00:28:56.083> use<00:28:56.748> Miro<00:28:57.381> quite<00:28:57.576> a<00:28:57.641> bit<00:28:58.112> for<00:28:58.209> whiteboarding.

00:28:58.777 --> 00:29:01.098
I<00:28:58.842> love<00:28:59.086> to<00:28:59.183> think<00:28:59.362> about<00:28:59.605> everything<00:28:59.913> as<00:29:00.011> a<00:29:00.076> system<00:29:00.498> or<00:29:00.563> a<00:29:00.725> diagram.

00:29:04.011 --> 00:29:06.157
Um,<00:29:04.031> but<00:29:04.176> yeah,<00:29:04.531> I<00:29:04.611> think<00:29:05.078> you<00:29:05.110> need<00:29:05.223> to<00:29:05.336> find<00:29:05.577> that<00:29:05.738> balance.

00:29:06.286 --> 00:29:07.301
Everything's<00:29:06.640> about<00:29:06.866> trade-offs.

00:29:07.429 --> 00:29:14.435
Everything's<00:29:07.816> about<00:29:07.993> finding<00:29:08.251> this<00:29:08.460> kind<00:29:08.573> of<00:29:08.798> Pareto<00:29:09.136> frontier<00:29:09.700> of<00:29:09.909> trade-offs<00:29:10.489> of<00:29:10.650> like,<00:29:11.520> you<00:29:11.616> need<00:29:11.761> to<00:29:11.794> be<00:29:11.939> productive<00:29:12.486> but<00:29:12.567> not<00:29:13.308> in<00:29:13.388> the<00:29:13.469> cult<00:29:13.726> of<00:29:13.871> productivity.

00:29:14.822 --> 00:29:19.492
You<00:29:14.854> need<00:29:14.999> to<00:29:15.095> like<00:29:15.724> push<00:29:15.949> forward<00:29:16.368> but<00:29:16.513> not<00:29:16.690> blindly<00:29:17.882> and<00:29:18.139> be<00:29:18.284> constantly<00:29:18.703> making<00:29:18.945> that<00:29:19.090> trade-off.

00:29:21.030 --> 00:29:34.931
This<00:29:21.126> was<00:29:21.239> a<00:29:21.448> weird<00:29:21.932> year<00:29:22.254> economically,<00:29:22.915> obviously,<00:29:23.478> and<00:29:23.801> it,<00:29:24.123> I<00:29:24.139> mean,<00:29:24.510> it's<00:29:24.832> really<00:29:25.331> impressive<00:29:25.976> that<00:29:26.137> you've<00:29:26.846> raised<00:29:27.264> in<00:29:27.345> this,<00:29:27.667> in<00:29:27.812> that<00:29:28.006> environment,<00:29:28.634> kind<00:29:28.811> of<00:29:28.956> the<00:29:29.101> uncertainty<00:29:30.727> And<00:29:31.690> obviously<00:29:32.091> the<00:29:32.171> election<00:29:32.572> was<00:29:32.909> uncertainty,<00:29:33.391> but<00:29:33.551> it<00:29:33.615> was<00:29:33.952> just<00:29:34.177> a<00:29:34.257> weird<00:29:34.450> economy.

00:29:34.995 --> 00:29:39.697
Like<00:29:35.236> some<00:29:36.359> industries<00:29:36.825> doing<00:29:37.065> really<00:29:37.242> well,<00:29:37.499> some<00:29:38.365> basically<00:29:39.007> in<00:29:39.071> a<00:29:39.168> recession.

00:29:39.809 --> 00:29:42.088
Like<00:29:40.034> it's<00:29:40.516> just<00:29:40.772> been<00:29:40.869> a<00:29:41.013> fascinating<00:29:41.960> year.

00:29:42.939 --> 00:29:47.929
And<00:29:43.019> now<00:29:43.260> we<00:29:43.484> are<00:29:43.581> going<00:29:43.901> into<00:29:44.238> 2025<00:29:45.586> with<00:29:45.763> a<00:29:45.907> little<00:29:46.068> bit<00:29:46.228> more<00:29:46.389> certainty,<00:29:47.576> right?

00:29:48.379 --> 00:29:54.059
How<00:29:48.491> are<00:29:48.699> you<00:29:48.940> looking<00:29:49.341> at<00:29:49.839> 2025<00:29:50.882> and<00:29:51.267> planning<00:29:51.604> for<00:29:51.764> that<00:29:52.166> and<00:29:52.326> think<00:29:52.503> about<00:29:52.727> it<00:29:52.791> from<00:29:52.968> like<00:29:53.225> a<00:29:53.289> macro<00:29:53.722> level?

00:29:55.519 --> 00:30:00.446
Yeah,<00:29:55.921> I<00:29:56.017> think<00:29:56.659> there's<00:29:56.819> multiple<00:29:57.477> macro<00:29:57.862> levels<00:29:58.328> that<00:29:58.488> I<00:29:58.504> think<00:29:58.809> are<00:29:59.146> going<00:29:59.242> to<00:29:59.290> end<00:29:59.451> up<00:29:59.547> being<00:29:59.788> tailwinds<00:30:00.333> here.

00:30:00.574 --> 00:30:04.942
One<00:30:00.751> is,<00:30:01.072> I<00:30:01.088> mean,<00:30:01.570> some<00:30:01.810> of<00:30:01.875> that<00:30:02.678> uncertainty<00:30:04.027> has<00:30:04.509> collapsed.

00:30:05.167 --> 00:30:15.013
I<00:30:05.183> mean,<00:30:05.328> there's<00:30:05.553> still<00:30:05.810> quite<00:30:06.035> a<00:30:06.067> bit<00:30:06.211> of<00:30:06.292> uncertainty,<00:30:06.998> but<00:30:07.111> I<00:30:07.175> think<00:30:07.898> I'm<00:30:08.058> hopeful<00:30:08.460> that<00:30:08.621> people<00:30:09.327> will<00:30:09.424> have<00:30:09.584> enough<00:30:09.761> certainty<00:30:10.162> now<00:30:10.403> to<00:30:10.532> start<00:30:10.869> making<00:30:11.190> longer-term<00:30:11.672> investments<00:30:12.379> because<00:30:12.700> they<00:30:12.732> know<00:30:12.957> at<00:30:13.118> least<00:30:13.278> directionally<00:30:13.985> where<00:30:14.081> things<00:30:14.403> might<00:30:14.563> be<00:30:14.708> going.

00:30:16.330 --> 00:30:23.831
I<00:30:16.410> think<00:30:16.956> when<00:30:17.053> it<00:30:17.133> comes<00:30:17.390> to<00:30:17.599> AI<00:30:18.017> specifically,<00:30:19.960> again,<00:30:20.827> a<00:30:20.972> year<00:30:21.052> or<00:30:21.229> two<00:30:21.470> ago,<00:30:21.871> people<00:30:22.176> were<00:30:22.433> all<00:30:22.594> about<00:30:22.915> like,<00:30:23.156> this<00:30:23.317> is<00:30:23.477> cool.

00:30:24.104 --> 00:30:24.987
I<00:30:24.120> want<00:30:24.297> to<00:30:24.361> try<00:30:24.843> it.

00:30:25.084 --> 00:30:25.903
I<00:30:25.148> want<00:30:25.244> to<00:30:25.389> learn<00:30:25.630> more.

00:30:26.192 --> 00:30:27.027
I<00:30:26.352> wonder<00:30:26.690> what<00:30:26.850> it<00:30:26.883> can<00:30:26.995> do.

00:30:27.140 --> 00:30:27.268
Do.

00:30:28.454 --> 00:30:32.349
And<00:30:28.534> I<00:30:28.614> think<00:30:29.095> this<00:30:29.319> next<00:30:29.640> year<00:30:29.832> is<00:30:29.977> about<00:30:30.762> how<00:30:30.858> do<00:30:31.003> I<00:30:31.355> extract<00:30:32.061> value?

00:30:32.622 --> 00:30:34.369
How<00:30:32.718> do<00:30:32.862> I<00:30:33.167> actually<00:30:33.728> put<00:30:33.840> it<00:30:33.984> to<00:30:34.208> work?

00:30:35.170 --> 00:30:39.947
And<00:30:35.667> shifting<00:30:36.308> from<00:30:36.869> prototypes<00:30:37.671> to<00:30:37.751> production<00:30:38.873> is<00:30:38.953> going<00:30:39.097> to<00:30:39.194> be<00:30:39.274> this<00:30:39.498> big<00:30:39.755> push.

00:30:41.037 --> 00:30:56.555
And<00:30:41.197> very<00:30:41.502> intentionally<00:30:42.159> when<00:30:42.303> we<00:30:42.383> started<00:30:42.736> the<00:30:42.864> company,<00:30:43.185> and<00:30:43.281> again,<00:30:43.586> some<00:30:43.778> of<00:30:43.842> this<00:30:44.002> was<00:30:44.163> because<00:30:44.884> we<00:30:45.669> got<00:30:45.846> to<00:30:45.926> see<00:30:46.246> this<00:30:46.952> with<00:30:47.369> extremely<00:30:48.234> elite<00:30:48.811> firms<00:30:49.132> over<00:30:49.276> the<00:30:49.372> last<00:30:49.693> decade,<00:30:50.158> was<00:30:51.440> we<00:30:51.536> were<00:30:51.680> trying<00:30:51.857> to<00:30:52.001> anticipate,<00:30:52.562> well,<00:30:52.818> what<00:30:52.979> problem<00:30:53.476> do<00:30:53.524> you<00:30:53.700> have<00:30:54.806> as<00:30:54.903> you<00:30:54.983> shift<00:30:55.304> from<00:30:55.464> prototyping<00:30:56.042> to<00:30:56.170> production.

00:30:57.454 --> 00:31:00.807
And<00:30:58.192> when<00:30:58.288> you're<00:30:58.432> prototyping,<00:30:59.058> you<00:30:59.090> need<00:30:59.299> a<00:30:59.315> lot<00:30:59.491> of<00:30:59.555> tools<00:30:59.940> for<00:31:00.117> like<00:31:00.277> getting<00:31:00.502> started.

00:31:00.855 --> 00:31:04.176
You<00:31:00.919> need<00:31:01.304> playgrounds,<00:31:01.962> you<00:31:02.042> need<00:31:02.283> like<00:31:02.524> debuggers,<00:31:03.149> you<00:31:03.181> need<00:31:03.486> intuition.

00:31:05.091 --> 00:31:07.032
And<00:31:05.203> there's<00:31:05.427> a<00:31:05.492> lot<00:31:05.732> of<00:31:05.796> companies<00:31:06.278> that<00:31:06.374> have<00:31:06.518> raised<00:31:06.839> money.

00:31:07.096 --> 00:31:08.973
There's<00:31:07.256> a<00:31:07.273> lot<00:31:07.417> of<00:31:07.465> companies<00:31:07.802> that<00:31:07.978> do<00:31:08.139> that<00:31:08.299> extremely<00:31:08.845> well.

00:31:09.406 --> 00:31:12.262
But<00:31:09.503> our<00:31:09.663> goal<00:31:09.968> always<00:31:10.289> was<00:31:10.465> like,<00:31:10.786> okay,<00:31:11.107> like,<00:31:11.348> what's<00:31:11.588> the<00:31:11.749> next<00:31:12.054> step?

00:31:12.391 --> 00:31:15.728
What's<00:31:12.551> that<00:31:12.792> next<00:31:13.112> bottleneck<00:31:13.594> that<00:31:13.738> everybody's<00:31:14.139> going<00:31:14.300> to<00:31:14.396> run<00:31:14.556> into<00:31:14.813> simultaneously?

00:31:16.482 --> 00:31:41.164
And<00:31:16.562> I<00:31:16.642> think<00:31:16.867> that<00:31:16.963> that's<00:31:17.348> where<00:31:18.230> this<00:31:18.391> production,<00:31:19.594> both<00:31:20.156> Making<00:31:20.493> that<00:31:20.734> jump<00:31:21.038> to<00:31:21.135> production,<00:31:21.680> like<00:31:21.841> crossing<00:31:22.338> this,<00:31:22.884> what<00:31:22.980> we're<00:31:23.141> coining<00:31:23.478> as<00:31:23.622> the<00:31:23.847> AI<00:31:24.104> confidence<00:31:24.650> gap,<00:31:25.291> to<00:31:25.388> even<00:31:25.693> have<00:31:25.885> enough<00:31:26.110> confidence<00:31:26.672> to<00:31:27.217> push<00:31:27.458> something<00:31:27.795> into<00:31:27.940> production<00:31:29.304> and<00:31:29.946> get<00:31:30.122> it<00:31:30.203> through<00:31:30.523> either<00:31:30.748> your<00:31:30.844> internal<00:31:31.246> compliance<00:31:31.727> processes<00:31:32.289> or<00:31:32.369> even<00:31:32.594> just<00:31:32.754> having<00:31:33.156> enough<00:31:33.477> internal<00:31:33.894> confidence<00:31:34.456> to<00:31:35.001> stick<00:31:35.242> your<00:31:35.338> neck<00:31:35.659> out<00:31:35.900> in<00:31:35.980> an<00:31:36.125> organization,<00:31:36.863> which<00:31:37.007> can<00:31:37.104> be<00:31:37.248> hard<00:31:37.441> to<00:31:37.585> say,<00:31:37.826> like,<00:31:38.789> I'm<00:31:38.885> going<00:31:39.655> to<00:31:39.688> sign<00:31:39.992> up<00:31:40.137> for<00:31:40.217> the<00:31:40.297> pager<00:31:40.618> duty<00:31:40.811> for<00:31:41.020> this.

00:31:41.357 --> 00:31:44.855
I'm<00:31:41.469> going<00:31:41.581> to<00:31:41.613> sign<00:31:41.983> up<00:31:42.159> for<00:31:42.945> what<00:31:43.122> goes<00:31:43.347> wrong<00:31:43.700> if<00:31:43.989> something<00:31:44.310> goes<00:31:44.566> wrong.

00:31:46.556 --> 00:31:55.136
But<00:31:46.636> then<00:31:46.797> once<00:31:46.957> something's<00:31:47.374> in<00:31:47.454> production,<00:31:48.561> fundamentally,<00:31:49.619> like<00:31:50.261> one<00:31:50.421> of<00:31:50.453> the<00:31:50.565> biggest<00:31:50.806> problems<00:31:51.207> with<00:31:51.255> software<00:31:52.009> isn't<00:31:52.410> writing<00:31:52.762> software,<00:31:53.644> it's<00:31:53.853> maintaining<00:31:54.607> software.

00:31:55.729 --> 00:31:57.221
It's<00:31:55.938> keeping<00:31:56.451> systems<00:31:57.076> up.

00:31:57.958 --> 00:32:03.507
Like<00:31:58.215> most<00:31:58.824> software<00:31:59.338> has<00:31:59.626> already<00:31:59.963> been<00:32:00.203> written<00:32:00.428> in<00:32:00.765> a<00:32:00.845> lot<00:32:01.021> of<00:32:01.086> these<00:32:01.326> organizations<00:32:02.288> and<00:32:02.433> for<00:32:02.513> many<00:32:02.834> applications.

00:32:03.892 --> 00:32:10.114
That<00:32:04.036> doesn't<00:32:04.213> mean<00:32:04.373> like<00:32:05.079> making<00:32:05.335> improvements<00:32:05.816> and<00:32:05.880> doing<00:32:06.041> new<00:32:06.137> things<00:32:06.378> isn't<00:32:06.602> good,<00:32:06.843> but<00:32:06.939> it's<00:32:07.083> like,<00:32:07.885> that<00:32:08.078> is<00:32:08.222> a<00:32:08.286> huge<00:32:08.879> organizational<00:32:09.585> problem.

00:32:11.173 --> 00:32:31.543
And<00:32:11.253> that's<00:32:11.574> where<00:32:11.814> you<00:32:11.910> need<00:32:12.536> better<00:32:12.777> confidence,<00:32:13.355> better<00:32:13.660> testing,<00:32:14.318> better<00:32:15.666> understanding<00:32:16.645> of<00:32:16.742> these<00:32:16.902> systems<00:32:18.235> as<00:32:18.636> not<00:32:18.877> only<00:32:19.134> you're<00:32:19.278> building<00:32:19.615> them<00:32:19.840> from<00:32:20.257> scratch,<00:32:21.365> but<00:32:21.606> making<00:32:22.023> iterative<00:32:22.360> improvements,<00:32:23.067> refactoring,<00:32:24.190> changing<00:32:24.736> out<00:32:24.913> components,<00:32:26.036> making<00:32:26.341> sure<00:32:26.518> it<00:32:26.598> doesn't<00:32:26.839> break<00:32:27.160> underneath<00:32:27.642> you,<00:32:28.669> having<00:32:29.006> it<00:32:29.231> weather<00:32:29.552> the<00:32:29.632> storm<00:32:30.435> of<00:32:30.772> real<00:32:31.157> usage.

00:32:32.522 --> 00:32:34.432
And<00:32:33.244> that's<00:32:33.806> where<00:32:33.919> we've<00:32:34.063> been<00:32:34.288> aiming.

00:32:34.480 --> 00:32:45.796
So<00:32:34.689> one<00:32:34.801> of<00:32:34.882> the<00:32:34.930> things<00:32:35.106> that<00:32:35.203> I'm<00:32:35.363> super<00:32:35.684> excited<00:32:36.166> about<00:32:36.407> is<00:32:36.551> this<00:32:36.969> macro<00:32:37.434> maturity<00:32:38.301> of<00:32:38.622> applications,<00:32:40.420> productionalizing<00:32:41.114> is<00:32:41.324> running<00:32:41.808> directly<00:32:42.389> into<00:32:42.632> the<00:32:42.712> type<00:32:42.971> of<00:32:43.035> problem<00:32:43.423> that<00:32:43.520> we're<00:32:43.600> trying<00:32:43.843> to<00:32:43.923> solve<00:32:44.731> with<00:32:44.828> respect<00:32:45.199> to<00:32:45.296> confidence.

00:32:47.550 --> 00:32:52.838
Because<00:32:47.729> in<00:32:47.810> all<00:32:47.989> honesty,<00:32:49.128> confidence<00:32:49.535> matters<00:32:49.860> a<00:32:49.877> little<00:32:50.088> bit<00:32:50.251> less<00:32:50.576> when<00:32:50.674> you're<00:32:50.755> just<00:32:50.918> playing<00:32:51.227> around,<00:32:52.057> because<00:32:52.204> you're<00:32:52.301> just<00:32:52.448> playing<00:32:52.692> around.

00:32:53.850 --> 00:32:57.594
There's<00:32:54.044> no<00:32:54.125> real<00:32:54.287> risk,<00:32:54.449> there's<00:32:54.595> no<00:32:54.693> real<00:32:54.936> downside,<00:32:56.541> there's<00:32:56.719> no<00:32:56.865> real<00:32:57.043> permanence.

00:32:58.162 --> 00:33:06.355
But<00:32:58.259> once<00:32:58.437> it<00:32:58.502> becomes<00:32:58.891> real,<00:32:59.870> there's<00:33:00.112> risk,<00:33:00.387> there's<00:33:00.630> downside,<00:33:01.179> there's<00:33:01.374> maintenance,<00:33:02.635> there's<00:33:03.298> all<00:33:03.444> of<00:33:03.557> that,<00:33:03.929> the<00:33:04.414> things<00:33:04.754> that<00:33:05.077> require<00:33:05.724> better<00:33:05.950> confidence.

00:33:07.509 --> 00:33:07.621
Yeah.

00:33:07.686 --> 00:33:12.535
Once<00:33:07.910> it<00:33:08.007> matters<00:33:08.569> and<00:33:08.713> once<00:33:09.147> somebody<00:33:09.612> is<00:33:09.853> relying<00:33:10.399> on<00:33:10.495> it,<00:33:11.459> it<00:33:11.555> starts<00:33:11.925> to<00:33:12.005> get<00:33:12.101> pretty<00:33:12.406> real.

00:33:13.289 --> 00:33:13.482
Yeah.

00:33:13.530 --> 00:33:13.803
Right.

00:33:14.413 --> 00:33:21.623
Once<00:33:14.606> you<00:33:14.686> have<00:33:14.815> to<00:33:14.911> hand<00:33:15.168> it<00:33:15.296> off<00:33:15.489> to<00:33:15.569> someone<00:33:15.939> else<00:33:16.115> too,<00:33:17.239> like<00:33:17.721> most<00:33:18.267> software<00:33:18.781> is<00:33:18.941> not<00:33:19.182> maintained<00:33:19.808> by<00:33:20.049> the<00:33:20.146> original<00:33:20.836> person<00:33:21.157> who<00:33:21.318> wrote<00:33:21.558> it.

00:33:22.538 --> 00:33:31.738
And<00:33:22.763> one<00:33:23.020> way<00:33:23.260> that<00:33:23.405> we've<00:33:23.501> been<00:33:23.662> able<00:33:23.855> to<00:33:23.983> do<00:33:24.128> that<00:33:24.304> with<00:33:24.384> traditional<00:33:24.802> software<00:33:25.348> is<00:33:25.573> through<00:33:25.830> having<00:33:26.697> robust<00:33:27.034> testing<00:33:27.997> because<00:33:28.238> it's<00:33:28.383> like<00:33:28.720> the<00:33:28.800> tests<00:33:29.201> will<00:33:29.346> let<00:33:29.507> me<00:33:29.539> know<00:33:29.699> if<00:33:29.779> something's<00:33:30.229> broken<00:33:30.630> because<00:33:31.032> I<00:33:31.128> didn't<00:33:31.369> write<00:33:31.610> this.

00:33:32.413 --> 00:33:36.119
And<00:33:32.493> then<00:33:32.798> when<00:33:32.894> something<00:33:33.279> does<00:33:33.536> break,<00:33:34.162> the<00:33:34.258> test<00:33:34.563> will<00:33:34.659> let<00:33:34.819> me<00:33:34.900> know<00:33:35.140> whether<00:33:35.349> or<00:33:35.381> not<00:33:35.686> I<00:33:35.718> fixed<00:33:36.071> it.

00:33:37.226 --> 00:33:39.633
And<00:33:37.291> we<00:33:37.387> just<00:33:38.013> don't<00:33:38.253> have<00:33:38.526> that<00:33:38.751> for<00:33:38.895> AI<00:33:39.232> today.

00:33:39.713 --> 00:33:47.031
And<00:33:39.794> that's<00:33:40.018> not<00:33:40.195> necessary<00:33:40.740> if<00:33:40.837> you're<00:33:40.981> the<00:33:41.222> only<00:33:41.543> one<00:33:41.783> playing<00:33:42.120> with<00:33:42.217> something<00:33:42.859> in<00:33:42.987> an<00:33:43.083> isolation,<00:33:44.592> but<00:33:44.672> that's<00:33:44.848> not<00:33:45.394> how<00:33:45.971> enterprise<00:33:46.292> software<00:33:46.758> works.

00:33:47.255 --> 00:33:48.394
It<00:33:47.496> grows<00:33:47.801> up<00:33:47.977> eventually.

00:33:48.459 --> 00:33:54.877
And<00:33:48.555> I<00:33:48.619> think<00:33:48.924> now<00:33:49.117> 2025<00:33:50.801> is<00:33:50.866> going<00:33:51.010> to<00:33:51.106> be<00:33:51.186> the<00:33:51.411> year<00:33:51.812> of<00:33:52.631> growing<00:33:53.096> up<00:33:53.593> in<00:33:53.674> the<00:33:53.754> enterprise<00:33:54.380> for<00:33:54.476> GenAI.

00:33:56.305 --> 00:33:56.498
Yeah.

00:33:57.364 --> 00:34:04.723
Finally,<00:33:58.327> we<00:33:58.471> end<00:33:58.632> every<00:33:58.792> interview<00:33:59.033> with<00:33:59.129> the<00:33:59.226> same<00:33:59.450> question,<00:34:00.076> And<00:34:00.172> that<00:34:00.348> is,<00:34:00.492> at<00:34:00.573> CEO.com,<00:34:01.374> we<00:34:01.534> believe<00:34:01.758> the<00:34:01.855> chances<00:34:02.319> one<00:34:02.560> gives<00:34:02.976> is<00:34:03.040> just<00:34:03.185> as<00:34:03.297> important<00:34:03.617> as<00:34:03.761> the<00:34:03.858> chances<00:34:04.258> one<00:34:04.418> takes.

00:34:05.444 --> 00:34:08.344
When<00:34:05.604> you<00:34:05.781> hear<00:34:06.005> that,<00:34:06.181> who<00:34:06.405> gave<00:34:06.646> you<00:34:06.758> a<00:34:06.822> chance<00:34:07.207> to<00:34:07.303> get<00:34:07.463> you<00:34:07.623> to<00:34:07.687> where<00:34:07.928> you<00:34:08.024> are<00:34:08.184> today?

00:34:09.130 --> 00:34:10.764
Oh<00:34:09.242> my<00:34:09.354> God,<00:34:09.947> so<00:34:10.203> many<00:34:10.428> people.

00:34:11.373 --> 00:34:21.757
Um,<00:34:12.671> I<00:34:12.751> would<00:34:12.911> say<00:34:13.296> one<00:34:13.472> of<00:34:13.504> the<00:34:13.568> things<00:34:13.937> that<00:34:14.033> I<00:34:14.177> love<00:34:14.434> the<00:34:14.514> most<00:34:14.850> about<00:34:15.011> Silicon<00:34:15.475> Valley<00:34:16.100> is<00:34:17.302> how<00:34:18.520> people<00:34:19.065> are—<00:34:20.123> they,<00:34:20.363> they<00:34:20.587> realize<00:34:21.084> that<00:34:21.164> they<00:34:21.324> got<00:34:21.645> there.

00:34:22.110 --> 00:34:25.186
Many<00:34:22.350> people<00:34:22.606> realize<00:34:22.991> that<00:34:23.087> they<00:34:23.231> got<00:34:23.424> there<00:34:24.113> by<00:34:24.273> other<00:34:24.449> people<00:34:24.754> helping<00:34:25.074> them.

00:34:25.250 --> 00:34:29.842
And<00:34:25.347> so<00:34:25.972> when<00:34:26.068> we<00:34:26.116> were<00:34:26.357> going<00:34:26.614> through<00:34:26.951> YC,<00:34:27.963> we<00:34:28.059> learned<00:34:28.300> a<00:34:28.316> ton<00:34:28.862> from<00:34:29.007> the<00:34:29.103> partners<00:34:29.585> there.

00:34:30.308 --> 00:34:37.020
Even<00:34:30.468> before<00:34:30.854> we<00:34:30.950> got<00:34:31.110> into<00:34:31.351> YC,<00:34:32.074> like<00:34:32.556> previous<00:34:32.957> alumni<00:34:33.744> helped<00:34:34.386> like<00:34:34.547> practice<00:34:34.949> interview<00:34:35.270> with<00:34:35.511> us<00:34:35.687> and<00:34:35.848> review<00:34:36.233> our<00:34:36.314> application.

00:34:37.277 --> 00:34:39.638
And<00:34:38.241> we<00:34:38.321> just<00:34:38.481> got<00:34:38.642> a<00:34:38.722> lot<00:34:38.899> of<00:34:38.963> really<00:34:39.140> good<00:34:39.236> advice.

00:34:39.702 --> 00:34:46.013
And<00:34:39.782> sometimes<00:34:40.184> it's<00:34:40.328> hard<00:34:40.826> to<00:34:40.906> get<00:34:41.308> feedback<00:34:41.774> where<00:34:41.870> they're<00:34:42.014> like,<00:34:42.255> hey,<00:34:43.797> you<00:34:43.941> can't<00:34:44.182> just<00:34:44.439> like<00:34:44.680> submit<00:34:45.082> your<00:34:45.226> PhD<00:34:45.708> thesis.

00:34:46.109 --> 00:34:47.523
You<00:34:46.142> have<00:34:46.286> to<00:34:46.318> have<00:34:46.463> like<00:34:46.623> a<00:34:46.688> product.

00:34:48.277 --> 00:34:50.526
Like<00:34:48.454> you<00:34:48.518> need<00:34:48.631> to<00:34:48.695> have<00:34:48.855> a<00:34:48.920> go-to-market<00:34:49.417> plan<00:34:49.674> and<00:34:49.803> things<00:34:50.044> like<00:34:50.220> that.

00:34:50.606 --> 00:34:57.868
But<00:34:51.329> it<00:34:51.409> helped<00:34:51.602> us<00:34:51.746> grow<00:34:52.051> up<00:34:52.148> in<00:34:52.244> a<00:34:52.308> way<00:34:52.469> that<00:34:52.629> we<00:34:52.790> were<00:34:53.351> I<00:34:53.432> think<00:34:54.303> every<00:34:54.819> founder<00:34:55.997> is<00:34:56.964> naive<00:34:57.400> in<00:34:57.545> some<00:34:57.787> way.

00:34:58.093 --> 00:35:00.626
Like<00:34:58.255> you<00:34:58.351> need<00:34:58.513> to<00:34:58.593> be<00:34:58.755> naive<00:34:59.239> to<00:34:59.335> be<00:34:59.480> able<00:34:59.642> to<00:34:59.964> take<00:35:00.206> that<00:35:00.448> leap.

00:35:02.319 --> 00:35:10.274
But<00:35:03.140> helping<00:35:03.542> us<00:35:03.881> kind<00:35:04.026> of<00:35:04.187> hone,<00:35:05.652> know<00:35:05.813> what<00:35:05.974> we<00:35:06.135> don't<00:35:06.441> know<00:35:06.634> and<00:35:06.763> know<00:35:06.860> where<00:35:07.021> we<00:35:07.118> need<00:35:07.263> to<00:35:07.343> get<00:35:07.504> better<00:35:07.762> is<00:35:08.003> something<00:35:08.374> that<00:35:08.470> was<00:35:08.937> really<00:35:09.372> important<00:35:10.016> to<00:35:10.145> us.

00:35:10.403 --> 00:35:12.786
So<00:35:11.031> we<00:35:11.111> saw<00:35:11.289> this<00:35:11.434> with<00:35:11.691> our<00:35:11.997> YC<00:35:12.319> partners.

00:35:12.883 --> 00:35:18.888
We<00:35:12.963> saw<00:35:13.140> this<00:35:13.382> with<00:35:14.750> Martin<00:35:15.491> as<00:35:15.796> we<00:35:15.893> were<00:35:16.022> his<00:35:16.183> first<00:35:16.440> board<00:35:16.923> as<00:35:17.551> he<00:35:17.648> was<00:35:17.809> joining<00:35:18.050> Andreessen<00:35:18.453> Horowitz.

00:35:18.936 --> 00:35:22.239
Like<00:35:19.403> we<00:35:19.805> all<00:35:20.224> kind<00:35:20.385> of<00:35:20.546> learned<00:35:20.884> and<00:35:21.029> grow<00:35:21.500> grew<00:35:21.692> together.

00:35:24.102 --> 00:35:32.964
But<00:35:24.247> yeah,<00:35:24.504> I'd<00:35:24.665> say<00:35:25.050> I<00:35:25.147> am<00:35:27.010> extremely<00:35:27.797> lucky<00:35:28.054> to<00:35:28.199> be<00:35:28.360> here,<00:35:28.906> and<00:35:30.352> a<00:35:30.432> lot<00:35:30.609> of<00:35:30.689> that<00:35:30.850> comes<00:35:31.139> from<00:35:31.220> the<00:35:31.316> people<00:35:31.639> who<00:35:31.720> helped<00:35:32.382> us<00:35:32.544> along<00:35:32.657> the<00:35:32.786> way.

00:35:34.224 --> 00:35:38.132
Scott,<00:35:34.579> thank<00:35:34.789> you<00:35:34.821> so<00:35:34.983> much<00:35:35.290> for<00:35:35.451> coming<00:35:35.758> on,<00:35:36.162> and<00:35:36.323> congratulations<00:35:37.163> on<00:35:37.292> everything<00:35:37.631> you're<00:35:37.777> doing.

00:35:38.520 --> 00:35:43.783
What<00:35:38.665> you're<00:35:38.762> doing,<00:35:39.069> I<00:35:39.166> think,<00:35:39.343> is<00:35:39.893> fundamental<00:35:41.329> if<00:35:41.426> we're<00:35:41.588> going<00:35:42.088> to<00:35:42.233> continue<00:35:42.621> to<00:35:42.701> make<00:35:43.121> strides<00:35:43.605> in<00:35:43.702> AI.

00:35:44.106 --> 00:35:48.739
So<00:35:45.139> well<00:35:45.300> done<00:35:45.543> there,<00:35:46.091> and<00:35:46.673> I'm<00:35:46.899> sure<00:35:47.060> we'll<00:35:47.205> have<00:35:47.302> you<00:35:47.334> back<00:35:47.706> on,<00:35:47.803> but<00:35:47.948> thanks<00:35:48.190> so<00:35:48.222> much<00:35:48.368> for<00:35:48.513> coming<00:35:48.690> on.

00:35:49.175 --> 00:35:50.236
Thank<00:35:49.338> you<00:35:49.420> so<00:35:49.583> much<00:35:49.811> for<00:35:49.893> having<00:35:50.138> me.

00:35:50.317 --> 00:35:51.198
I<00:35:50.382> really<00:35:50.627> appreciated<00:35:51.116> it.

```