What this covers

This is an eight-minute clip from a longer interview Joe Lonsdale conducted with Scott Wu and Russell Kaplan of Cognition. The speakers are unlabeled in the transcript, but the material suggests one voice describing how Cognition hires and where AI is heading, while a second, an investor and former Palantir recruiter, adds background on talent and small business. The central line of inquiry is what happens to work as models run unattended for longer. The evidence is the METR autonomy metric, which the speaker says rose from about ten seconds to 18-20 hours in models like Opus 4.6. That figure is measured in equivalent human time, with actual execution roughly a hundred times faster.

The clip then moves through several linked areas. One is the form factor of human-AI interaction: as autonomy grows, real-time supervision gives way to delegating whole projects with periodic checkpoints, and later to proactive, event-driven work in which AI notices something like a security vulnerability and responds on its own. Another is organizational urgency. Large companies and governments, the speaker claims, can't stay five times slower than competitors, while a core group of agent forward engineers already reports 5x to 10x productivity gains.

The hiring discussion is the most concrete. Cognition's interviews reverse the old ban on AI: candidates must use it freely to build a full product in a few hours, and are judged on product judgment, trade-offs, and information gathering rather than tool skill. The company also recruits heavily from former founders, one team entirely so, pitching software engineering as a problem big enough to absorb high-agency people. The second speaker adds a personal aside about misjudging Scott's CEO potential early on. The clip closes on AI as a small business enabling technology, cheapening legal, financial, and engineering labor, with a side project aimed at creating many new small business owners.

Sharpest takeaway

AI capabilities are advancing exponentially (doubling every 2-3 months), fundamentally changing how work is organized and who can accomplish it, particularly enabling small business founders and requiring organizations to shift from managing constant human supervision to delegating autonomous multi-day projects.

  • METR autonomy metric has grown from 10 seconds to 18+ hours in just 2-3 years, doubling roughly every 2-3 months
  • As AI autonomy increases, the form factor of human-AI interaction shifts from real-time shepherding to high-level task delegation with periodic checkpoints
  • AI acts as a small business enabling technology by providing quality legal, financial, and technical capabilities that previously required expensive specialized staff

The claims · ranked15 claims · weighted by value

This asset isn't compiled yet

You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.

0.64

Top AI lab leaders predict continued high-pace capability improvements for at least two to three years, making it difficult to predict exactly what the field will look like five years ahead but clear that things will change substantially.

forecasthigh valueestablishednovelty 1/4durability 2/4· Unknown Speaker 1

“I talked to Yeah. some of the people running the top labs. They're pretty convinced things are going to keep getting better at a pretty high pace for at least two or three years. It's like kind of like with Moore's law, you can't really see out what's going to get there five years, but but see it feels like things are going to change a lot.”

0.57

As AI autonomy increases from performing seconds of work to performing days or weeks of work, the optimal form factor for human-AI interaction changes fundamentally: from real-time shepherding where humans oversee every step, to high-level task delegation where humans set objectives and AI autonomously completes entire projects with periodic verification checkpoints.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Unknown Speaker 1

“when we're saying, okay, you do 10 seconds of work obviously the answer is you as a human need to be staring at your file of code and like shepherding it and hand-holding it with every single step, right? If you're talking about it's doing days of work or weeks of work, now you're actually giving it whole, you know, output level tasks of like, hey, you know, we we really need to go make this app much faster. Can you go and run this whole thing and then, you know, do a smoke test and make sure all of the changes look right?”

0.57

What Cognition actually tests for in interviews is not tool proficiency per se, but rather judgment and decision-making: what the candidate believes is the right thing to build, how they make product trade-offs, and how they gather information to determine what should be built.

normativehigh valuespeaker onlynovelty 3/4durability 3/4· Unknown Speaker 1

“what we actually want to test is is yeah, what what what do you think is the right thing to build or how do you make these product decisions? How do you make these trade-offs? How do you decide or collect information about what you should be doing?”

0.55

Large organizations, companies, and governments are realizing they cannot remain five times slower than competitors and are now urgently prioritizing learning how to effectively use AI systems.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Unknown Speaker 1

“all of these big organizations or these these companies or governments or things like that are are seeing those same results and realizing wait, we can't just sit here and be five times slower and you know, we have to go learn how to do this right now.”

0.52

AI is an extremely effective small business enabling technology because it addresses the core constraint of small business formation: the inability to afford specialized labor across different functions (legal, financial, technical), and AI-powered tools can now provide high-quality analysis, legal review, and software development at negligible marginal cost.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Unknown Speaker 1

“I think there's going to be an explosion in small businesses. Um I think AI is actually like an extremely small business enabling technology in particular where you think about what's hard about about building or starting small business, it's, you know, you don't have the resources of a large company, that's specialization of labor in each part of the in each part of the process and I think AI it's extraordinarily enabling, right?”

0.52

The interview and hiring philosophy has fundamentally shifted from preventing AI use during interviews to explicitly allowing and expecting AI use, because evaluating people on tasks that AI can already perform is measuring the wrong thing.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Unknown Speaker 1

“one or two years ago I think a lot of people had this mentality in terms of how you interview of like, okay, there's all these AI tools, how do we interview people in a way that um you know, how how do we make sure that people aren't using AI while we're interviewing them cuz then we're and and I think that is totally flipped. And and honestly I think that was wrong, right?”

0.52

The value proposition of Cognition to potential founder-employees is that the problem space (solving software engineering) is sufficiently large and ambitious that a high-agency person can accomplish more by collaborating on this one problem than by starting their own independent company.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Unknown Speaker 1

“I think the the the play solving software engineering feels like a big enough one, you know, that that we can all do that together and and and I think that's a lot of what it comes down to honestly is just like um yeah, are are we working on something and doing something that's that's that's exciting for for folks who to your point are are, you know, I I I I think a ton of the folks at Cognition are could could very very easily go off and start their own companies and get funded and build their teams and so on and and and the question I think for us always has been about um how how do we make this the place that that that makes more sense for them to do?”

0.49

There exists a core group of 'agent forward engineers' and 'agent forward companies' that understand how to use AI effectively and are currently seeing 5x to 10x productivity gains as a result, while most other organizations have not yet adopted these practices.

factualhigh valuespeaker onlynovelty 2/4durability 2/4· Unknown Speaker 1

“you know, you have this core group of we'll call it like agent forward engineers, right? Or agent forward companies that understand how to use this and they are seeing these, you know, 5x, 10x productivity gains as a result.”

0.49

AI-enabled autonomous work will shift from reactive human-triggered tasks to proactive self-initiated work and event-driven automation, where AI detects situations (like security vulnerabilities) and autonomously responds without requiring human moderation at every step.

forecasthigh valuespeaker onlynovelty 2/4durability 2/4· Unknown Speaker 1

“You're going to see a lot more I think proactive work. Um You're going to see a lot more basically like event-driven kickoff work where it's it doesn't have to be human that's that's moderating it every second.”

0.49

Cognition's interview process requires candidates to use AI freely to build a complete product in a few hours—a task that would be impossible to complete by hand in that timeframe, requiring candidates to demonstrate AI tool proficiency, product decision-making, trade-off analysis, and information-gathering ability.

factualhigh valuespeaker onlynovelty 2/4durability 2/4· Unknown Speaker 1

“for us and this has always been the case for us, you know, our interview process has always been you can use as much AI as you want to use, you know, it's just like we're going to give you a few hours just build build your whole own product surface, right?”

0.48

The METR autonomy metric, where the 18-hour figure represents the equivalent human time to complete the work (not the actual AI execution time), has been doubling approximately every 2-3 months for the past few years, with actual AI execution time being 100 times faster than the human equivalent (e.g., AI completes an 18-hour human task in 1-2 hours).

factualhigh valuespeaker onlynovelty 3/4durability 2/4· Unknown Speaker 1

“This is always very weird to me cuz it's implying that models work in human time though, right? Which is so weird. Like why are they 100 times faster? Yeah, so so so the answer the the they do typically do the tasks in less time than a human would and then the 18 hours is basically this is how long it would take a human to do that amount of work”

0.48

The METR (Methodology for Evaluating Transformative Robotics) autonomy metric measures how much human work an AI model can perform automatically before requiring human correction, and this metric has grown from approximately 10 seconds two to three years ago to 18-20 hours in the latest models like Claude Opus 4.6.

factualhigh valuespeaker onlynovelty 3/4durability 2/4· Unknown Speaker 1

“one of the stats that people talk about a lot is this METR report which basically says for um each different model that comes out roughly how much human work can it do uh in an automated fashion before you have to go interrupt it and say oh that was wrong, let's go do this, right?”

0.46

Cognition (the AI company) has hired a significant percentage of its team from former founders, with one entire team (special projects engineers) consisting exclusively of former founders, which reflects a deliberate strategy to attract high-agency individuals.

factualhigh valuespeaker onlynovelty 2/4durability 2/4· Unknown Speaker 2

“one thing you guys have done is like a significant percent of the Cognition hires are actually former founders. So not only are you hiring the best people in the world, you're hiring a ton of former founders.”

0.29

Second speaker hired many of the first few hundred people at Palantir and spent significant time on talent acquisition, including hiring Scott at an earlier company with Bio's assistance, but at that time did not perceive Scott as having CEO-level capability, though Scott has since grown substantially into the role.

factualspeaker onlynovelty 0/4durability 4/4· Unknown Speaker 2

“So I I hired a lot of the first few hundred people at Palantir and I spent a lot of time on talent. Obviously even hired Scott at one point a long time ago with one of my companies with with Bio's help... He's very very very very impressive. I but I did not I did not see him at at the time as someone who is like a CEO person. So he really grew a lot which is good.”

0.17

The speaker personally has a side project focused on helping create tens of thousands of small business owners, indicating direct alignment with and action toward leveraging AI for small business enablement.

factualspeaker onlynovelty 0/4durability 2/4· Unknown Speaker 2

“I actually I love this. I actually have a a small thing on the side where I'm trying to help create tens of thousands of small business owners.”