"Money isn't the problem. The question is: will this product sell? Who is your customer? Have you talked to them?" 1:07:02
From the episode
Are you fit to be a founder? (1:06:37)
- Patent holders often lead with "I need X million and I'll change humanity". They haven't asked whether anyone will buy it (1:06:54).
- KnowledgeX's second product, which he calls "Core" (spelt KORE on the site), uses psychometric assessments to test founder fit (1:07:24). "We didn't invent the tests; we invented a different way of analysing them."
- It looks for three personas: the Engineer (builds the product), the Hustler (sells, hears "no" and keeps going) and the Artist (1:08:27). He says research shows that having all three in the founding team raises the odds of success from 40% to 55% (1:08:11).
- It also maps team synergy, meaning where clashes are likely. It has been used with corporate teams and a government team reporting to a deputy minister (1:09:05).
Does AI make research less valuable? (1:10:28)
- Partly, he says. AI is strong at finding and analysing published sources (1:10:59).
- But research is more than that. His example is a neurosurgeon's patented brain-biopsy guide. It sets the needle's angle and depth before drilling, where the old method meant positioning by hand and re-drilling if the angle was wrong (1:11:20). AI could not have invented that, because it only knows what is published (1:13:37).
- Core thinking stays human, especially leaps in mathematics (1:13:55). He mentions recent news of AI agents cracking a hard physics problem about friction in water, then solving it again faster (1:14:03). To him this shows reasoning, but only over what has been published.
The idea, explained
Founder fit. The persona idea used in the episode is a variant of the well-known Hacker, Hustler, Hipster (builder, seller, designer) shorthand from startup culture. The strongest academic study we found used personality data from about 25,000 to 33,000 founders. It found that firms with three or more founders were more than twice as likely to succeed as solo founders, and that personality-diverse teams did better. Its six types are named differently from the episode's three, but the message is the same: the mix of people matters (McCarthy et al., 2023).
Psychometrics measures traits through structured questionnaires. Big Five personality tests are the most widely used in research. They are tools for conversation and team design, not verdicts.
AI and research, as of late 2026. Recent events show both sides of this point. In September 2026 OpenAI announced that its AI agents, running about 10,000 at a time, had resolved a version of the Navier–Stokes problem, a famous open question in fluid dynamics. The proof was confirmed by the Lean proof checker, though human review was still under way (Science News). The work built on human mathematicians' progress on a closely related problem. The lesson matches the episode: AI is becoming a powerful research tool, while choosing the problem, designing the instrument and owning the result stay human.
Try it
Tap the triangle to place up to four founders. The closer a founder sits to a corner, the stronger that trait. It is a conversation starter, not a test.
Closing
The episode ends at 1:15:39. Its thread through all eight chapters: knowledge you own is what lets a company last. Getting there takes patient money, the right pipeline from paper to patent, and people who can translate research for those who act on it.