Covers 1:18:32–end. Two research areas, his personal stance on AI tools, and career advice.

Research area 1: local-first software

  • Collaborative software (like Google Docs, Figma) that better protects users' data: less dependent on a single cloud provider that can lock you out of your files, giving users greater agency and autonomy. ~10 years of work combining open source, algorithm development, and formal verification. 1:18:32 1:18:52

Research area 2: cryptography for the physical world

  • Brand new area: using cryptography to prove things about the physical world, focused on sustainability. 1:18:52
  • Problem 1 — carbon emissions: for emissions numbers to inform purchasing decisions they must be correct, but today incentives favor lying, creative accounting, and greenwashing. 1:19:22
  • Problem 2 — EU deforestation regulation: importers of coffee, cocoa, palm oil, etc. must prove which plot of land products came from and check against satellite imagery that it wasn't recently deforested. 1:19:57
  • The approach: cryptographic proofs about supply chains without revealing commercially sensitive information (e.g. who a company's suppliers are, which could leak a secret recipe) — prove the accounting was done correctly while keeping supplier/customer data private. 1:19:57 1:20:30

Martin's own use of AI

  • He writes very little code these days and hasn't personally needed AI agents much; sees the impact through collaborators using AI heavily for software development. 1:20:58
  • For the book's prose, he writes every word by hand — not on principle, but because writing is how he figures things out; the thinking is the goal. AI is productive for getting feedback on ideas and testing whether they hold up to scrutiny, in both industry and academia. 1:21:25 1:21:51 1:22:14

Career advice: industry vs. academia

  • They're not mutually exclusive. Some of the best PhD students he's worked with spent a few years in industry first — real-world engineering experience gives breadth that straight-to-PhD students can lack. 1:22:36 1:23:18
  • In the other direction, academia teaches nuanced, critical, first-principles thinking — an antidote to industry's "short-circuit reasoning" ("I heard this at a conference talk, I'll just go with that"). 1:23:42
  • Recommendation: weave in and out of industry and academia rather than treating them as exclusive career paths. 1:24:09