Bread makers don't knead dough the way a human baker does. They get to the same result via a completely different route — and that turns out to be a pretty good map for what's actually happening with AI right now.

We start with brioche, end up in a crisis of professional tacit knowledge, and find a surprisingly useful frame for thinking about what machines can and can't do — and what that costs us in the long run.

• Why "harness the model" is the real skill — and what that actually looks like in practice
• The 50/30/10 split: surprisingly good, fine, and catastrophically bad — in the same tool (and no the maths doesn't add up)
• Why test-driven development went from "boring best practice nobody does" to "the only way this works at all"
• The tacit knowledge cliff: what happens in 20 years when there are no senior lawyers who did the grunt work
• Explicit → explicable → tacit → relational: a spectrum that explains where AI taps out
• Why scarcity and skin in the game are the two things a language model structurally cannot fake
• Artisanal lawyers, peak athletes, and the industrial revolution: why commodification always leaves a pocket for the handmade

For anyone trying to think clearly about AI without falling for either the hype or the backlash.

References

• Luca Dellanna's piece on what makes humans different from AI
• Nonaka & Takeuchi (1995) – the bread maker as an example for how knowledge is encoded in processes and organisations https://www.scribd.com/document/258487259/Nonaka-I-and-Takeuchi-H-1995
• (Found after we recorded) a critique of Nonaka & Takeuchi's work on bread making machines - bread maker not as incorporation of tacit knowledge, but as fitting a social prosthesis into a rearranged world: https://www.researchgate.net/publication/215439406_The_Bread-Making_Machine_Tacit_Knowledge_and_Two_Types_of_Action

Find out more about us and our work at crownandreach.com

Hosted on Acast. See acast.com/privacy for more information.