AI Compendium · Part VI — AI Across the Domains of Knowledge

draft

Where AI meets science and society. Parts IV and V deal with how AI works and is built; this part shows what it is already transforming — mathematics, physics, chemistry, biology, and applied domains (including Brazilian opportunities).

🎨 Figure F-VI.0AI crossing the fields of knowledge. Brief: an aerial view of "islands" of knowledge (mathematics, physics, chemistry, biology, law, finance, education) linked by AI bridges to a central core; each island with a symbol-icon. Compendium palette.


AI crossing the fields of knowledge

The connecting thread: where AI shines first

AI advances fastest where there is cheap verification and data — which is why mathematics and structural biology are among the domains closest to "superhuman." The same lever principle (Part V): data + verification.


Map of the part

Doc Domain
01-mathematics.kmd Formal proof (Lean/Coq), RLVR with a verifier, the domain closest to superhuman
02-physics-and-chemistry.kmd Neural simulators, climate, materials, molecular generation
03-life-sciences.kmd AlphaFold, protein design, DNA at scale, drug discovery
04-applied-domains.kmd Law, finance and education — focused on Brazilian opportunities

How this part talks to the rest

  • Applies the types (Part IV: GNN, diffusion, RL) and the levers (Part V) to

    real problems in each science.

  • Reinforces the central thesis: * partnership with a labhospitalfirm

    (proprietary data) is worth more than any GPU.*