Essays & Analysis
Notes from the field on AI-first engineering, drug discovery, and what changes when intelligent systems take work off your hands.
OpenFold3 predicts the 3D structure of any protein complex. NVIDIA AI Enterprise makes it callable as an API. I wired them together into a voice-guided molecule viewer — so anyone, not just structural biologists, can explore a folded protein.
Today's devices were built for the classical era — search, files, manual tasks. The intelligence era needs new hardware and OSes built for agents.
Tylenol is one of the most thoroughly mapped drugs in human biology. Watch the safe and toxic pathways branch, saturate, and fail — and see why AI drug discovery needs to see all of it at once.
AI drug discovery isn't waiting on better models — it's waiting on biology to catch up. The 'table of context' for a single molecule, and what it means for timelines.
AI drug discovery's biggest gap isn't models — it's pathway infrastructure. Six investments biology needs before AI can deliver on its promise.
AMD was the cost-effective alternative for a generation of tech kids. Now NVIDIA leads AI and Intel is courting Musk. Five strategic moves AMD needs to make.
The next AI winners won't be the broadest chatbots — they'll be the products that deeply understand specific professions. Why domain knowledge is becoming the moat.
AI didn't just speed up coding — it changed how we plan, build, and deliver software. The shift from ticket-driven to system-driven engineering.
When your SDK, design system, and brand guide become machine-readable curriculum, AI coding assistants stop building into a void — and the SDLC's compression gaps disappear.
Graphics cards turned game worlds into clarity. The same engine could turn drug discovery into a fluid, real-time visual workflow. The Scientific Visual OS thesis.