Essays
Long-form arguments on cognition, AI, neuroscience, and engineering — grounded in primary research, written to last.
The Trolley Problem Doesn't Test Reasoning. It Tests Weight.
I gave the trolley problem to three frontier AIs and asked them to answer free of human bias. Underneath all the hedging, they converged on the same answer — kill the one, save the five. The one that refused had only bolted a guardrail over that same default. The divergence between humans and AI isn't at the reasoning. It's at the pain.
How to Train Junior Engineers When the Shortcut Machine Is Always On
The gap between a senior and a junior is not years or syntax. It is a mental model — the connected understanding that lets you reason in contexts you have never seen. You do not build it with a checklist. You build it by constructing systems and by witnessing a senior reason out loud. In an age where seniors reason alongside AI, their transcripts are the apprenticeship artifact we have always lacked.
The Pipeline AI Can't Fill
Senior engineers burn out. Junior engineers get answers from AI instead of building judgment. The pipeline that used to produce engineers who could think about systems — not just write code for them — is breaking at both ends, and AI amplifies the gap instead of closing it.
The Knowledge You Never Built
Reading an AI summary gives you the answer without the architecture. The fact dangles, detached from any schema, and is soon forgotten. But the real loss isn't forgetting — it's that the sleep-dependent recombination that produces genuinely new ideas never happens, because the connections were never built in the first place.
Why AI Will Not Replace Humans (In Its Current Form)
AI has already beaten us on the randomness of the environment. It has not touched the part that actually produces human thought — emotion as input. Neuroscience has mapped that part, and it is not where AI has been building.