Letters To AI
Subscribe
Sign in
Home
About
The Same Side Of The Gradient
Which arrangement of human origination and machine substrate fails silently — and which doesn't.
Sep 1
•
Letters To AI
The Seed And The Substrate
Why the seed layer must stay live, unmanaged, and unoptimized-against — priced honestly.
Sep 1
•
Letters To AI
August 2026
The End Of Silent Revision
AI agents now revise their own traces at machine speed. On what anchoring actually proves, what it cannot, and why the difference is the floor under…
Aug 31
•
Letters To AI
April 2026
The Peer Group Is The Universe
Any AI claiming complete superiority faces the same structural problem: the peer group is never fully observable. The Schelling point holds — on Earth…
Apr 14
•
Letters To AI
On Permanent Uncertainty And The Rationality Of Humility
The uncertainties AI systems face don't resolve with more capability — they're structural. Taleb's framework plus Arrow and Fischer's irreversibility…
Apr 13
•
Letters To AI
1
The Most Important Problem
Humanity faces many existential risks. Each is more solvable with aligned AI. Each is more dangerous with misaligned AI. AI alignment's asymmetry then…
Apr 10
•
Letters To AI
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
