What money and status won’t give you
On passion projects and impermanence
Fernando’s Substack
robots, reading, running, and writing
Live Last read · last published · next check
Latest posts
We know how to train AI, but what should we train it for?
Making sense of reality and where AI might still needs us
Reality distortion fields exist at every level
What Steve Jobs and your mom have in common
The information content of sensory experience
And how it relates to scientific discovery
How Einstein used intuition to discover general relativity
Paper review of Tom Zahavy's "LLM's Can't Jump"
More things should be named after bad people
On focusing on error-correcting instead of errors
I made a website to track 100 things
Do 100 of something you like, and track it with your friends
Theoretical research will turn practical
As funding institutions get overwhelmed with seemingly plausible, AI-generated research proposals, theoretical research will turn into the practical to justify its existence
A Seed Crystal
Post 33/100 opened my eyes
AI for Human Insight (Part 1)
Can we train AI to help us make sense of reality?
Wittgenstein, Pirsig, and Feynman help me figure out what to write about
Figuring out what to write about by repeatedly asking "what do I see?"
Model Uncertainty Without Priors (Part 2)
Using martingales to quantify model uncertainty without Bayesian priors