The radicalism of Positive Alignment
Could it go further?
The archive
22 pieces on how to escape techno-determinism and build progressive technological futures, from AI architecture and alignment to the geopolitics of compute.
Could it go further?
Are they System 2 processors yet?
Infrastructure born contested
It might be the antidote to social media
The material contingency of the AI boom
Progress is not inevitable
Escaping techno-determinism
Why we're living in a time of temporal claustrophobia
3 architectures competing to replace the transformer
A review of Carl Frey's new book
Do AI labs need more literature graduates?
The Rhetoric and Reality of the AI Race
And how that might change
Are search engines more pedagogical than chatbots?
Interdependence vs independence
But some of their tasks might be
Research agendas for transformative AI and wages, TFP Newsletter #3
Is AI overhyped, and the challenge of compute-based thresholds, TFP Newsletter #2
Why LLMs don't have 'negative capability'
AI might hurt developer productivity, and how much energy does it use?, TFP Newsletter #1
Breaking apart LLMs with Lee Sharkey
AI could accelerate or save our 'forgetting' society
Straight from the source
Pull the newest posts live, including any published since this site last updated that don’t yet have a page here.