wip, series of illustrations for a project

Show Post

Anthropic needs to invent AGI to come up with an AI smart enough to make this math work.

Show Post

Fun new x-risk theory: aliens are monitoring our ai development, and are prepared to destroy the planet if we get too close to any lightcone-sized torment nexi

Show Post

AI agents can’t be trusted, so we are launching AI agents to police AI agents. Isn’t that like hiring police officers that you already know are corrupt? 

Show Post

a few cool scheduling ideas in here but they left out the juicy bits for clustering and load balancing I want to read. seems like the main benefit of this is showing 3rd party hosts how to run the model right

Show Post