After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
Google and Demis want different things
Google needs a frontier LLM that can compete w OpenAI and Anthropic. Demis never really wanted to build that.
He wants to build AGI and solve scientific problems They'd both be better off if he went and built another DeepMind
Attended America’s largest AI conference this week
The most inspiring talk was from Noland Arbaugh, Neuralink’s first BCI recipient. He cut through all the frothy business talk around AI—ROI, governance, security, etc
Noland reminded us that we build AI to improve human lives
@SCHIZO_FREQ I feel like they intentionally do this because the user is forced to linger on the post to count the zeros
More user seconds spent on reading the post tricks the algorithm into thinking the post is higher quality than it is