Pinned post
Julius Adebayo
Verified account This account is verified. Learn more@juliusadml
building interpretable intelligence @guidelabsai
1K Following 2.5K Followers 2.8K Posts
Latest posts 5
the overgeneralization and bastardization of the bitter lesson essay is grating. Attention is definitely not all your need. Just because bubble sort is simple doesn't mean you should scale it. Scaling actually demands new algorithmic primitives.
The counter reformation
Brute force computation is not enough for intelligence. Clever new methods are needed. The science of AI is far from finished. IMHOView quoted post on X
"I could not get innovation into use (or even seriously considered) when I was in MAI," incredible quote about the state of 'research' in this era of scaling.
The End of AI Frontiers hunch.net/?p=13764006 .View quoted post on X
There's a bit of a train wreck here that's painful to express. Nevertheless, it's a significant event for me and quite a few other folks involved so I wanted to memorialize a bit.
impressive performance engineering. RSI I can actually understand
Running AI is expensive! @AnthropicAI's amended S-1 today shows $895B committed to compute and infrastructure but with Q3 revenue $17B and $567M loss, mostly from short-term compute.
@makermaker_ai can make it cheaper, better and faster with RSI. makermaker.ai/breakthroughs/… 🧵1/7View quoted post on X
