Barriers to trying out new ideas to build on top of LLMs are quickly falling. 

Approaches that seem like wasting inferences are the way forward to finding new powerful and innovative uses for generative AI.

https://www.sublayer.com/blog/posts/waste-inferences

LLM costs are falling. what’s the stupidest/funnest thing we should be building now?

enjoy cats, running, bread, and a bit of w̶e̶b̶3̶ onchain data

Still thinking about that idea I shared with you of a private reddit-style social network filled with AI personalities to discuss links with...

CEO @ Sublayer, previously CTO at Sayspring(acq by Adobe)