Hacker News new | ask | show | jobs
The Inference Shift (stratechery.com)
20 points by swolpers 32 days ago
1 comments

If it's that much faster then perhaps it can come up with 10 solutions and rank them and still have a total better performance than a larger LLM.