HNHacker News
TopNewBestAskShowJobs

yuhfdr

7 karma · joined October 14, 2024

submissionscomments
yuhfdr··on DeepSeek: Advancing theorem proving in LLMs through large-scale synthetic data
Sloppyjoes always assume the best case for llms and worst case for their counter example of choice.

10 trillion llms, powered by a Dyson sphere, that still output slop is still worse than one unappreciated post doc.

yuhfdr··on DeepSeek: Advancing theorem proving in LLMs through large-scale synthetic data
Same with code generation.

But generating useless code, or proofs, just to discard them is hardly a consequence and externality free effort.