OpenAI thinks they've figured out how to avoid inadvertently making the hallucinations worse with bad benchmarks. It seems like progress? openai.com/index/why-la...
sep 9, 2025, 6:19 am • 0 0 • view
OpenAI thinks they've figured out how to avoid inadvertently making the hallucinations worse with bad benchmarks. It seems like progress? openai.com/index/why-la...
the question of if one trusts an LLM circles back to if one should trust a poem
Look up is just a search engine... What am I missing? I've seen Elon avoid how many Bs by triggering a programmatic routine. But then it's no AI. And Copilot gives search results. Which is not AI.