https://hackerpaper.kaipfister.de/p/2407.03321/Collected HN comments and submissions linking this paper.enMon, 02 Nov 2026 11:16:22 GMThttps://news.ycombinator.com/item?id=41290812urn:hackerpapers:mention:2407.03321:41290812Mon, 19 Aug 2024 13:19:26 GMT<p>You clearly didn&#x27;t read any of these papers. Quote from one of them &quot;Our results support the hypothesis that GPT-4, perhaps the most capable “general” LLM currenly available, is still not able to robustly form abstractions and reason about basic core concepts in contexts not previously seen in its training data&quot; Another, recent, good one https://arxiv.org/abs/2407.03321 EDIT: For people who don&#x27;t want to read the papers, here is a blog post that explains what I&#x27;m arguing in more accessible terms https://cacm.acm.org/blogcacm/can-llms-really-reason-and-pla...</p><p><a href="https://hackerpaper.kaipfister.de/p/2407.03321/">Paper and discussion history</a>. A mention is not an endorsement.</p>cs.CL