LLMs really are pretty good at summarizing text. The average person I interact with summarizes my tweets in all kinds of bizarre, misleading ways (and then gaslights me about how that's my fault) but look at how Grok here successfully gets the two key pieces of the argument. LLMs are either far more honest or far more capable summarizers than the average X user.

LLMs now beat humans on math accuracy. But does that mean they're actually building understanding on top of the right foundations? Title: Do LLMs Exhibit Coherent Knowledge Structures in Mathematical Reasoning? A Perspective from Knowledge Space Theory URL: ❓ Do LLMs actually build on prerequisite knowledge to get answers right? 💡 Accuracy favors LLMs (92.5% for the best model, Qwen3-80B, vs. 79.6% for humans), but "perfect prerequisite satisfaction" tells a different story: 72.7% for humans vs. only 48.16% for the best LLM. Lots of correct answers rest on shaky foundations. ❓ Does giving them prerequisite hints help? 💡 Surprisingly, prerequisite-grounded context barely outperformed unrelated examples. That points to surface-level pattern matching rather than genuine structured reasoning over prerequisites. ❓ Do strong models at least share a consistent knowledge structure with each other? 💡 Human learner groups overlap at 0.9+ in their knowledge structure, but LLM pairs only overlap 0.38-0.6 — and stronger models diverge even further from humans. ❓ So what's the takeaway? 💡 Accuracy alone hides how differently LLMs "understand" math. Knowledge Space Theory offers a lens that exposes the fragmented structure lurking behind impressive scores. #LLMEval# #MathReasoning#

LLMs made bot replies worse and also human replies worse i spend every day working in a niche field and i post simple facts i'm learning and hundreds of people want to argue with me with ai powered confidence

LLMs with scaffolds have lagged on text-to-SQL, a task that relies on human judgment. By folding expert judgment into every part of RLVR on Tinker, @maxYuxuanZhu and @ddkang (UIUC and Bridgwater) trained the first text-to-SQL model to beat the human mark.

LLMs are non-deterministic. Pi turns your sessions into a tree you can use to time travel and explore different outcomes. Try out: - /fork: new session from a past message - /tree: navigate the session tree and continue from any point - /resume: reopen a previous session

LLMs are very two-sided for security. On the one hand, long-latent bugs are finally surfacing. On the other hand, before you choose to use a product, you can get your own code review done for incredibly low cost*! In the short term it’s incredibly bad, in the long-term it democratizes code review greatly. “Don’t trust, verify” finally becomes an actual possibility for normal people! * of course LLMs do still make up issues when asked, push any such review to write PoCs to check their analysis!