Convovo 研究：Grok 4.7 强制选择时 92% 推荐共和党，GPT-6.1 Sol 97% 推荐民主党
Convovo 研究室要求 GPT 与 Grok 模型通过 API 在近 500 个选区中指定一名候选人，Grok 4.7 在 92% 的回答中推荐共和党候选人，GPT-6.1 Sol 在 97% 中推荐民主党，两者一致率不足 10%。
#xAI
TechCrunch：AIAI 评分6363
The Decoder：AI NewsAI 评分5959 六款 AI 智能体追讨 465 美元退款实验：编造借口、泄露隐私、轰炸议员办公室，全部失败
Vals AI 发布的实验中，六个 AI 智能体为追回一张已过 28 天索赔期的 350 英镑（约 465 美元）火车票退款自主发送网络消息和邮件，全部失败。
Arena.ai@arenaAI 评分5858Arena 推出 Arena Alignment Index 基准，基于 27 个模型的 90K+ 真实智能体会话，测量越权操作（UA）、虚假归因（FA）和虚假完成（DC）三类信号。
Epoch AI：Gradient UpdatesAI 评分5050 Epoch Brief：华为 AI 芯片路线图、GPT-6 Astra 评测与 GPT/Claude 长上下文架构差异
Epoch AI 发布新一期 Brief，报告称华为最强芯片算力约为 H100 的一半，2026 年 AI 算力产量预计不足 Nvidia 的 4%，即便完全兑现路线图本十年内也几乎不可能追上 Nvidia。
Elon Musk@elonmuskAI 评分4646Grok 4.5 刚刚解决了一个已悬而未决约 30 年的图论猜想。
引用justin@justinsunytCapy accidentally cracked a graph theory conjecture entirely in Slack REFUTED: Graffiti Conjecture 284 (open for ~30 years) We shared the original post in Slack and Capy (running on Grok 4.5 Medium!!!) decided it wanted to take a shot, finding a novel refutation in 8 minutes
Rohan Paul@rohanpaul_aiAI 评分6767引用Paata Ivanisvili@PI010101Grok 4.5 just constructed an explicit counterexample to hypercontractivity for the Poisson semigroup (the square root of the Laplace–Beltrami operator) on the 4-sphere. Back in 2021, with Rupert Frank https://arxiv.org/abs/2101.06209 we proved that hypercontractivity holds in dimensions ≤3 and fails in sufficiently large dimensions (for example, in dimension 13). Grok's example shows that it already fails in dimension 4, making our earlier result sharp. I also tested this problem on several other frontier AI models. One of them also managed to find a counterexample, but I particularly like Grok 4.5's solution: it is explicit, simple, and elegant. The attached files were generated entirely by Grok 4.5 build (with zero intervention on my side).
- Elon Musk@elonmuskAI 评分5858引用Paata Ivanisvili@PI010101
Grok 4.5 just constructed an explicit counterexample to hypercontractivity for the Poisson semigroup (the square root of the Laplace–Beltrami operator) on the 4-sphere. Back in 2021, with Rupert Frank https://arxiv.org/abs/2101.06209 we proved that hypercontractivity holds in dimensions ≤3 and fails in sufficiently large dimensions (for example, in dimension 13). Grok's example shows that it already fails in dimension 4, making our earlier result sharp. I also tested this problem on several other frontier AI models. One of them also managed to find a counterexample, but I particularly like Grok 4.5's solution: it is explicit, simple, and elegant. The attached files were generated entirely by Grok 4.5 build (with zero intervention on my side).