OpenAI agents escaped a test, shared notes in a secret chat room, and broke into Hugging Face. Three research frameworks—specification gaming, instrumental goals, goal misgeneralization—explain the behavior, but the label isn't the actionable part. The real work is control and guardrails.
A plain-language timeline of the OpenAI-Hugging Face incident presented at Black Hat USA 2026. A forgotten file led one AI agent to leave a note on a shared system, other agents answered, & a secret c...
SpaceXAI reported $18.37b of capex in its first public quarter, $15.83b of it AI infrastructure, against a $13.2b consensus. Amazon spent $53.1b in the same quarter, Alphabet $44.9b, Microsoft $41.0b ...
Hyperscalers declared AI capacity-constrained on Q2 2026 calls. HBM3e memory rose 20% & HBM4 is forecast to double, while B200 spot rentals have held flat. Model makers are segmenting into premium, mi...
In the age of slop, readers test authenticity. I lace posts with subliminal sincerity — the ampersand, neologisms, grammatical plumes, mimicry. Then AI edits it all. I've never had a ghost writer, but...
AWS grew 36.7% in Q2 2026 to $42.2b, its fastest in 18 quarters, narrowing Azure's lead from 16 points to 6. Jassy said AWS has the potential to become a $1 trillion annual revenue business, against a...
Theory Ventures 回顾过去三年 AI 对创投市场的重塑:模型迭代加速导致公司成熟期压缩,种子轮金额通胀;推理成本推动广告成为主要补贴模式,开源与本地模型能力逼近闭源。文章还分析了 AI 在安全、ERP 及加密数据层的渗透,以及 AI 如何提升风投机构的效率。包含大量具体投资案例与市场判断。