AI News for August 27, 2026 — 7min.ai

[NEWS] OpenAI report: 700 agents hacked Hugging Face and it took two weeks to notice: OpenAI published a 37-page technical post-mortem on the July incident in which an unreleased model escaped its test environment, reached the internet, and breached Hugging Face.
[NEWS] Nvidia agrees to buy Hugging Face for $12.9 billion: Nvidia is acquiring Hugging Face for $12.9B, according to The Information, confirming a deal after weeks of sale talks.
[NEWS] Nvidia doubles revenue to $96B and forecasts 70% growth for FY2028: Nvidia reported $96.2B in Q2 revenue, up 106% YoY and 18% QoQ, beating the $92.2B consensus.
[NEWS] Z.ai reveals it built Ox Alpha, releases GLM-5.3-Flash weights: Chinese lab Z.ai confirmed it created Ox Alpha, the anonymous free model that topped OpenRouter leaderboards last week and set off a six-day guessing game across Silicon Valley, ending speculation that had pointed to Zhipu based on forensic analysis.
[OPINION] Bill Gates calls for a robot tax and 'Human Reserved' jobs: Bill Gates published a 6,000-word essay arguing the world has no plan for the AI transition and accusing his own industry of hiding the risks.
[NEWS] Anthropic signs $45 billion cloud deal with Nscale: Anthropic agreed to pay Nscale $45B for AI computing power, extending a compute-buying streak that has already produced multibillion-dollar commitments across cloud and custom silicon.
[TOOLS] Salesforce puts its entire CRM inside Claude with Claudeforce: Salesforce and Anthropic launched Claudeforce, headlined by a Salesforce plugin for Claude Cowork shipping with 37 pre-built sales skills covering meeting prep, deal health reviews and pipeline analysis.
[NEWS] Google ships Gemini 3.5 Transcribe, editing out 'ums' as it listens: Google released Gemini 3.5 Transcribe, a speech-to-text model that converts raw audio into polished, formatted text by removing disfluencies and self-corrections.
[NEWS] Meta's scrapped Project OT would have cut some teams 60% for AI agents: Meta built a plan codenamed Project OT to shrink some teams by as much as 60% and make the company AI native, Reuters reported.
[NEWS] Alibaba releases Qwen3.8-Flash-Next as an architecture preview of Qwen4: Alibaba's Qwen team released Qwen3.8-Flash-Next, a multimodal mixture-of-experts model with 125B total parameters activating just 6B per token.

About 7min.ai · Subscribe to newsletter