AI News for April 14, 2026 — 7min.ai

[NEWS] Goldman Sachs CEO confirms banks are 'hyper-aware' of Mythos as European regulators join alarm: Goldman Sachs CEO David Solomon publicly stated his bank is "hyper-aware" of Anthropic's Mythos capabilities and "working closely with Anthropic," marking the first on-record confirmation of bank-level operational engagement since the emergency Treasury meeting.
[NEWS] OpenAI's leaked CRO memo reveals 'Spud' model, Frontier agent platform, and sharp attack on Anthropic: OpenAI Chief Revenue Officer Denise Dresser sent a four-page internal memo laying out Q2 strategy: a new model codenamed "Spud" that will make all products "significantly better," an agent platform called "Frontier," and an expanded Amazon partnership after Microsoft "limited our ability" to reach clients.
[NEWS] Molotov suspect faces federal charges; manifesto listed names and addresses of other AI CEO targets: Daniel Moreno-Gama, 20, now faces federal charges of attempted property destruction by explosives and possession of an unregistered firearm, carrying maximum sentences of 20 and 10 years.
[NEWS] Meta building AI clone of Zuckerberg for employee interactions: Meta is training a photorealistic AI avatar on Zuckerberg's image, voice, mannerisms, and public statements so employees can interact with a digital version of the CEO.
[RESEARCH] AI proves new mathematical results at rapid pace, marking a turning point for research: Mathematicians are using AI to discover and prove new results, compressing weeks of work into single days.
[NEWS] GPU prices hit 48%, Anthropic uptime falls to 98.95%, and Sora was killed to free compute: GPU prices climbed 48% per the Ornn Compute Price Index, up from the 20-22% gains reported last week.
[NEWS] Novo Nordisk partners with OpenAI to accelerate obesity drug development: Novo Nordisk will integrate OpenAI's AI across the company to speed up drug development, including its blockbuster obesity drug pipeline.
[NEWS] Stanford 2026 AI Index finds growing disconnect between AI insiders and the public: Stanford's annual AI Index report documents diverging views between AI experts and the general public, with rising anxiety around jobs, medical care, and the economy.
[NEWS] Microsoft tests OpenClaw-like AI agents for 365 Copilot enterprise customers: Microsoft is exploring integrating OpenClaw-style features into 365 Copilot, aiming for agents that "run autonomously around the clock." The always-on Copilot could monitor Outlook inboxes and calendars, serving up daily task lists.
[NEWS] Users accuse Anthropic of degrading Claude performance, calling it 'AI shrinkflation': Developers and power users across GitHub, X, and Reddit are accusing Anthropic of degrading Claude Opus 4.6 and Claude Code, citing worse sustained reasoning, mid-task abandonment, and increased hallucinations.

About 7min.ai · Subscribe to newsletter