Published today, sourced from stories within the last 24 hours
🧠 OpenAI Model Solves 80-Year-Old Math Problem
An internal OpenAI model has reportedly disproved the Erdős unit distance conjecture, a famous problem in discrete geometry that stumped human mathematicians for eight decades. Fields Medalist Tim Gowers called it “a milestone in AI mathematics,” while University of Toronto professor Daniel Litt said it’s “the first example of a result produced autonomously by an AI that I find exciting in itself.” The AI cleverly combined existing techniques from multiple subfields to construct a full proof — though it didn’t pioneer entirely new methods. Human mathematicians have since cleaned up and extended the result. The breakthrough signals a future where AI and human mathematicians complement each other, though the rapid pace of improvement raises questions about the long-term role of humans in mathematical research.
💡 Intel Unveils “Crescent Island” AI Chip — Cheaper & Cooler Than Rivals
Intel plans to ship its new Crescent Island GPU by the end of 2026, targeting the AI inference market with a chip that uses cheaper LPDDR5 memory and air cooling instead of the expensive HBM and liquid cooling required by Nvidia’s Blackwell and AMD’s offerings. Kevork Kechichian, who leads Intel’s data center group, said the company is “starting with the basics” after its Gaudi training chip flopped. Intel is also evaluating whether a compliant version of the chip could be sold in China despite US export controls. The move is Intel’s first major AI infrastructure push under CEO Lip-Bu Tan, and investors are buying in — Intel shares are up more than 200% this year.
⚖️ Erin Brockovich Takes On Data Center Secrecy
Environmental activist Erin Brockovich has launched a new campaign demanding transparency from data center operators. Her new website features a crowdsourced map of US data centers, and she says she received nearly 4,000 community submissions in the first month alone. The top concern isn’t noise, water usage, or utility bills — it’s transparency. Brockovich describes a pattern of “projects announced after permits are already secured, developers who don’t return calls, local officials who signed NDAs before their neighbors knew a project was being considered.” She’s not anti-AI, she says, but anti-secrecy.
🤯 The Debate Over “AI Psychosis” Heats Up
Box CEO Aaron Levie sparked a viral conversation by suggesting tech CEOs are “uniquely prone to AI psychosis.” The phrase captures a growing tension: leaders are simultaneously over-investing in AI while their users are pushing back. DuckDuckGo installs surged 30% as users reject being “force-fed” Google’s AI search. Graduating students are booing AI mentions. The backlash is real, but so is the adoption — and the industry is struggling to reconcile both realities. TechCrunch’s Equity podcast unpacked the phenomenon, questioning whether this anti-AI moment could actually be an opportunity for startups.
💰 SoftBank to Invest €75 Billion in French Data Centers
SoftBank announced plans to invest up to €75 billion to build data centers in France, a massive commitment that underscores the global infrastructure race driven by AI demand. The investment would significantly expand Europe’s AI computing capacity and represents one of the largest single data center investment pledges to date.
🤖 Microsoft’s AI “Super App” Reportedly in the Works
Microsoft is reportedly building an AI “super app” that would combine GitHub Copilot, Copilot Chatbot, Copilot Cowork, and a new agentic workflow capability internally named “Autopilot” into a single interface. The move mirrors OpenAI’s own super app ambitions and could be revealed at Microsoft Build next week. It signals Microsoft’s push to consolidate its fragmented AI tooling into one cohesive experience.
📱 Meta Reportedly Developing an AI Pendant
Meta is reportedly working on an AI-powered pendant — a wearable device that could bring AI assistance into a hands-free, always-available form factor. Details are still emerging, but the move would put Meta in competition with Humane, Rabbit, and other hardware startups chasing the AI wearable space.
🏥 Microsoft Launches Copilot Health AI Preview
Microsoft has opened the preview of Copilot Health to Microsoft 365 subscribers. The AI can analyze your medical records, connect data from wearables and apps like Apple Health, and help you find doctors. First announced in March, this puts Microsoft in direct competition with health AI offerings from OpenAI and Anthropic.
💸 GitHub Copilot’s Token-Based Billing Angers Developers
GitHub’s new token-based billing model for Copilot is drawing heavy criticism from developers, with many calling it “What a joke.” The shift from flat-rate subscriptions to consumption-based pricing could significantly increase costs for heavy users and teams, and the developer community is pushing back hard.
🔬 AI Agent Bottleneck: It’s Not Models — It’s Permissions
Workday built its Sana AI agent platform to solve what it sees as the real bottleneck for enterprise AI agents: permissions. While the industry focuses on model performance, Workday argues that without proper access controls at the system-of-record layer, AI agents can’t safely operate in production. The approach addresses a growing reliability problem as enterprises move agents from pilot to production.
Published June 1, 2026. All stories sourced from reports published within the last 24 hours. Links to original sources are available from the publishers mentioned above.
Leave a Reply