🛎️ OpenAI Bags a Fields Medalist

Good Morning, AI Enthusiasts!

The smartest model in the room might not win. The winner might be whoever controls the money, machines, and millions of users.

FAME

OpenAI Bags a Fields Medalist

👀 What's happening: Hours after winning the Fields Medal for work on the four-decade-old André-Oort conjecture, Jacob Tsimerman announced he would join OpenAI’s safety team. The University of Toronto mathematician has warned that AI could wipe out humanity, co-authored a taxonomy of AI extinction scenarios, and predicted models could surpass human mathematicians at theorem proving within two years. OpenAI researchers immediately amplified the hire online.

🌍 How this hits reality: OpenAI got more than a safety researcher. It got the Fields Medal’s shine at exactly the moment it needs investors, regulators, and customers to believe its breakneck commercialization comes with serious intellectual supervision. The company did not earn the medal, but it can borrow its authority. A mathematician’s warning about extinction became a polished trust badge for the lab accelerating the technology he fears.

🛎️ Key takeaway: OpenAI put the fire alarm on payroll, borrowed its medal for the pitch deck, and called the whole thing responsible AI.

NEW TECH

BigMac Stops GPUs Waiting

👀 What's happening: Xiaohongshu’s (RedNote) dots infra team has open-sourced BigMac, a multimodal training system designed to kill a brutal tradeoff: train faster and activations blow through GPU memory, or save memory and leave costly chips idle. BigMac preserves the LLM pipeline while nesting vision encoders and media generators around it. On Qwen3-30B-A3B workloads, it raised training speed by 1.08x to 1.9x while keeping peak memory stable as batch sizes grew.

🌍 How this hits reality: Multimodal AI has been charging customers for infrastructure waste. Labs bought more GPUs to compensate for pipelines that made chips wait, then passed the bill into token prices. BigMac turns idle time and bloated activation storage into a scheduling problem. If it holds up in commercial stacks, the same cluster can train larger multimodal models faster, run more workloads, and cut the infrastructure cost behind every generated token.

🛎️ Key takeaway: BigMac did not create cheaper intelligence. It found the GPU waste inflating its price, and gave token economics a reason to fall by multiples.

LLM

Kimi Exposed the Valley's Mess

👀 What's happening: Moonshot released Kimi K3, an open-weight Chinese model that rivals leading US systems while costing far less to use. Within days, the release set off four reactions: Washington began examining whether Moonshot accessed restricted Nvidia chips or distilled US models; nearly 200 US startups asked the White House not to ban it; Nvidia defended American companies’ right to use it; and Kimi’s own API reportedly shut down under surging demand.

🌍 How this hits reality: Kimi made the valley admit what it is. Washington sees espionage. OpenAI allies see a threat to the API tollbooths funding their GPU binge. Startups see the cheaper intelligence they were promised. Nvidia sees more users who will eventually need more chips. Everyone claims to be defending innovation. Everyone means a different revenue stream.

🛎️ Key takeaway: Kimi did not reveal a Chinese threat. It revealed Silicon Valley has no shared answer for AI except that somebody else should pay more for it.

MILESTONE

Gemini Passed 950 Million

👀 What's happening: Google said Gemini has passed 950 million monthly users, nearly tripling its audience in a year and closing in on ChatGPT’s billion-user mark. The milestone comes as Google’s next frontier model release has dragged on and rivals dominate the AI launch cycle. Gemini still rides Android, Search, Chrome, Gmail, and Workspace. Google’s AI Mode has also crossed one billion users and is driving more search queries.

🌍 How this hits reality: Gemini's milestones prove that consumers are not grading model releases. They are choosing the AI already inside the phone, browser, inbox, and search box they use all day. Gemini does not need to beat OpenAI or Anthropic in every benchmark if Google can make switching feel unnecessary. The frontier model is the industry’s prestige contest. Distribution is the consumer product.

🛎️ Key takeaway: Gemini’s 950 million users do not prove Google built the best AI. They prove the best AI loses when it asks people to go somewhere else.

TOGETHER WITH MYCLAW

Run Best-in-Class AI Agents — In Seconds, Always On

MyClaw is a best-in-class AI agent platform that brings together OpenClaw, Hermes Agent, and commercial infrastructure to help businesses launch intelligent agents that understand goals, use tools, automate workflows, and deliver finished results.

We already host OpenClaw and Hermes — and we're just getting started. Claude Code, Codex, and every agent worth running are coming next. One home for all of them, always on.

Stop babysitting infrastructure. Spin up your agent, walk away, and let it work around the clock.

MyClaw — Your Agents, Hosted Right.

DAILY TL;DR

• OpenAI made ChatGPT Health available to all U.S. adults, letting users connect medical records and personal health data inside ChatGPT.
• Anthropic brought Claude voice mode to Opus and Sonnet, with app actions across Gmail, Slack, Canva, and Calendar.
• Amazon is revamping Prime Video’s homepage, recommendations, and voice control to make AI more central to streaming.
• US lawmakers are proposing mandatory AI kill switches that would let the government shut down models in severe loss-of-control scenarios.
• Uber cut 10% of customer service roles while tying the restructuring to its broader AI push.
• Patreon laid off about 20% of staff as its CEO said AI has changed how the company must operate.
• Runway launched a generative-media router that chooses image, video, or audio models based on quality, speed, and cost.
• Nvidia is sending Jetson chips to the lunar surface for GPU-powered lidar control on a future moon rover.
• Meta released an AI optimism ad soundtracked by a song about human extinction, making the campaign’s message land awkwardly.

READ MORE

Let the Future Come to Your Inbox

Stay ahead without drowning in information. We turn the most important signals across AI, tech, marketing, and future products into 5-minute reads you can actually finish.

• AI Secret uncovers what really matters in AI
• Bay Area Letters decodes tech and business shifts from Silicon Valley
• Robotics Herald tracks how robots move from labs into daily life
• Marketing Secret breaks down real growth and go-to-market playbooks
• The Hardwire explores hardware, consumer tech, and what’s coming next

TOGETHER WITH US

AI Secret Media Group is the world’s #1 AI & Tech Newsletter Group, reaching over 2 million leaders across the global innovation ecosystem, from OpenAI, Anthropic, Google, and Microsoft to top AI labs, VCs, and fast-growing startups.

We've helped promote over 500 Tech Brands. Will yours be the next?

Email our co-founder Mark directly at mark@aisecret.us if the button fails.