AI Circle Daily
https://aicircle.news/archive/2026-09-10?category=leaders&lang=en
Scan to read this page
After OpenAI claimed 10,000 of its agents solved the 90-year-old Navier-Stokes problem in 88 hours and a mathematician challenged the claim, Sam Altman issued a public statement on the dispute.
Take: This is the week's biggest claim about AI's limits: if true, math gets rewritten; if not, OpenAI's credibility takes the hit. Altman stepping in shows he knows the strategic stakes.
OpenAI says 10,000 of its agents cracked the 90-year-old Navier-Stokes problem in 88 hours, but a mathematician has already challenged the claim.
Take: Whoever proves this first owns the narrative that AI can do real science, not just code. The pushback is already here, so expect a credibility war, not a victory lap.
Jacob Coxon, who just quit Anthropic, told WIRED there's a "mini Manhattan project" inside the lab and that AI labs have only a few years left to make their systems safe, calling it "crunch time for humanity."
Take: A "mini Manhattan project" line from an insider cuts ten times deeper than outside criticism. The alignment camp is using exit narratives to pressure the labs, and neither OpenAI nor Anthropic escapes.
Jacob Coxon posted on X that Anthropic and OpenAI are "gambling with our lives," putting the odds of AI "killing all humans" above 10%, prompting experts to weigh in.
Take: 10% is not a rounding error, and it's the lever that pushes safety from seminar rooms to front pages. Lab comms teams are not sleeping well tonight.
Senator Ted Cruz said Elon Musk personally told him AI has a 10% to 20% chance of destroying humanity.
Take: Musk moving the risk number from private chats into Senate lore hands regulators a knife. The ones squirming are the labs pushing compute hardest.
Nvidia CEO Jensen Huang, seated next to US Commerce Secretary Howard Lutnick on a G20 stage, said AI is like the new electricity and the coming spend is "not tens of billions, but tens of trillions."
Take: This elevates the AI capex story to national-infrastructure scale. Get it and you see Nvidia isn't selling chips, it's selling the utility's seat.
Jensen Huang has declared AGI is here, but the market doesn't care, and economists say real interest rates would show it if AI were truly about to unleash explosive growth.
Take: The CEO shouts AGI and the bond market yawns, which is the most honest signal of the year. The gap between declaration and pricing is the real bubble debate.
Paul Christiano, an influential researcher focused on alignment, is joining the OpenAI Foundation's board of directors.
Take: Putting the best-known AI doomer into governance is both defense and signal: OpenAI is handing regulators and the public a safety card. Safety now has a board seat, and the accelerationists feel the heat.
Automattic's board forced CEO Matt Mullenweg into a leave of absence, and Mullenweg said in a company Slack message that it was against his will.
Take: A founder-versus-board fight in the open is both a governance crisis and a power signal. The WordPress ecosystem and Automattic's next move both hinge on how this brawl ends.
In an internal message, Mullenweg claimed Automattic CFO Mark Davies "conspired" with board members to place him on paid leave.
Take: "Conspired" is a nuclear-grade accusation that drags an internal governance war into public view. What investors fear now isn't earnings, it's a control vacuum.