2026-11-22 Updated

Economy, markets, tech & the world — what moves your money

S&P 500 6,910 ▼ 3.70%Nasdaq 22,618 ▼ 5.28%Dow Jones 48,147 ▼ 1.93%IPC México 68,629 ▼ 0.12%Gold / Oro 3,515 ▼ 4.62%WTI Oil 83.41 ▲ 1.18%EUR/USD 1.0975 ▲ 0.01%USD/MXN 19.14 ▲ 0.46%US 10Y 4.9120 ▼ 2.77%BTC 60,874 ▲ 1.41%ETH 1,715 ▲ 2.16%SOL 70.38 ▼ 0.22%S&P 500 6,910 ▼ 3.70%Nasdaq 22,618 ▼ 5.28%Dow Jones 48,147 ▼ 1.93%IPC México 68,629 ▼ 0.12%Gold / Oro 3,515 ▼ 4.62%WTI Oil 83.41 ▲ 1.18%EUR/USD 1.0975 ▲ 0.01%USD/MXN 19.14 ▲ 0.46%US 10Y 4.9120 ▼ 2.77%BTC 60,874 ▲ 1.41%ETH 1,715 ▲ 2.16%SOL 70.38 ▼ 0.22%
Breaking
COP31 Climate Talks in Turkey Collapse as Oil Producers and Europe Clash Over Fossil Fuels•Binding of Isaac's AI Music Scandal: Tracks Pulled, Steam Reviews Still Furious•Private equity's hospice shopping spree comes with a bigger Medicare bill, research finds•Intelligence chief says Putin is 'not America's friend in Ukraine,' even if Trump calls him one•Suspected Sea Drone Hits Russian Vessel at Libya's Benghazi Naval Base•Las Vegas Could Become the Ultimate Robotaxi vs. Human Driver Test•Intelligence chief Clayton tells Americans to expect 'bumps in the road' as aging systems get upgrades•Nvidia jitters tempting you to swap stocks for real estate? Take a breath first•Switch 2 Black Friday: $20 off Donkey Kong Bananza and more big Nintendo games•Old GoPro in a drawer? Turn it into a dash cam or webcam•Trump Calls Wind Power Worthless, But a Bill He Backs Could Help Build It•COP31 Ends in Disaster, Even as a Super El Niño Makes Climate Change Hard to Ignore•Israeli settlers attack several Palestinian villages in the West Bank, Al Jazeera reports•Seniors Who Shouldn't Drive Are Buying Teslas for 'Full Self-Driving' That Still Needs a Driver•COP31 Climate Talks in Turkey Collapse as Oil Producers and Europe Clash Over Fossil Fuels•Binding of Isaac's AI Music Scandal: Tracks Pulled, Steam Reviews Still Furious•Private equity's hospice shopping spree comes with a bigger Medicare bill, research finds•Intelligence chief says Putin is 'not America's friend in Ukraine,' even if Trump calls him one•Suspected Sea Drone Hits Russian Vessel at Libya's Benghazi Naval Base•Las Vegas Could Become the Ultimate Robotaxi vs. Human Driver Test•Intelligence chief Clayton tells Americans to expect 'bumps in the road' as aging systems get upgrades•Nvidia jitters tempting you to swap stocks for real estate? Take a breath first•Switch 2 Black Friday: $20 off Donkey Kong Bananza and more big Nintendo games•Old GoPro in a drawer? Turn it into a dash cam or webcam•Trump Calls Wind Power Worthless, But a Bill He Backs Could Help Build It•COP31 Ends in Disaster, Even as a Super El Niño Makes Climate Change Hard to Ignore•Israeli settlers attack several Palestinian villages in the West Bank, Al Jazeera reports•Seniors Who Shouldn't Drive Are Buying Teslas for 'Full Self-Driving' That Still Needs a Driver
Economy

Mike Winer and Team Race to Build a Math Safety Net for A.I.

The New York Times
Economy

According to The New York Times, Mike Winer and his colleagues are working on a mathematical safeguard meant to keep artificial intelligence under human control. The Times reports that they worry time is running short.

The material available is thin: we have the headline and a short teaser, not the full article. That means details such as who funds the effort, how the safeguard would work, or what timeline the team has in mind are not known here, and we won't guess.

What the teaser does tell us is the shape of the idea. Rather than relying only on testing a model and hoping it behaves, the team is pursuing a formula, a mathematical way to guarantee humans stay in charge. The framing suggests the people involved see the stakes as extremely high.

The broader context is well established. As A.I. systems grow more capable, researchers and policymakers have debated how to make sure they do what people intend. Approaches range from rules and audits to technical methods that try to prove a system will stay within limits. Proof-style approaches are attractive because they promise certainty, but they are hard to apply to large, complex software.

The urgency the team feels reflects a familiar tension: capabilities tend to advance quickly, while safety methods often lag behind. Whether this particular formula can be built, and built in time, is the open question the Times piece appears to explore.

Why it matters

A.I. is being built into products, markets and infrastructure, so credible ways to keep it controllable could shape regulation, investor confidence and which companies win trust. If safety methods lag behind capabilities, the economic and political costs could be large.

AdvertisementAdSense · article

Key points

  • Mike Winer and colleagues are developing a mathematical safeguard for A.I.
  • The goal is to keep A.I. under human control.
  • The team fears it is running out of time, per The New York Times.

Written by the PixelContent newsroom from public reports. Original reporting: The New York Times.

#A.I.#AI safety#technology#research

More stories

Link copied