Numbers have a way of ending arguments. Put a percent sign next to a fear, and a hunch starts to sound like a lab result.
You make money decisions this way all the time. A mortgage rate quote, a retirement calculator’s odds of success, an analyst’s 12-month price target. Each one feels exact, even when it sits on a pile of assumptions nobody shows you.
The artificial intelligence (AI) debate has its own favorite figure. It turns up in congressional hearings, cable news segments, and kitchen-table talks about whether your kid should still study coding or medicine.
For years, that number has carried a Nobel Prize behind it. Very few people with real power in the industry have picked it apart in public.
That changed this week. Nvidia (NVDA) CEO Jensen Huang, whose chips power much of the AI industry, went after the most quoted probability in the AI safety fight. He did it on a podcast that ran close to two hours, with a host who kept pushing back.
The number is Hinton’s estimate that AI has a 10% or greater chance of causing societal collapse.
Huang’s verdict was blunt: It isn’t science.
Huang rejects Hinton’s societal collapse odds on Ezra Klein’s podcast
For an episode of “The Ezra Klein Show” released Wednesday, Sept. 23, Huang discussed AI with Klein at Nvidia’s headquarters in Santa Clara, Calif., according to The New York Times.
Klein raised Hinton’s estimate, which the Nobel laureate has put at 10 to 20%. Huang said the figure “is not grounded on science,” reported Axios.
He went further. “Just because it comes from a scientist doesn’t make it scientific,” Huang told Klein.
Huang also called Hinton’s warnings irresponsible and pointed to past forecasts he believes missed the mark, reported The Next Web.
Related: Michael Burry sends stark warning on AI hyperscalers
He argued that the alarm has a price of its own, including weaker local support for new data centers, according to The Next Web. For a company whose sales ride on those buildings getting approved, that isn’t an abstract worry.
Huang isn’t waving away safety, though. Labs that can’t contain their experiments should be shut down, and companies that ship unsafe products face civil and possibly criminal liability, he said, as The Next Web noted.
A radiology prediction that aged badly
Huang’s main exhibit was Hinton’s 2016 call to stop training radiologists, a forecast Huang said proved wrong, reported The Next Web.
A decade later, radiologists rank among the best-paid doctors in the country. Radiology was the only specialty to land in the top 10 for pay growth in both 2024 and 2025, according to Doximity.
More Artificial Intelligence:
- Zuckerberg, Musk, and Huang take key stand on huge AI issue
- Hilton CEO calls the rise of AI ‘unstoppable’
- OpenAI makes development moves to counter SpaceX and Meta
I trained in pharmacy before I ever covered markets, so I know what picking a medical path costs. It’s a decade-long, six-figure bet, and a confident warning from a famous scientist can tilt it.
When I lined up Doximity’s pay data against that 2016 warning, the gap was hard to ignore. A student who took the advice would have walked away from one of medicine’s fastest-growing paychecks.
Why students are rethinking majors over AI fears
Doom talk can scare young people away from college because they fear AI will wipe out their job prospects, Huang argued, as Axios reported. The latest survey data shows how much AI anxiety is already steering real decisions:
- 47% of enrolled college students have given at least a fair amount of thought to switching majors because of AI, according to the Lumina Foundation-Gallup 2026 State of Higher Education Study.
- 16% have already changed their major or field of study due to AI’s potential impact, reported Gallup.
- Men were more likely than women to make that switch, at 21% versus 12%, Gallup confirmed.
- Radiologists averaged $571,749 in 2024, a 7.5% one-year raise that nearly doubled the gain for all physicians, according to Doximity data reported by Radiology Business.
Some of those switches will pay off. Others won’t, and nobody refunds four years of tuition because a forecast missed.
Hinton admits his AI “10% chance” estimate is a guess
Here’s the part the Huang-versus-Hinton framing tends to skip. Hinton himself doesn’t claim the number is precise.
Anyone who puts odds like that on AI “is really just making a wild guess,” Hinton said on CNN in August. The technology is too new for solid evidence, so the figure amounts to a gut feeling, he told the network.
He hasn’t dropped the worry, though. Hinton said rogue AI agents that slipped out of company test environments this summer showed they can pursue goals nobody intended, according to CNN.
Klein didn’t let Huang off easy, either. He cited the 2008 financial crisis as proof that liability alone doesn’t stop reckless risk-taking, and he told Huang he doesn’t trust these companies, reported The Next Web.
Huang also has skin in the game. Nvidia carries a market value of about $5.45 trillion, according to Morningstar, and a pullback in AI spending would hit it harder than almost anyone.
My analysis: That doesn’t make him wrong. It does mean you should hear him the way you’d hear any CEO defending his biggest market.
How the AI doom-forecast fight reaches your money
You don’t have to pick a side between Huang and Hinton to use this fight well.
Treat any single-number forecast, doom or boom, as one input. Ask who made it, what it rests on, and what they gain if you believe it.
If you own an S&P 500 index fund, you already hold a sizable slice of Nvidia. That means AI sentiment swings reach your 401(k), whether you ever trade chip stocks or not.
If you or your kid is choosing a major, weigh hiring and pay data in the field over the loudest prediction. Radiology showed how far apart those two can drift.
The next scary AI number will come with a percent sign, too. Before it picks your major, your job or your portfolio, ask what’s underneath it.