When ‘Godfather of AI’ agreed with Elon Musk; said billionaires like him are right about future of work; but there is a big problem and that is…
Geoffrey Hinton and Elon Musk do not usually land on the same side of an argument. But asked whether the tech billionaires forecasting the end of work were being melodramatic, the Nobel laureate widely called the “Godfather of AI” said they were probably right. Musk has said most people won’t have to work at all inside two decades. Similarly, Bill Gates has suggested humans may soon not be needed for most things. Hinton, in a discussion with Senator Bernie Sanders last year at Georgetown University, said those predictions are not fringe. They are likely.The catch is everything that follows. Hinton’s worry is not that AI is bad technology. It is that the economy sitting underneath it was not built to absorb what is coming. He expects “massive unemployment caused by AI”, and he does not think the people writing the cheques have worked through the consequences. His blunt line to Sanders was that if workers stop earning, “there’s nobody to buy their products”.
Why Geoffrey Hinton thinks the trillion-dollar AI spending spree runs on job cuts
Follow the money, Hinton argued, and the intent becomes obvious. Somebody has to recover the roughly trillion dollars flowing into data centres and chips, and subscription fees for chatbots won’t cover it. The return has to come from selling companies AI that does a worker’s job for less. Big Tech, in his reading, is betting squarely on replacement. The financial pressure is real enough: HSBC estimated OpenAI won’t turn a profit before 2030 and may need over $207 billion to keep scaling.Hinton also keeps pointing out that this is a design choice, not destiny. AI could do enormous good in healthcare and education. What makes it dangerous is distribution. Musk gets richer, a lot of people lose work, and as Hinton put it, “that’s on how we organize society”.
AI will create new jobs, but Hinton says nowhere near enough to replace what it kills
The standard rebuttal is that every technology shock eventually produces new categories of work. Hinton accepts the first half and rejects the second. Yes, roles like prompt engineer appear. No, he does not believe they arrive in anything like the volume of the jobs being erased. The old escape route mattered: a ditch digger could move to a call centre. Those call centre desks are the ones going first.He is careful about his own confidence, though. Predicting AI, he told Sanders, is like driving in fog, where the next hundred yards are clear and the two hundred yard mark is invisible. A year or two out, he can see. Ten years out, nobody can. At the Ai4 conference in Las Vegas this month, he went further on capability, saying AI could surpass humans at everything within twenty years, and probably a good deal sooner.
Bernie Sanders warns 100 million US jobs face AI automation risk
Sanders has tried to size the damage. An report last year from his office estimated nearly 100 million US jobs are exposed to automation, covering fast food, customer service and manual labour, but also accounting, software development and nursing. Senator Mark Warner has warned recent graduates could face unemployment near 25% within three years.Hinton’s fog is still there. The 175,479 tech jobs cut in this year are not, and the names on that list run through Amazon, Meta, Microsoft, Oracle, and plenty of others.
