For Silicon Valley, AI isn't just about replacing some jobs. It's about replacing all of them

The Guardian

12-05-2025

I recently found myself at a dinner in an upstairs room at a restaurant in San Francisco hosted by a venture capital firm. The after-dinner speaker was a tech veteran who, having sold his AI company for hundreds of millions of dollars, has now turned his hand to investing. He had a simple message for the assembled startup founders: the money you can make in AI isn't limited to the paltry market sizes of previous technology waves. You can replace the world's workers – which means you can capture their salaries. All of them.
Replacing all human labour with AI sounds like the stuff of science fiction. But it is the explicit aim of a growing number of the tech elite – and these are people who lack neither drive nor resources, who have deep pockets and even deeper determination. If they say they want to automate all labour, we should take them at their word.
This is generally an aim that's only admitted to behind closed doors, for obvious reasons. There's little that will summon the pitchforks quicker than telling people you're trying to take away their jobs. But a company called Mechanize last month bucked the trend and said the quiet part out loud. Their vision is 'the full automation of the economy', a vision they've convinced some of the biggest names in Silicon Valley to fund, including Google's chief scientist, Jeff Dean, and popular podcaster Dwarkesh Patel.
Is automating all jobs really feasible? Elon Musk certainly thinks so. The rise of AI and robotics will mean 'probably none of us will have a job', he said last year. Bill Gates thinks humans soon won't be needed for 'most things'. Massive labour replacement has also been predicted by godfather of AI Geoffrey Hinton and billionaire investor Vinod Khosla. These are hardly fringe voices that have no idea what they're talking about.
Some careers are obviously safe from robot takeover. Taylor Swift is not in danger. Nor is Harry Kane. Nor, for that matter, is Keir Starmer, or the as-yet-unnamed next archbishop of Canterbury. Famous artist, sportsperson, politician, priest – perhaps the four jobs that are the most resistant to automation. Unfortunately they're not open to all of us.
Today's technology cannot replace all human labour. AI makes mistakes. Robots lack coordination, dexterity, versatility. So that's something. But there is lots that cutting-edge technology can already do. And there are good reasons to think it will continue to improve – fast.
GPT-4, one of OpenAI's large language models, was already scoring in the top 10% on the bar exam back in 2023. Their more recent models are better at coding than their own chief scientist. Freelance writing jobs plummeted when ChatGPT was released; the same happened to graphic design jobs with the arrival of AI image generators. Driverless cars are everywhere in San Francisco. As Sam Altman himself said: 'Jobs are definitely going to go away, full stop.'
While AI grabs most of the headlines, robots are advancing rapidly too. And where AI threatens white-collar jobs, robots target physical labour. One type of humanoid robot is already being tested in BMW factories; another managed to master more than 100 tasks that would usually be done by human store workers. Companies plan to start testing robots in the home this year. The Silicon Valley vision for the labour market is remarkably simple: AI does the thinking, robots do the doing. What place do humans have in this arrangement?
Up until very recently, AI researchers thought that artificial general intelligence (AGI) – that is, AI that can perform essentially all cognitive tasks at human level – was a long way off. Not any more. Demis Hassabis, the head of Google DeepMind, now thinks 'it's coming very soon' – less than five to 10 years wouldn't surprise him.
Of course these predictions may be wrong. Perhaps we're headed for another AI winter; perhaps the chatbots will stop improving, the robots will keep falling over, the funding will move on to the next big thing in tech. I don't think so, but it's possible. But that's not the point. The question here isn't whether the legions of tech CEOs and billions of dollars of funding being poured into near-total labour automation will achieve what they're trying to achieve. The question is why they're trying to achieve it at all, and how the rest of us feel about it.
The generous answer is that they genuinely believe a post-labour economy will mean huge economic growth and vastly improved global living standards. The obvious question is what, historically speaking, suggests that the benefits of this growth would be distributed evenly.
The less generous answer is that it's about what it's always about: money. Venture capitalist Marc Andreessen once famously said: 'Software is eating the world.' Up until now there's only been so much it could eat. Whatever software you built, you still needed people to do most of the world's work, with the labour market itself tantalisingly out of reach for ambitious tech execs. But now Silicon Valley sees an opening. A chance to own the entire means of production. And it wouldn't be Silicon Valley if it didn't try to seize that chance.
Ed Newton-Rex is the founder of Fairly Trained, a non-profit that certifies generative AI companies that respect creators' rights, and a visiting scholar at Stanford University