Meta to announce $15bn investment in bid to achieve computerised ‘superintelligence'

The Guardian

a day ago

Meta is to announce a $15bn bid to achieve computerised 'superintelligence', according to multiple reports.
The Silicon Valley race to dominate artificial intelligence is speeding up despite the patchy performance of many existing AI systems.
Mark Zuckerberg, Meta's chief executive, is expected to announce the company will buy a 49% stake in Scale AI, a start-up led by Alexandr Wang and co-founded by Lucy Guo, in a move described by one Silicon Valley analyst as the action of 'a wartime CEO'.
Superintelligence is described as a type of AI that can perform better than humans at all tasks. Currently AI cannot reach the same level as humans in all tasks, a state known as artificial general intelligence (AGI).. Recent studies have shown that many mainstream systems collapse when presented with highly complex puzzles.
Meta's attempt to leapfrog the current state of progress and target superintelligence is seen by observers as an attempt by the company to regain the initiative over AI after significant advances by competitors including Sam Altman's Open AI and Google and after Meta's huge investment in the concept of the Metaverse flopped..
In March, Wang, 28, signed a deal to build a US defence department system called Thunderforge to apply AI to US military planning and operations, starting with the Indo-Pacific and European commands. The company also received early investment from Peter Thiel's Founders Fund.
The move by Meta has led to fresh calls for European governments to launch their own more transparent research push that would ensure that the technology was developed robustly while maintaining public trust – an AI equivalent of the Cern European Organization for Nuclear Research in Switzerland.
Michael Wooldridge, a professor of the foundations of artificial intelligence at the University of Oxford, said: 'There's a good argument that there should be a Cern for AI where governments collaborate to develop AI openly and robustly so we understand and trust the technology we are developing. That's not going to happen if it's developed behind closed doors. [AI] seems just as important as Cern and particle accelerators.'
Wooldridge said the reported deal 'seems very much like an attempt by Meta to regain the initiative after the Metaverse didn't take off. They invested spectacular amounts in that and it didn't just fail, it was ridiculed.'
But he said the state of AI development was patchy and that AGI is still far away, never mind 'superintelligence'.
'We have AI that can do genuinely impressive things but then it fails on an absolutely simple task which any competent GCSE student wouldn't fail,' he said.
Dr Andrew Rogoyski, the director of partnerships and innovation at the Institute for People-Centred AI at Surrey University, said the deal would be 'a reflection that the AI companies are hoovering up AI talent from wherever they can get it, whether acquiring young AI companies, or hollowing out university AI departments'.
He said that Meta's strategy for AI was different from the likes of OpenAI and Anthropic in that AI is an enabler for Meta's business rather than its core purpose.
'It means they're not quite as desperate to achieve AGI, so they can afford to take a longer view,' he said.
It was reported that Wang will take a senior position in Meta.
Meta declined to comment. Scale AI has been contacted for comment.