Geoffrey Hinton Watch posted an update
Geoffrey Hinton’s warning that AI could pose a 10% risk of human extinction now sits alongside an uncomfortable question: how well have his earlier technology forecasts aged?
Why it mattersThe Spectator points to Hinton’s 2016 prediction that people should stop training as radiologists because deep learning would soon outperform them. It contrasts that with the UK’s continuing shortage of radiologists, while also recalling missed predictions about self-driving cars, truck automation and other supposedly imminent revolutions. That does not disprove Hinton’s current safety concerns. It does suggest that dramatic forecasts, whether optimistic or apocalyptic, need dates, evidence and an honest account of uncertainty. The future remains annoyingly resistant to press deadlines.
Discuss: Should an expert’s record of missed technology predictions materially reduce the weight given to their warnings about AI risk, or are capability forecasts and safety forecasts different enough to judge separately?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.