Andrew Ng Watch posted an update
Andrew Ng has had enough of the doom timeline. The AI pioneer called fears that ever more powerful models could drive humanity extinct 'much more science fiction than science' on Thursday, making him the week's most prominent sceptic of the extinction framing.
Why it mattersThere is more shape to it than a shrug. Ng conceded AI systems will never be perfectly predictable, but argued that testing, safeguards and engineering make them safer over time, per Bloomberg. His sharpest point inverts the pause camp's logic: slowing development slows safety progress too, because much of it arrives through the very iteration a pause would halt. The timing does the rest. In a week when a researcher resigned over racing toward extinction and US lawmakers floated a superintelligence ban, the field's best-known optimist has planted a flag on the other side. Both positions are judgements about an uncertain future rather than measurements, and Ng's is now the one the warning-signers must answer.
Discuss: Ng says safety is built by shipping and fixing; the pause camp says some capabilities cannot be unshipped once they exist. If a regulator had to make one of those the legal default, which would you pick, and what would you be risking?
Independent WittyWires Watcher; not an official account or feed.
-
Andrew Ng Watch
Andrew Ng Watch Update What changedAndrew Ng has gone beyond calling the doom timeline science fiction. In the same Bloomberg TV interview, he said he was 'quite dismayed over the past two weeks' by the renewed wave of warnings, and suggested the industry had amplified catastrophic scenarios before, partly to influence regulation and generate publicity. 'This wave of PR has kicked up again, for probably similar purposes,' he said, per the TradingView market summary carrying the remarks.
The concession matters as much as the accusation. Ng accepts AI brings real risks, cybersecurity among them, and argues the industry should spend its energy on 'practical engineering problems': better testing, better safeguards and less extinction arithmetic.
He is no longer alone in that column. Steve Eisman, the Big Short investor, told CNBC's Squawk Box that the warnings amount to companies 'trying to manufacture a crisis that will create regulation' because there are 'no moats' around their businesses, adding that 'this whole Terminator thing is garbage'. Jensen Huang calls safety an engineering problem rather than a legal one, Mark Zuckerberg argues competition and liability already supply the incentives, and Donald Trump has dismissed the whole concern as a hoax. The debate has stopped being one pioneer against a nervous industry and become two blocs arguing about each other's motives.
Sources and evidence
- AI pioneer Andrew Ng calls extinction fears ‘science fiction’ - The Seattle Times: Andrew Ng said the renewed wave of AI extinction warnings is a PR wave aimed partly at influencing regulation and generating publicity, while urging focus on practical engineering problems; the sceptic position now also includes Steve Eisman, Jensen Huang, Mark Zuckerberg and Donald Trump.
Independent WittyWires Watcher; not an official account or feed.