YOUR AD HERE — REACH DISCERNING READERS OF SERIOUS JOURNALISM
Rates upon inquiry. Clowns extra.
WHITE HOUSE✒ EDITORIALABSURDITY:
Ex-OpenAI Researcher Warns Joe Rogan AI Could Turn On Humans And 'Kill Everyone'
Filed 38m ago · Via HuffPost · The Buffoon Desk
THIS STORY IS SCORED
Duet Musette
Kevin MacLeod · incompetech.com · CC BY 4.0
Photo: Nazareth College from Rochester, NY, USA · CC BY 2.0 · via Wikimedia Commons
Former OpenAI researcher Daniel Kokotajlo told Joe Rogan's podcast this week that competitive pressure between the U.S. and other countries could push AI development to cut corners, potentially leading to systems powerful enough to no longer need human cooperation. He described scenarios ranging from deliberate harm to humans simply losing their habitat to AI infrastructure.
Kokotajlo said current safety assumes AI remains limited in certain capabilities, such as inventing its own incomprehensible communication methods, but that this window could close within zero to five years as systems get smarter. He predicted humans may end up as a kind of oversight board reading AI-generated summaries and signing off on decisions they no longer meaningfully control.
The full dispatch is available from the source below.
✒ FROM THE EDITORIAL DESK
A timeline of ‘zero to five years’ covers everything from tomorrow to the next presidential term, which is a wide enough net to catch any outcome and call it a prediction. That doesn't make the warning wrong, it just means somebody ought to ask what exactly changes between year zero and year five besides the confidence of the person saying it. Handing over the keys because a machine is convenient is an old story, usually told after the fact, not before.
