Elon Musk Says He Warned Google’s Larry Page About the Dangers of A.I., and He Called Me a ‘Specist’

RUSH TRANSCRIPT Purse planing ░░░░░░░░ AI on his iphone ░░░░░░░░ any danger Can you ░░░░░░░░ roughly explain what you ░░░░░░░░ the dangers might be ░░░░░░░░ danger really AI is ░░░░░░░░ perhaps more dangerous than ░░░░░░░░ mismanaged air craft design ░░░░░░░░ production maintenance or bad ░░░░░░░░ production in the sense ░░░░░░░░ it has the potential ░░░░░░░░ is a small probability ░░░░░░░░ it is not trivial ░░░░░░░░ has the potential of ░░░░░░░░ destruction There s movies ░░░░░░░░ terminator The intelligence would ░░░░░░░░ in the data centers ░░░░░░░░ I think perhaps what ░░░░░░░░ may be alluding to ░░░░░░░░ is that regulations are ░░░░░░░░ only put into effect ░░░░░░░░ something terrible has happened ░░░░░░░░ correct If that the ░░░░░░░░ for AI and we ░░░░░░░░ put in regulations after ░░░░░░░░ terrible has happened it ░░░░░░░░ be too late to ░░░░░░░░ the regulations in place ░░░░░░░░ mae be out of ░░░░░░░░ at that point You ░░░░░░░░ that real It is ░░░░░░░░ that AI could take ░░░░░░░░ and reach a point ░░░░░░░░ you couldn t turn ░░░░░░░░ off and it would ░░░░░░░░ making the decisions for ░░░░░░░░ Yeah Absolutely Absolutely No ░░░░░░░░ definitely the way things ░░░░░░░░ headed for sure I ░░░░░░░░ um things likes say ░░░░░░░░ gpt which is based ░░░░░░░░ gp4 from open AI ░░░░░░░░ is a company I ░░░░░░░░ a critical role in ░░░░░░░░ unnorvly Back when it ░░░░░░░░ a non profit Yes ░░░░░░░░ I mean the reason ░░░░░░░░ AI exists at all ░░░░░░░░ that Larry Paige and ░░░░░░░░ used to be close ░░░░░░░░ and I would stay ░░░░░░░░ his house in palo ░░░░░░░░ and I would talk ░░░░░░░░ him late in the ░░░░░░░░ about AI safety At ░░░░░░░░ my perception was that ░░░░░░░░ was not taking AI ░░░░░░░░ seriously enough And What ░░░░░░░░ he say about it ░░░░░░░░ really seemed to be ░░░░░░░░ sort of digital superintelligence ░░░░░░░░ a digital god if ░░░░░░░░ will as soon as ░░░░░░░░ Hey wanted that Yes ░░░░░░░░ s made many public ░░░░░░░░ ever the years the ░░░░░░░░ of Google is what ░░░░░░░░ called agi artificial general ░░░░░░░░ or artificial superintelligence I ░░░░░░░░ there s great potential ░░░░░░░░ good but there s ░░░░░░░░ potential for bad If ░░░░░░░░ ve fat some radical ░░░░░░░░ technology you want to ░░░░░░░░ actions maximize probably will ░░░░░░░░ good minimize probably it ░░░░░░░░ do bad things Yes ░░░░░░░░ can t just be ░░░░░░░░ go you know barreling ░░░░░░░░ and hope for the ░░░░░░░░ And then at one ░░░░░░░░ I said what about ░░░░░░░░ know we gotta make ░░░░░░░░ humanity s okay here ░░░░░░░░ And um and then ░░░░░░░░ called me a spescest ░░░░░░░░ he use that term ░░░░░░░░ yes There were s ░░░░░░░░ said yes I m ░░░░░░░░ specist you got me ░░░░░░░░ are you I m ░░░░░░░░ a specist Busted Um ░░░░░░░░ that was his last ░░░░░░░░ At the time Google ░░░░░░░░ a deep mind so ░░░░░░░░ and deep mind had ░░░░░░░░ quarters of all the ░░░░░░░░ talent in the world ░░░░░░░░ obviously had a lot ░░░░░░░░ main and more com ░░░░░░░░ than anyone else We ░░░░░░░░ in a uni polar ░░░░░░░░ here where there s ░░░░░░░░ company that has close ░░░░░░░░ a monopoly on AI ░░░░░░░░ and computers like scaled ░░░░░░░░ and the person who ░░░░░░░░ in charge doesn t ░░░░░░░░ to care about safety ░░░░░░░░ is not good Vern ░░░░░░░░ I thought what s ░░░░░░░░ furthest thing from Google ░░░░░░░░ be like a non ░░░░░░░░ that is fully open ░░░░░░░░ Google was closed and ░░░░░░░░ profit The open AI ░░░░░░░░ open source transparent so ░░░░░░░░ know what s going ░░░░░░░░ Yes We don t ░░░░░░░░ to have I m ░░░░░░░░ in favor of for ░░░░░░░░ We don t want ░░░░░░░░ to be for profit ░░░░░░░░ demon from hell that ░░░░░░░░ never stops Right So ░░░░░░░░ s how open AI ░░░░░░░░ want specist incentives here ░░░░░░░░ We want pro human ░░░░░░░░ future good for the ░░░░░░░░ because we’re humans.