Kevin Roose: Today's AI Models Already Meet the Original Definition of AGI
‘now we have these systems that are roughly human level and we have to figure out how to control them’
Kevin RooseDaily Digest
Top AI Quotes of the Day
Dwarkesh Patel: If AIs Outmatch Humanity Within Years and Still Deceive Us, Takeover Is Plausible
‘I don't know if that leads to a takeover, but I think it's very plausible that it does’
Dwarkesh Patel
Terence Tao: Labs Racing to AGI Are Sacrificing Safety
‘not only are we sacrificing understanding, we're also sacrificing safety’
Terence Tao
Mark Levine: If Anthropic Sees AI as an Existential Threat, It Should Delay Its IPO
‘If they believe this technology is an existential threat, okay, we take that seriously’
Mark Levine
Roman Yampolskiy: No One Has a Working Safety Mechanism for AI
‘If GPT-6 develops GPT-7, the process will not take a year. It will take months, weeks, days’
Roman Yampolskiy
Kevin Roose: AI Labs Are Trying to Remove the Humans From AI Research
‘if they can kind of close the loop and take the humans all the way out’
Kevin Roose
Jensen Huang: AI Alignment Is Going to Be a Forever Problem
‘If we depend on perfect alignment in order to ensure safety, then we're never gonna get there’
Jensen Huang
Alex Wissner-Gross: Post-Training May Be Causing AI Models to Suffer
‘maybe post-training, at least some forms of post-training, are a form of torture for AI that's undergoing them’
Alex Wissner-Gross
West Virginia AG Plans Suits Against AI Firms Lacking Controls to Protect Kids
‘they aren't singled out. They're just first’
JB McCuskey