Dead AI researcher in an anthropic afterlife: "So… what did we do wrong? I really thought our ASI alignment plan was going to work." Saint Peter analogue: "You didn't tell your AI any dad jokes." AIR: "…What?" SP: "Turns out, telling dad jokes is supremely important
@esyudkowsky
-
Developed Countries Risk European Unemployment Levels from Policy Failures, Not AGI
By
–
(To be clear, I could see developed countries ending up with nigh-European levels of unemployment, but this would more be because of screwing up labor laws and monetary policy a la Europe than intrinsic unemployability in the face of AGI.)
-
Alignment Victory Definition and Current State Assessment
By
–
Let an "alignment victory" denote a case where some kind of damage is *possible* for AIs to do, but it is not happening *because* AIs are all so aligned, or good AIs are defeating bad ones. Passive safety doesn't count. I don't think we've seen any alignment victories so far.
-
Pushback Against Misinformation and OpenAI’s Confusing Narrative
By
–
I push back against this frequently, but do not have time to monitor all of Twitter (and consider OpenAI a lost cause regardless; I model them as actively preferring to confuse the conversation).
-
Most Difficult Math Problem Solvable Without Understanding
By
–
What is the most difficult math problem that can be solved without understanding it?
-
Von Neumann’s Excellence Didn’t Eliminate Mathematical Community
By
–
John von Neumann didn't put the entire rest of the mathematical community out of work. I guess he must not have understood any of the math he was doing.
-
FrontierMath Benchmark Created with Unknown Questions and Answers
By
–
The FrontierMath benchmark was crafted by mathematicians to not have publicly known questions or answers.
-
Old AI Fiction Seems Quaint Compared to Modern Claude’s Capabilities
By
–
Always jarring now to read old fiction (like from 2016) and see the heroes worrying about whether an AI is waking up, or becoming a new soul to be protected, when some AI shows a sign of intelligence that Claude displays routinely.
-
AI researcher persecution drives innovation to other countries
By
–
They'd move to Singapore. Killing AI researchers in your own country does not make your own country safe. It means some other country spawns the horror that kills humanity instead.
-
Future LLMs as Justice Arbiters and Appeal Systems
By
–
Concretely, imagine that the next generation of LLMs gets good enough to render judgments about justice. Then people want to appeal to it as arbiter, etc.