Eliezer Yudkowsky

From Remilia Wiki
Jump to navigation Jump to search
Eliezer Yudkowsky
Eliezer Yudkowsky
NationalityAmerican
OccupationAI researcher, writer
Known forMachine Intelligence Research Institute, LessWrong, AI alignment, Harry Potter and the Methods of Rationality

Eliezer Yudkowsky is an American artificial intelligence researcher and writer known for his warnings about the risks of superhuman AI and for founding the rationalist community around the website LessWrong. In 2000, he co-founded the Singularity Institute for Artificial Intelligence with Brian and Sabine Atkins; it was renamed the Machine Intelligence Research Institute (MIRI) in 2013.[1]

Writing

Yudkowsky's blog posts on reasoning, cognitive bias and AI, collected as "the Sequences" and later published as Rationality: From AI to Zombies, formed the core of LessWrong, which he launched in 2009.[2] His fan fiction Harry Potter and the Methods of Rationality, serialized online from 2010 to 2015, introduced many readers to the community.[3] The thought experiment known as Roko's basilisk, posted to LessWrong in 2010, is often associated with the community.

AI risk

Yudkowsky was an early advocate of AI alignment research, the effort to make advanced AI systems pursue human goals, and argues that building superhuman AI without solving alignment would likely cause human extinction. In a March 2023 essay in Time, he argued that a proposed pause on AI development did not go far enough and called for large AI training runs to be shut down.[4] In September 2025, he and MIRI president Nate Soares published If Anyone Builds It, Everyone Dies, which makes the same case at book length.[1]

See also

References

  1. ↑ 1.0 1.1 "About the Machine Intelligence Research Institute". Machine Intelligence Research Institute. Retrieved November 11, 2026.
  2. ↑ "Welcome to LessWrong!". LessWrong. Retrieved November 11, 2026.
  3. ↑ "Harry Potter and the Methods of Rationality". HPMOR. Retrieved November 11, 2026.
  4. ↑ Eliezer Yudkowsky (March 29, 2023). "Pausing AI Developments Isn't Enough. We Need to Shut it All Down". Time. Retrieved November 11, 2026.