Wei Dai on 2012

This AI Prediction was made by Wei Dai in 2012.


Predicted time for AGI / HLMI / transformative AI:

(Hover for explanation)Types of advanced artificial intelligence: AGI (AI that can perform many tasks at a human-level), HLMI (more advanced AI that surpasses human intelligence in specific areas), and Transformative AI (AI that could significantly impact society and the world)




Opinion about the Intelligence Explosion from Wei Dai:

my overall view as being around 50/50 uncertain about whether the Singularity will be Yudkowsky-style (fast local FOOM) or Hanson-style (slower distributed FOOM)


Flycer’s explanation for better understanding:

The author is uncertain about whether the Singularity will be Yudkowsky-style or Hanson-style. The Singularity could be a fast local FOOM or a slower distributed FOOM. The author’s overall view is around 50/50 uncertain.



The future of humanity with AGI / HLMI / transformative AI:

Conditioning on Yudkowsky-style Singularity, I agree with Eliezer that the default outcome is probably a paperclipper-style UFAI, and disagree with him on how hard the FAI problems are (I think they are harder). Conditioning on Hanson-style Singularity, I agree with Hanson’s competitive evolution / burning of the cosmic commons scenario, but disagree with him in that I think that would be a terrible outcome rather than an ok outcome.


Flycer’s Secondary Explanation:

The author agrees with Eliezer’s prediction of a paperclipper-style UFAI as the default outcome in a Yudkowsky-style Singularity, but disagrees with him on the difficulty of the FAI problems. In a Hanson-style Singularity, the author agrees with the competitive evolution/burning of the cosmic commons scenario but disagrees with Hanson on the outcome, believing it would be terrible rather than okay.




Wei Dai is a renowned computer scientist and cryptographer, known for his contributions to the development of digital currencies and cryptographic protocols. Born in China in 1972, Dai moved to the United States at the age of four and grew up in the Washington D.C. area.

He attended the University of Maryland, where he earned a Bachelor’s degree in Computer Science in 1993.After graduation, Dai worked as a software engineer at Microsoft for several years, where he developed a keen interest in cryptography and digital currencies. In 1998, he published a paper titled “B-Money”, which proposed a decentralized digital currency system that would allow users to conduct transactions without the need for a central authority. This paper is widely regarded as one of the earliest precursors to Bitcoin and other cryptocurrencies.

Dai continued to work on cryptographic protocols and digital currencies throughout the early 2000s, publishing several influential papers on topics such as secure multiparty computation and digital signatures. In 2009, he co-founded the cryptocurrency project “Bitcoin Spinoff”, which aimed to create a more secure and scalable version of Bitcoin.

Today, Dai is recognized as one of the pioneers of the cryptocurrency movement, and his work continues to inspire and inform the development of new digital currencies and blockchain technologies. He remains an active member of the cryptography and computer science communities, and is widely respected for his contributions to the field.





Source: https://www.lesswrong.com/posts/WCYK7B28SZ7uJxftD/how-can-we-ensure-that-a-friendly-ai-team-will-be-sane 


Keywords: Singularity, UFAI, FAI problems