Eric Drexler
Article
Eric Drexler is a recurring person in the Astral Codex Ten archive, appearing 4 times across 4 issues between June 04, 2021 and February 29, 2024. The archive places it in contexts such as “although it was popularized by Eric Drexler”; “go Eric Drexler’s argument”; “Eric Drexler’s CAIS framework”. It most often appears alongside Astralcodexten Com, Eliezer Yudkowsky, FAA.
Metadata
- Category: People
- Mention count: 4
- Issue count: 4
- First seen: June 04, 2021
- Last seen: February 29, 2024
Appears In
- Your Book Review: Where’s My Flying Car?
- Practically-A-Book Review: Yudkowsky Contra Ngo On Agents
- Open Thread 208
- Links For February 2024
Related Pages
-
- Astralcodexten Com (2 shared issues)
-
- Eliezer Yudkowsky (2 shared issues)
-
- FAA (2 shared issues)
-
- Richard Ngo (2 shared issues)
-
- @BoyanSlat (1 shared issues)
-
- @eigenrobot (1 shared issues)
-
- @JackTindale (1 shared issues)
-
- @literalbanana (1 shared issues)
-
- @NiohBerg (1 shared issues)
-
- @seanw_m (1 shared issues)
-
- A.I.M. (1 shared issues)
-
- A16Z (1 shared issues)
External Links
Source Context
Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.
Besides flying cars and nuclear energy, the last main thread of the book is nanotechnology. The idea is due to Feynman, although it was popularized by Eric Drexler. Basically: we would benefit a lot from being able to do atomic-scale manufacturing. But our factories are human scale; the tools are way too big to deal with atoms. What if you used our human scale factories to make a one-quarter-human-scale factory? Then you could use the quarter-scale-factory to make a one-sixteenth-scale factory. Repeat a few times, and you’re down to atom-scale. Nanotech, Hall says, is to nuclear energy as the steam engine was to coal - the technology that will unlock the potential of a new energy source. But like nuclear energy, nanotech has languished.
Inline links: Feynman, Eric Drexler
Richard's side of the argument is in some ways a recapitulation of Eric Drexler's argument about tool AIs. This convinced me when I first read it, but Eliezer's counterargument here has unconvinced me. Let's go over it again.
Inline links: convinced me when I first read it
The opposite of a tool AI is an agent AI. An agent AI sits and thinks and plans the same way we do. It tries to achieve goals. You might think its goal is to win a chess game, but actually its goal is to convert the world to paperclips, or whatever. These (go Eric Drexler's argument, which Richard flirts with a few times here) are the really dangerous ones.
1. Scott describes my position as similar to Eric Drexler's CAIS framework. But Drexler's main focus is modularity, which he claims leads to composite systems that aren't dangerously agentic. Whereas I instead expect unified non-modular AGIs; for more, see https://www.alignmentforum.org/posts/HvNAmkXPTSoA4dvzv/comments-on-cais
54: Eric Drexler’s 2019 report Reframing Superintelligence is one of the works of AI futurism that’s aged the best in our current LLM era, and I give him lots of credit for his successful prediction. Now he has a new AI futurism blog on Substack, AI Prospects.
Inline links: Reframing Superintelligence, AI Prospects