5 ms·

I was very involved with the Less Wrong community starting when it was the joint Overcoming Bias blog between Yudkowski and Hanson through the splinter off and

I was very involved with the Less Wrong community starting when it was the joint Overcoming Bias blog between Yudkowski and Hanson through the splinter off and into the formation of MIRI (fmr: Singularity institute) around 2008-2011 or so. I think it comes down to a few factors. 1. Deep Learning took off and took the narrative as the "right path" to "General AI." Back in its heyday all the LW folks (including myself) were really focused on approaching AI from the perspective of Pearl's Causality and the work of Ray Solomonoff 2. There were no practical products, tools, frameworks or research that came out of that group that could either be easily applied to existing business or demonstrated SOTA on any computing domain. 3. AGI/GAI/HLAI/etc... is generally still (unfairly IMO) seen as the domain for cranks and zealots - which was the primary end-goal of the LW crowd, where "rationality" was the practical middle-step. 4. It kind of turned into a mini-cult around Yudkowski, which IMO turned people off [0] https://intelligence.org/ https://intelligence.org/

Verdex 6y ago
> It kind of turned into a mini-cult around Yudkowski I think this transformation had already happened when I first heard of rationalists. Some of the things that they were saying seemed vaguely interesting. But I couldn't quite put my finger on something being off. At some point I came to the conclusion that Yudkowski was some sort of cult leader like figure. A cult around rational thinking. Isn't that something. Then later tiger king came out and I realized that apparently everything can devolve into forming cults.
- TheOtherHobbes 6y ago
  Humans can turn (almost) anything/anyone into a tribal fetish.
mordymoop 6y ago
The rationalist community and the number of people who call themselves rationalists has actually grown enormously (exponentially, in fact) since those early days. Modern rationalists exist in all walks of life. The early Yudkowsky-centered rationalist movement is a historical, not a modern depiction.
saalweachter 6y ago
I would also add in the toolmaker's problem. Rational thinking, like programming, is a tool. When you turn programming on itself, you get metaprogramming and new programming languages. Which is cool for programmers, and many a very skilled programmer has fallen down that rabbit-hole, but neither particularly profitable nor necessarily all that useful. To get the big bucks and make a big -- direct -- difference in the world, you need to combine the tool, programming, with domain knowledge and real-world problems. Likewise, if you turn rational thinking on itself, you get metacognition, thinking about thinking, and rationalists. It's neat, but at the end of the day...

Verdex 6y ago

> It kind of turned into a mini-cult around Yudkowski I think this transformation had already happened when I first heard of rationalists. Some of the things that they were saying seemed vaguely interesting. But I couldn't quite put my finger on something being off. At some point I came to the conclusion that Yudkowski was some sort of cult leader like figure. A cult around rational thinking. Isn't that something. Then later tiger king came out and I realized that apparently everything can devolve into forming cults.

TheOtherHobbes 6y ago
Humans can turn (almost) anything/anyone into a tribal fetish.

mordymoop 6y ago

The rationalist community and the number of people who call themselves rationalists has actually grown enormously (exponentially, in fact) since those early days. Modern rationalists exist in all walks of life. The early Yudkowsky-centered rationalist movement is a historical, not a modern depiction.

saalweachter 6y ago

I would also add in the toolmaker's problem. Rational thinking, like programming, is a tool. When you turn programming on itself, you get metaprogramming and new programming languages. Which is cool for programmers, and many a very skilled programmer has fallen down that rabbit-hole, but neither particularly profitable nor necessarily all that useful. To get the big bucks and make a big -- direct -- difference in the world, you need to combine the tool, programming, with domain knowledge and real-world problems. Likewise, if you turn rational thinking on itself, you get metacognition, thinking about thinking, and rationalists. It's neat, but at the end of the day...