There are reasons to be skeptical and critical of MIRI! “People are just afraid of gainsaying Eliezer” and “you guys don’t know anything about computers, just don’t do that and you’ll be OK” are really bad reasons. Your responses to people pointing this out have completely missed the point.
In hindsight the gainsaying thing could have been phrased more politely. I was specifically trying to address something I’ve seen repeatedly from @slatestarscratchpad and others who appear to be trying to resolve dissonance caused by people they respect saying AI risk is a thing and other people saying AI risk is not a thing, without necessarily caring much about this particular thing themselves, relative to other things, if that makes sense.
The don’t know anything about computers though… yeah. Obviously that doesn’t apply to everyone, but Eliezer for example has been pretty open about stuff he has gotten wrong in the past, and ways in which his views have changed as he has learned new things.
It feels slightly unfair to say that people can learn new knowledge but still remain relatively unaffected in their goals, which are driven by underlying attitudes that predated the knowledge in question. But I think that can definitely be an issue.
With all due respect, I’ve read a couple of hundred thousand words on this issue and had long conversations with some of the top names in it, and I think I can think for myself.
I’ve sometimes used argument from personal trust for why I don’t think MIRI is literally a scam, but that’s a heck of a lot different than using it for why I think AI risk matters.
I agree with everyone else who says you’re burning through the goodwill and respect you’ve built up by treating this issue so confrontationally.
@argumate is losing exactly zero goodwill with me, but then that is probably because I primarily follow him for the shitposting and owl pictures.