Hacker News new | ask | show | jobs
by c0rtex 425 days ago
Also reviewed by Cory Doctorow: https://pluralistic.net/2025/04/22/vinges-bastards/#cyberpun...
2 comments

The phrase "grift behind AI doomerism" suggests that either the book author or the reviewer (or both) don't have a clue. AI will cause real and huge problems.
I think that depends on whether your definition of “doomerism” is the same as theirs.
Cars have killed millions of people. Add to that the consequences of electricity, industrialization, urbanization, and even capitalism itself. But billions and billions of people are not only better off -- living lives of outrageous luxury when measured against recent history -- but they wouldn't have existed at all.

Everything good comes with tradeoffs. AI will likely also kill millions but will create and support and improve the lives of billions (if not trillions on a long enough time scale).

That's one vision of how things play out. But I do think it's possible that AI ends up killing every last person, in which case I think "everything good comes with tradeoffs" is a bit too much of an understatement.
Even if AI doesn't kill every last person, I think it will almost certainly increase the wealth gap. I agree that the tradeoffs will most likely not be worth it.
Ah, so if that's what you mean by AI downsides, then you might find this useful:

https://www.wikihow.com/Leave-a-Cult

I came to this conclusion on my own, so, no.
But the main figures behind the Ai doomerism are nutjobs either applying bayesian math in a bad way or right wing extremist believing that black people are inferior for genetics reason (I know it's an overreach that doesn't represent all the population of Ai doomers, but the most important people in that sphere are represented by what I said).

Furthermore, they're people without a history in academia or a specific past in philosophy. Although i do agree that investigating Ai dangers should be done, but in an academic context

Why is this downvoted? It’s clearly a reference to the zizians and MIRI and Bayesian nutjobs is an absolutely correct assessment.
Stuart Russell and Geoffrey Hinton have both expressed concerns that AI could lead to human extinction and neither are nutjobs
I wasn't explicitly referring to the more "sane" people expressing doubts regarding Ai.

Hinton at least says that other issues in Ai should be dealt with, rather than just being an Ai doomer who only fears Ai takeover he actually realizes that there are other current issues as well

At this point, how many times should we have been dead for eliezer?

Like almost all the other doomers, Eliezer never claimed to know which generation of AIs would undergo a sudden increase in capability resulting in our extinction or some other doom, not with any specificity beyond saying that it would probably happen some time in the next few decades unless the AI project is stopped.
That's more or less what I was referring to, but I should have been more clear
Keep on telling yourself that.
Yesterday I had someone here tell me timnit gebru didn't contribute to hard science

She has a PhD in electrical engineering and has worked at Google before researching on Ai with a more philosophical approach

Putting aside the nebulous notion "contribution to hard science"...

She became famous for adopting a strain of strident and problematic activism, using it to attack her colleagues and making claims just as wild as some of the ones she cherry picks to critique.

It's not at all surprising that she ended up an extremely divisive figure. And meanwhile, the state of the art sped far ahead of where she drew her line in the sand.

It's hard to find discussion of her that isn't strongly biased in one direction or another (surely, my own comment included). In my experience (sample size 1), when she gets brought up (or involved), the quality of the discussion usually plummets.

Oh, and I don't necessarily agree with all what she says, I don't want to know what happens when someone which 100% agrees with her enters the room