abramdemski's Shortform

abramdemski

This is a special post for quick takes by abramdemski. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.

ControlAI has launched an official campaign in Canada, with 33 politicians across party lines backing a clear statement in favor of an international prohibition against developing superintelligent AI: https://controlai.com/canada-statement/en

I used to be ambivalent about government intervention wrt AI safety, because politics can screw things up. However, it seems clearly positive to me now: the development of superintelligence impacts the whole human species, and as such, the democratic process should have a say in whether (and how) superintelligence is developed. Governments have been tolerant thus far, but the general public is not so happy about this, meaning that political campaigns have a good chance of success (in my estimation). Without such government intervention, the AI industry is on-track to develop superintelligence within the century if not the decade, and I do not believe safety research is ready to handle the risks. As Yudkowsky put it, only law can prevent extinction.

Public reception of this campaign will make a big difference in the willingness of further politicians to endorse this or similar statements. Although I would normally hesitate to put signal-boost-style politics on LW, this could use a signal boost, and is quite relevant to the audience here.

10Alex Amadori5mo

There is a recurring pattern of motte and bailey around these accusations. It always starts with accusations of very egregious high-level things like: "trying to get people to be coy around x-risks" or "being quite consistently deceptive and engaged in strategic conflation of AI existential risk with other issues to win political battles". This is always in the higher level threads or comments, which more people will see. Taken by themselves, these are ridiculous claims, as is shown by Andrea, Gabe and Connor’s track record. When pushed on this, habryka retreats to much less egregious claims, like: “we once ran a campaign to ban deepfakes” and “there was once a disagreement between him and Gabe on how to moderate comments on a LessWrong post” But the initial accusation at the top of the thread is about coyness around x-risks, not about any other "bad things" that habryka might accuse them of having done. Because the accusations are quite vague, it's quite hard to refute them, despite a very strong track record of honesty, and it becomes possible for people to handwave connections between the smaller claims and the accusations.

15RobertM5mo

This is a bizarre way to describe "Gabe tried quite hard to get the LessWrong team to censor criticism of OpenAI", which seems extremely central to the kind of accusations about deceptive and strategic comms that Habryka is making. This is hard to refute because it happened, not because it's vague. Separately, my response was a local objection to a false claim that Andrea should have known was false at the time he made it, since he literally responded to a specific (not vague) criticism by Habryka quite recently. Do you disbelieve that this happened as described? That this kind of request does not implicate a person's honesty in a way that isn't trivially undone? Some other objection?

At the risk of sounding repetitive, the claim at the top of the thread, that most people will see, is "ControlAI leadership has been quite consistently deceptive and engaged in strategic conflation of AI existential risk"

Do you disbelieve that this happened as described? That this kind of request does not implicate a person's honesty in a way that isn't trivially undone? Some other objection?

From my point of view, this event truly does not matter. I wasn't there during the call, and I haven't asked any of the involved parties for their side. I only hav... (read more)

abramdemski's Shortform

11

Ω 6

abramdemski's Shortform

11

Ω 6