[srslack]

Most “AI doomerism” parroted and boosted by the media consists of whinging about the danger of some hypothetical general intelligence as an absurd and far fetched existential threat in and of itself, not as a socioeconomic threat or threat fundamental to our humanity (differentiated from a threat to our species, to be clear.)

In fact, at least from my perspective, most of those AI doomers are the “elite” that Kaczynski talks about. They’re looking to control it, have the academics that Kaczynski despises on their side, are good at bullshitting, etc., and a good way to control it and keep power in your hands is to paper over the real risks and perform regulatory capture.

That is to say, they want that system of technological progress to continue, but they want it to continue in a way that is controlled by them, centralization.

So, yes, but no. He absolutely despised those people. To clarify, I don't think you'll find many who want the whole system to stop, the most extreme you find are those who still want the system to continue, but want only “responsible” and “safe” companies to have control over such things, flip flopping between that and the possibility of air striking datacenters if some multi-modal model convinces them that it's sentient.

[thumbuddy]

In my mind the risk AGI poses is actually lower than that of current LLMs for the most part. Current use cases for LLMs are well aligned with the weaponization and exploitation of social and infrastructure flaws then they are for societal good.

People bringing up the hazards of AGI as a hyperbole for ignoring discussions of risk/merit of current AI technologies are acting disingenuously.

We really shouldn't distract ourselves from the matter at hand and divide people over the reality of the issue(s) that do currently exist.