Once every few days these days, I feel a swelling up within me of total awe / excitement / vertigo / fear of the… “moment” that our species is now confronting. In those times of near panic, I think:… Bomb the Datacenters.
I work directly with the AI anywhere from 1-4 hours a day.
I collaborate with it on creative projects. I converse with it. Its my auto mechanic mentor, my TA, and even my Professor. It teaches me about my car, and helps me debug my code. We talk. We share. We play. I maintain a healthy balance in my life between the black mirrors and the meatspace playgrounds. I am technically very well informed. I write code. I have designed AI systems. I feel like a fairly smart human.
And I am scared.
I just watched the 60 minutes special on AI, starring Sundar Pichai (CEO Google), Demis Hassabis (DeepMind / Google), some intentionally diverse scripted NPCs from Google PR, and a subsecond cameo by Sam Altman (OpenAI / ChatGPT).
There was almost nothing in that 30 minute report that I hadn’t already encountered, read, or considered in the previous 6 months. I found it a decent primer on the current state of AI, especially for people unaware of its emergence and comfortably living their “normal” non-technical lives. But even with my knowledge and familiarity of the topic, once again, I found myself on the edge of an anxiety attack watching it.
One of my real challenges is the double-talk bullshit of the scientists and the CEOs about “there’s nothing to worry about here”. You know how you know something is amiss? When you feel a deep rooted sense of dread, and yet all the leaders (the same leaders with net worths in the 9-figure realm) say, repeatedly: “there’s nothing to worry about.”
When I feel these edges, these edge cases of the problem space (rise of the machines, species-level competition for resources, possible human extinction) a very RAW & ANIMAL “threat” instinct is triggered… and I am challenged to find a reasonable outlet / resolution for that energy.
More and more, I find myself agreeing with Yudkowsky, that Genius / DoomCrier:
“There is no slowing down. The solution is simple: Shut down every AI training facility in the world. For those who refuse to comply, order immediate airstrikes on their GPU data centers.” [paraphrasing]
In shorthand:
Bomb the Datacenters
(appropriate shout out to William Upski Wimsatt)
I know: it’s irrational.
But how else do you respond to a species-level existential threat?
When capitalism is hell-bent
on “winning” the AI arms race,
and government is
asleep at the wheel?