This resonates with me, as I was just last night telling a friend about the grift that is “the singularity”. Remember that? The tech bros were really concerned about that for a hot moment, then their stock went up… …but nobody seems worried about the singularity anymore.
A lot of us still are. Just because the news cycle moved, doesn't mean Yudkowsky et al isn't still screaming into the void.
Two weeks ago: https://youtu.be/K8p8_VlFHUk?si=RpatWnlb1z8aKzhN
only a few days ago somebody posted a plot of how much of their code for an AI helper is written by AI and I think it reached 80% or something.
We're not quite at the point where deepseek comes up with the next version of deepseek, but as more intelligence gets distilled into smaller models, we can expect more and more people to start trying to build a better AI and eventually we should expect take off to happen.
To me it was just an interesting bit of speculation popularised by an academic and sci-fi author (Vernor Vinge) - which is the very thing such people should be doing imo.
And the history of it goes back to Ulam and Von Neumann, and maybe Turing. Hardly grifters.
Musk can't have it both ways, "AI is summoning the demon, and also I own a humanoid robot factory, AI for autonomous cars, and am training my own LLM now that a company I part funded then stopped supporting made a famous one before I did".
But there's still a lot of alignment research going on, to make sure we can even tell if the AI we have now has the equivalent of an evil bit (which they definitely observably and reproducible demonstrate having) that is set to "evil".