argomo
Joined 7 years ago
Joined 7 years ago
-- Lord Kelvin, 1895
Are you saying that superintelligence is impossible?
Are you saying that the alignment problem will certainty be solved before superintelligence emerges?
Are you saying that a superintelligent being connected to the internet would be unable to gain resources such as GPU time, money, and social influence?
Are you saying that a superintelligent being would for some reason be incapable of deception and cunning?
Are you saying that a superintelligent being would necessarily regard human flourishing as a prime objective to be prized above it's own goals and ambitions?
If it's really just doomerism we should be able to point to the flaws in his argument instead of making ad hominem attacks.
If Anyone Builds It, Everyone Dies https://en.wikipedia.org/wiki/If_Anyone_Builds_It,_Everyone_...
One can easily imagine an LLM-enabled database that lets a wider audience build meat-and-potatoes line-of-business apps for small team use with minimal compliance concerns.
Putin did it better; he kept the military weak and aggressively managed the risk via the FSB.
I'm also finding that the AI-cited links are often more helpful and authoritative than the top search results.