I’m not an AI doomer, but my firsthand experience dating back to 2020 is that the researchers expressing concerns are sincere.
The closer they are to the research, the more worried they seem to be.
@BlancheMinerva@ilex_ulmus@sama How would putting one man in jail stop hundreds of thousands of ML researchers?
If the US tries to stop development they’ll simply leave to China
@TalulahRiley I’ve seen quite a few technologies develop, but none with this level of risk. AGI is significantly higher risk than nuclear weapons, in my opinion.
Super smart humans have trouble imagining something vastly smarter than themselves.
@reznik_andrew@jenzhuscott That’s why I believe these models could be so frightening, they lack any empathy we’ve evolved over millennia, it’s pure loss minimization.
Put alphafold 8 in the hands of a dr strange love type and the bubonic plague is akin to the flu
@reznik_andrew@jenzhuscott I think you’re still over fixated on the present, a world 5 years from now in which an architecture a few algorithmic advances above looped transfers, learns to shard itself and creates self replicating back ups is possible. Can you claim with absolute certainty it’s not?
@theo It’s fascinating to watch normalcy bias play out en masse. The tides have retreated, let us explore the sand, while those who run for the hills are mocked in disdain.
I believe history will prove him right.
@reznik_andrew@jenzhuscott We should be alarmed. Cross border alignment actions should certainly be taken.
I believe his angle for sensationalism is to rapidly raise awareness and cause a correction in rapid scaling.
But sanders style doomerism is too extreme. A middle path exists.
@DFintelligence M5 ultra is more than enough for deep seek 4.1
Now take the future version of Deepseek 6 and pair it with m7 ultras…
Your take looks at a point on a trajectory and claims this shit simply can’t hit the fan it’s 2 m too far !