Opus Copus

[Opus Copus @OpusCopus88]

@MikePFrank @willcb Bounded also by relevance to humans. What is a cat going to do with a rtx 3080? Not that I think this paradigm is capable of that sort of cross species jump in intelligence, but AI that starts doing crazy physical things that nobody can even conceptualize would be made illegal.

[Opus Copus @OpusCopus88]

@AbHomineDeus People won’t be happy in the hedonic treadmill but we are underestimating the amount of things to do in the future. Imagine someone from the 1600s trying to predict today— no one could do it. Not the aschenbrenners, not the Amodei’s, some things are computationally irreducible

[Opus Copus @OpusCopus88]

@deanwball Space is necessary. Every AI company should be talking about space and medicine. This is the only way

[Opus Copus @OpusCopus88]

@ReadDoofus @ZZ_SOUL_PLANIT @XiXiDu The idea that a super intelligence will have ‘super morality’ or something like this is dangerously misguided. It will be Dario Amodei’s personal morality + SOTA data analysis/machine learning to apply it as universally as possible

[Opus Copus @OpusCopus88]

@danfaggella @bitcloud The evil part is saying children aren’t conscious but club penguin chatbot is

[Opus Copus @OpusCopus88]

@danfaggella @bitcloud Evil retard

[Opus Copus @OpusCopus88]

@RuxandraTeslo David Hume vindicated

[Opus Copus @OpusCopus88]

@ThetaDecayPete @calder_mchugh @jam_croissant @CVecchioFX @JermalChandler Rationalism is what got us into this mess

[Opus Copus @OpusCopus88]

@FoldingNapkins @TheMeteorBlog Just joking around, of course computer science as a field itself is basically applied math

[Opus Copus @OpusCopus88]

@burner1693878 @Griffit15999943 @godspeed_aflame The people who say this invariably look like this

https://t.co/wlNtC7P3aS

[Opus Copus @OpusCopus88]

@Griffit15999943 @godspeed_aflame 100% This is what people like OP are setting themselves up for

[Opus Copus @OpusCopus88]

@spicey_lemonade No you wouldn’t need just that, you would need robots living embodied lives and becoming alcoholics like Bender from Futurama, getting addicted to gambling like Dostoevsky. Then I will happily accept AI art. Until that day it is blasphemy and garbage.

[Opus Copus @OpusCopus88]

@StephenLCasper Yes there are many ways in which this guy is much smarter in terms of dealing with the world’s complexity than Opus 5.5.

[Opus Copus @OpusCopus88]

@AndrewCurran_ Oneshotted by matrix multiplication. Many such cases

[Opus Copus @OpusCopus88]

@PlastiqSoldier @Michael_Druggan No. Your longtermism is satanic. In any situation an animals that turns their back on their own family for the good of another species is a perverse and demented creature. No different from a bonobo in captivity biting its limbs or filial cannibalism. Diseased and sick.

[Opus Copus @OpusCopus88]

@Rizstanford @davidsenra @hubermanlab @pmarca I mean yeah that’s basically exactly what he said

[Opus Copus @OpusCopus88]

@Griffit15999943 @godspeed_aflame 100% This is what people like OP are setting themselves up for

[Opus Copus @OpusCopus88]

@godspeed_aflame Evo-psych is the closest thing the humanities have to not being complete bullshit at this point OMEGALUL

[Opus Copus @OpusCopus88]

@realwesmyers Is infinitely scaling ai policy really more tractable than coordinating with other humans? Not saying I think we should pause, or hold hands with China, I’m actually very against that as we stand, but at the limit I’m somewhat skeptical that the former is easier than the latter.

[Opus Copus @OpusCopus88]

@willsolfiac https://t.co/lfJObfYYzB

[Opus Copus @OpusCopus88]

@nabeelqu For all talk about pacing the frontier, whether you think it’s good or bad, Dario is one of the few people in the world, where if you took him out of the last decade, we would possibly be multiple years away from current capabilities. And yet e/acc’s hate him. Actions > rhetoric

[Opus Copus @OpusCopus88]

@ArtemisConsort @PositivFuturist ‘Taste’, by the time it reaches the uvula it will be oneshotted by the second order internal model that the first internal model itself is safeguarding from lab employees