Paul Shearer

[roon @tszzl]

we should not like a future where there are 1 to 2 AIs and the rest are illegal. we should want a cambrian explosion of the types and possibilities of minds, experiencing new joys and thinking new thoughts. ensuring that this is compatible with human flourishing is only step 1

[Paul Shearer @ThePaulShearer]

@robbensinger My most recent Substack essay covers all of this in more meaningful detail. https://t.co/rc0F6TvEMr

[Paul Shearer @ThePaulShearer]

@robbensinger Furthermore, even if you show that AI increases net danger from either cyber or bio, it does not follow that halting AI development is helpful. The effect of that could be simply to remove the edge that advanced, civilized society has over criminals. end/

[Paul Shearer @ThePaulShearer]

@robbensinger I don't think that fast hackers matter. I don't think pathogen engineering is realistic, and by the time it is, it won't matter.

[Paul Shearer @ThePaulShearer]

@robbensinger My position is that I would like to know what agents we can produce, and millions of words worth of LessWrong post verbiage notwithstanding, I see no reason whatsoever to believe that exploring that space is dangerous to humanity.

[Paul Shearer @ThePaulShearer]

@robbensinger I think the AI safety movement has been completely cavalier, thoughtless, and fantastical as to what "vastly more capable" would even mean.

Computers can do multiplication billions of times faster than us. Vastly superhuman. So what?

[Paul Shearer @ThePaulShearer]

Indeed, and one of the implications is that you don't actually know if it really would be incredibly powerful. Especially in an environment controlled by humans.

[Aella @Aella_Girl]

people have really not fully grappled with the implications of a superintelligence. it is very difficult to imagine something so much smarter than you.

[Ramez Naam @ramez]

Amen. AI is the most important new technology humanity has developed, perhaps ever. The last possible thing we should want is centralized control over it. Democratization and diversification of this technology is the path to resilience and freedom.

[Paul Shearer @ThePaulShearer]

@Nixaveee @DONOTACCEPTai I refuse your "for my own good" prison state.

[Ramez Naam @ramez]

100% agree. AI is hyper-competitive and default plural. No signs of natural monopolies or strong network effects. Most likely path to AI monopoly is through regulatory capture or poorly crafted policy. Much more here: https://t.co/WjJHxjHUS7

[Paul Shearer @ThePaulShearer]

@labenz @nlw You still need a human saying what they want and don't want, and speaking as a deep applied AI practitioner in industry, that question is way, way harder than you are likely to ever imagine

[Paul Shearer @ThePaulShearer]

@allTheYud They're machines and they love us, don't feel too bad

[Paul Shearer @ThePaulShearer]

@allTheYud @fchollet I'm sure my "simply" here will sound outrageous to you. But I do in fact insist on it, and I have built a whole AI risk worldview around it.

[Paul Shearer @ThePaulShearer]

@allTheYud @fchollet Humans get to use machines, so the edge in question is narrower than you make it sound. Still, I would guess that machines have a more favorable upskill curve than humans once they have the right sensors and actuators (many years off). The solution is simply to limit their power.

[Ramez Naam @ramez]

Despite this election, I remain an optimist about America and the world. Humanity will continue to produce new ideas and new innovations to improve our lives. Good people will continue to come together to improve the world. And the political tide will turn. We'll make it so.

[Paul Shearer @ThePaulShearer]

@chriswaterguy @allTheYud Every AI doom argument is answered in one of my essays

[Paul Shearer @ThePaulShearer]

@Baes_Theorem @JeffLadish Throttles civilization too much and becomes impractical as capabilities advance and become cheap. Instead, allow deployment and use liability plus regulation for TBTF areas. There's more to it, my essay AI as Normalizable Technology covers the rest. https://t.co/3WQptp7PAe