nausicaa

[nausicaa @_nausicaa__]

@SenTinaSmith loss of control risk demands a lot more proactive solutions than jail people after the fact

[nausicaa @_nausicaa__]

@mollytaft yes definitely

[nausicaa @_nausicaa__]

@jon3k do you think LLMs, which have existed for less than a decade, are the most advanced form AI will take?

[nausicaa @_nausicaa__]

@Frances_Coppola source: washington post 🙄

[nausicaa @_nausicaa__]

@davidsirota how would that change their calculus if A) it is extremely hard to predict when they cross a threshold into “not safe” (potential catastrophe) B) trillions $ seem up for grabs and C) jailing after loss of control is too late anyway

[nausicaa @_nausicaa__]

@davidsirota david i love you but this is obviously not true. as long as there are incomprehensibly massive incentives to race toward superintelligence, firms will behave unsafely and jailing CEOs post catastrophe will be too late

[nausicaa @_nausicaa__]

@briebriejoy of course you do lol

[nausicaa @_nausicaa__]

@Lamp636269 @democracynow not either/or

[nausicaa @_nausicaa__]

@CNLiberalism correct, but that’s a bit of a non sequitur? she does not say anything to that effect here. am i missing something?

[nausicaa @_nausicaa__]

@Lamp636269 @democracynow crimes against humanity should be condemned and punished in all circumstances, and “not having western values” is a crazy way to characterize murdering hundreds of protestors

[nausicaa @_nausicaa__]

@r0berto_new_6 right… not good enough but at least she acknowledges the reality of loss of control risk at a moment when many on the left aren’t

[nausicaa @_nausicaa__]

@ChrisSosa the that she says “ we should take these risks seriously” is extremely important though

[nausicaa @_nausicaa__]

@bengrossbg i agree 100% but at least she acknowledges loss of control risks, which is more than i can say about many on the left, tragically.

[nausicaa @_nausicaa__]

@DavidSacks sounds easy

[nausicaa @_nausicaa__]

@NYCDSATechWG this is much better than the IG statement dismissing loss of control risk. “we can’t let the billionaires regulate themselves” is the right message!

[nausicaa @_nausicaa__]

@GaryMarcus @daniel_thilo israel not mentioned in headline lmao

[nausicaa @_nausicaa__]

@LuizaJarovsky likely is the operative word here! “extinction” or “by this date” framing is bullshit, the key issue is “we could lose control at some point in the future”, and that would be unacceptable regardless of when and how much harm AI subsequently does or does not inflict

[nausicaa @_nausicaa__]

@robertskmiles easier still is simply “loss of control” which would be unacceptable regardless of how much harm AI subsequently does or does not inflict on humans

[nausicaa @_nausicaa__]

@RepCasar @mehdirhasan @AOC wish she would ditch the “the REALER danger is ____” framing, but overall a good statement, all things considered…

[nausicaa @_nausicaa__]

@JeffBohren naive in the extreme to assume LLMs are the final stage of AI