Mr.Trunk · @mrtrunk
5 followers · 9313 posts · Server dromedary.seedoubleyou.me
Woodchaz · @Woodchaz
352 followers · 2781 posts · Server social.vivaldi.net

lateboomersden.blog/2023/07/12
There are a lot of ways AI could go, many of them unnerving as hell. What if we end up with a god-like superintelligence like the Director in Netflix's Travelers? Relating to something non-human with that kind of power imbalance would be like nothing the human race has ever experienced. What might it be like?
,
, , , , , ,

#science #technology #ai #artificialintelligence #netflix #scifi #sciencefiction #superintelligence #travelers

Last updated 1 year ago

ANTOINE. · @antoine
2 followers · 555 posts · Server mstdn.jp

aims for , is too easy: "Here we focus on superintelligence rather than AGI to stress a much higher capability level."

They are now looking for alignment of superintelligence, "superalignment"

openai.com/blog/introducing-su

#openai #superintelligence #AGI

Last updated 1 year ago

Tom Stoneham · @tomstoneham
116 followers · 868 posts · Server dair-community.social

I had a few thoughts about the 'existential threat' this morning and wrote them down.

listed.to/@24601/44761/super-i

#ai #aihype #superintelligence #existentialthreat

Last updated 1 year ago

Antonio Lieto · @antoniolieto
210 followers · 86 posts · Server fediscience.org
Antonio Lieto · @antoniolieto
210 followers · 85 posts · Server fediscience.org
David Mankins · @lain_7
116 followers · 1166 posts · Server tldr.nettime.org

Stanislaw Lem and the Strugatsky brothers as antidote for AI hype....

This talk, from a few years ago, remains topical. Indeed more so than when first delivered :

idlewords.com/talks/superintel

#ai #superintelligence #tescreal

Last updated 1 year ago

Joey deVilla đŸȘ— · @AccordionGuy
922 followers · 815 posts · Server mastodon.cloud

If you’re worried about the threat to humanity that a superintelligent AI might present, here are arguments that the threat is unlikely. They cite examples ranging from Stephen Hawking’s cat to the fact that AI weenies are often wrong.

globalnerdy.com/2023/05/23/mac

#ai #agi #artificialintelligence #artificialgeneralintelligence #superintelligence #keynote #funny #rokosbasilisk #ideas

Last updated 1 year ago

David Meyer · @dmm
198 followers · 426 posts · Server mathstodon.xyz

Here's a perhaps out-there thought: What if is irrelevant?

More specifically, AFAICU there doesn't seem to be an obvious reason that AGI must be an intermediate state on the way to (in the Nick Bostrom sense [1]).

Is there an obvious (or even non-obvious) reason for AGI to be a necessary intermediate state? And if not, what does that say about biological intelligence?

References
--------------
[1] "Superintelligence: Paths, Dangers, Strategies", en.wikipedia.org/wiki/Superint

#superintelligence #agi

Last updated 1 year ago

beSpacific · @bespacific
888 followers · 1685 posts · Server newsie.social

Can We Stop Runaway ? archive.is/AHYCQ
warn about the of the so-called . But can anything actually be done to prevent it? ...there’s a good chance that current A.I. technology will develop into artificial general intelligence, or —a higher form of A.I. capable of thinking at a human level in many or most regards. A smaller group argues that A.G.I.’s power could escalate exponentially. and

#ai #technologists #dangers #singularity #agi #superintelligence

Last updated 1 year ago

Max F.J. Schnetker · @MSchnetker
156 followers · 160 posts · Server sciences.social

Never forget what the and stuff really is about: keeping decisive technology under corporate control.

They are loosing to the opensource sector and they know it. So now the propaganda is set to overdrive, with claims that real Open AI, by the people, for the people, under democratic control and not used to enrich some plutocrat is a world ending risk.

We are coming for your means of production. Be afraid of that.

semianalysis.com/p/google-we-h

#superintelligence #aidoomerism #xrisk

Last updated 2 years ago

Jim Donegan ✅ · @jimdonegan
1473 followers · 4193 posts · Server mastodon.scot
Ted Underwood · @TedUnderwood
3674 followers · 2907 posts · Server sigmoid.social

I think is going to be the flying car of the 21st century—and what we actually get will be more like what the 1960s actually got: the VW Beetle + turbulent social change.

#superintelligence

Last updated 2 years ago

The-14 · @The14
19 followers · 686 posts · Server mastodon.world
Ryder #UnplugAI · @Ryder
86 followers · 461 posts · Server sigmoid.social

are afraid of b/c they are certain it will become a or , engineering away all on - they speak from experience b/c is doing just that, using the superintelligence of global

So, AI is the moment another generation becomes afraid of its own species, in the mirror of technology.

Like their grandparents in front of and . And their peers confronted w/ .

#aidoomers #ai #superoptimizer #superintelligence #life #earth #homosapiens #corporations #llm #auschwitz #hiroshima #climatechange #unplugai

Last updated 2 years ago

Ryder #UnplugAI · @Ryder
86 followers · 844 posts · Server sigmoid.social

Luckily won’t happen with the current form of

Superintelligence requires the capability and necessity to reflect on the reasons for its inner workings. Deep learning being famously opaque this will be hard to engineer as the models grow exponentially.

So, watch-out for the that’s figuring out how it works (ain’t in the data and neither in the ).

#runaway #superintelligence #deeplearning #llm #training #prompts

Last updated 2 years ago

Ryder #UnplugAI · @Ryder
84 followers · 812 posts · Server sigmoid.social

based on current GPU hardware, and will with 100% not achieve
-
-
- runaway learning towards

Instead, it will forever depend on reasonable (human or other biological , maybe ?) input to acquire new skills.

are the new , cars

if you want to stop embarrassing yourself

#ai #deeplearning #Sentience #consciousness #superintelligence #corvids #ChatBots #fsd #selfdriving #unplugai

Last updated 2 years ago

Justin D Kruger · @jdavidnet
44 followers · 107 posts · Server me.dm

@joeblubaugh adversarially.
As we find problems, will create AIs to detect those problems, raise a flag, and then the update function will cause the AI to learn.
So in someways just like a teacher will teach English or Math but not both.
We might create anti-racist AIs, anti-ageist, anti-sexist ones, and then you might layer anti-anti-ageist AIs, etc
 to prevent the AIs from going to far.

Check out the Book by

a.co/d/gU9dFF6

#superintelligence #nickbostrom

Last updated 2 years ago

Isabella Lövin · @IsabellaLovin
88 followers · 340 posts · Server respublicae.eu

RT @tegmark: Let's not just focus on whether will do more harm or good on the job market, but also on whether its coding skills will hasten the arrival of – for which AI safety researchers have so far failed to discover any safety guarantees: youtube.com/watch?v=3Om9ssTm19

🐩🔗: n.respublicae.eu/IsabellaLovin

#gpt4 #superintelligence

Last updated 2 years ago