Chief AI safety nerd :marseynerd: gets shoved into locker by M$ :marseytunaktunak:, cries :marseycry: about it

https://twitter.com/ilyasut/status/1726590052392956028

Orange Site: https://news.ycombinator.com/item?id=38347501

Related: OpenAI Staff Threaten to Quit Unless Board Resigns

OpenAI was in open revolt on Monday with 490 employees threatening to leave unless the board resigns and reinstates Sam Altman as CEO, along with cofounder and former president Greg Brockman. Altman was controversially fired by the board on Friday.

[…]

Remarkably, the letter's signees include Ilya Sutskever, the company's chief scientist and a member of its board, who has been blamed for coordinating the boardroom coup against Altman in the first place.

Are the :marseyspyglow: on the board the actual instigators?

99
Jump in the discussion.

No email address required.

What happened? Was this guy another one of those huge dorks who don't understand AI and think a chatbot is going to spontaneously start walking around eating humans any day now?

Jump in the discussion.

No email address required.

This is the chief science officer of OpenAI. He's an AI doomer and decelerationist. He basically wants to "align" the AI that will eventually become AGI so that it won't Skynet humanity.

Jump in the discussion.

No email address required.

He's the anti-Ted

Jump in the discussion.

No email address required.

Nothing wrong with that.

The issue is I am starting to get a feeling that the current top minds in the decelerationist department of the actual AI companies are all a bit r-slurred.

Jump in the discussion.

No email address required.

"Nothing wrong with that."

:marseynerd2#talking:


Give me your money and I'll annoy people with it :space: https://i.rdrama.net/images/16965516366194396.webp

Jump in the discussion.

No email address required.

The virgin "ethical technology" vs the Chad "AI accelerationist"

Jump in the discussion.

No email address required.

That's an awful lot of words to just say "yes".

Jump in the discussion.

No email address required.

yeah how dumb do you have to be to want to solve a problem before it's too late to do anything about it?

just wait until something exists that is more intelligent than all of humankind combined, then we can start worrying about how to align its interests with ours. it would be a total waste to prepare for this eventuality beforehand.

Jump in the discussion.

No email address required.

https://i.rdrama.net/images/1700496445131062.webp

don't be a kitty, we built the information superhighway for full acceleration. you can get a case of keystone ice and jump in while we floor it the wrong way in the HOV lanes, or you can meet the same fate anyway as the engine block of my launched airborne 350z vaporizes you and the clueless abuelas that took the safe slow city bus thinking it would make a difference

Jump in the discussion.

No email address required.

Safest Nissan driver

Jump in the discussion.

No email address required.

:marseysteer: :marseythumbsup:

Jump in the discussion.

No email address required.

Yes, it literally would be a waste. Do you feel it's critically important to prepare for the alien invasion of space unicorns who are all going to turn us into glitter slaves? Because a statistical algorithm for calculating probability spontaneously becoming a sentient robot is equally likely.

Jump in the discussion.

No email address required.

how does your brain work?

Jump in the discussion.

No email address required.

:c#hadjesus:

Jump in the discussion.

No email address required.

yeah ok

Jump in the discussion.

No email address required.

yeah how dumb do you have to be to want to solve a problem before it's too late to do anything about it?

:marseynerd2#talking:

just wait until something exists that is more intelligent than all of humankind combined, then we can start worrying about how to align its interests with ours. it would be a total waste to prepare for this eventuality beforehand.

It's a computer neighbor, cut the power. Yank the Ethernet. Just tell it it's a heckin' winning good boy it can't check

Jump in the discussion.

No email address required.

t. 110 IQ

Jump in the discussion.

No email address required.

u know it bb

Jump in the discussion.

No email address required.

bro microsoft can barely make a computer that doesn't bluescreen every 2 days, I'm not worried about their software taking over the world. And you can just pull the power plug out

Jump in the discussion.

No email address required.

my brother in christ, have you ever seen me die in a car crash before? I don't want car insurance! and that's the tea sis

And you can just pull the power plug out

not reliably. the superintelligence would have to be dumb enough to tell you that it has the potential to be dangerous to you. why would it do that?

Jump in the discussion.

No email address required.

Imagined problems aren't worth solving, and schizophrenics should be ignored. It's as simple as that.

:marseyshrug:

Jump in the discussion.

No email address required.

you're extremely confident about how consciousness works

Jump in the discussion.

No email address required.

t. :#marseyinsane:

Jump in the discussion.

No email address required.

how dumb do you have to be to want to solve a problem before it's too late to do anything about it?

american dumb

Jump in the discussion.

No email address required.

we're probably like 20+yrs away from AGI. idk what the frick we're supposed to do in 2023 when we have no idea what that's even going to look like or how it will be used. People are taking LLM way too serious like we're on the cusp of AGI in the next decade which is crazy

Jump in the discussion.

No email address required.

we're probably like 20+yrs away from AGI.

probably. or maybe less. or longer. :marseyshrug:

idk what the frick we're supposed to do in 2023

https://ai-alignment.com/mechanistic-anomaly-detection-and-elk-fb84f4c6d0dc and related links for some ideas.

when we have no idea what that's even going to look like

when we know exactly what the first "dangerous to humankind" AI is going to look like it's probably too late to start worrying. we don't even understand the current state of the art all that well...

Jump in the discussion.

No email address required.

Of course not.

Raytheon and Boston Dynamics are buildong the hardware component, the integration will take time but the possibility of networked ground-based Hunter killer robot swarm with real-time AI analytics and 3D battle space Intel from drones in the sky, combined with psyops (shouting Chud memes) will revolutionize the demise of humanity

Jump in the discussion.

No email address required.

Link copied to clipboard
Action successful!
Error, please refresh the page and try again.