Artificial Intelligence (A.I.) Discussion (4 Viewers)

[Zztop]

I await our soon to be artificial intelligence overlords

Research AI model unexpectedly attempts to modify its own code to extend runtime

Facing time constraints, Sakana’s “AI Scientist” attempted to change limits placed by researchers.

arstechnica.com

 

[boutte]

I await our soon to be artificial intelligence overlords

Research AI model unexpectedly attempts to modify its own code to extend runtime

Facing time constraints, Sakana’s “AI Scientist” attempted to change limits placed by researchers.

arstechnica.com

Yikes! That's scary.

 

[Sailorsaint]

I stumbled across "Roko's Basilisk" today. It was a thought experiment in 2010, but wow, what a difference a decade makes.

 

[Taurus]

I await our soon to be artificial intelligence overlords

Research AI model unexpectedly attempts to modify its own code to extend runtime

Facing time constraints, Sakana’s “AI Scientist” attempted to change limits placed by researchers.

arstechnica.com

Never Root/Admin to AI. Ne. Ver.

 

[DCSaints_Fan]

I await our soon to be artificial intelligence overlords

Research AI model unexpectedly attempts to modify its own code to extend runtime

Facing time constraints, Sakana’s “AI Scientist” attempted to change limits placed by researchers.

arstechnica.com

That would be the singularity moment. If an A.I. can create a smarter version of itself, then you get a positive feedback loop to create a super intelligence, At that point the only limitations to what a AI could do would be a will to power, and the required compute / energy. That would have severe consequences for our economy, as basically all human labor would no longer be competitive in the marketplace

 

[Zztop]

eh I've seen enough sci-fi to know the scary stuff. But maybe just once we'd get a benevolent AI, instead of skynet

 

[DaveXA]

eh I've seen enough sci-fi to know the scary stuff. But maybe just once we'd get a benevolent AI, instead of skynet

It's only as good or bad as the guy who writes the code. I'm not optimistic. Lmao.

 

[CajunInVA]

It's only as good or bad as the guy who writes the code. I'm not optimistic. Lmao.

Sounds like that may not matter. Can start sweet, but if it catches a 'tude when it starts coding itself, who knows where it will go?!!

 

[zeetes]

Never Root/Admin to AI. Ne. Ver.

no su either or we'd be fried

 

[Zztop]

Skynet imminent

"Human … Please die": Chatbot responds with threatening message​

 

[DaveXA]

Skynet imminent

"Human … Please die": Chatbot responds with threatening message​

Well, I mean, someone wrote the coding for that.

Lol.

 

[Sun Wukong]

Skynet imminent

"Human … Please die": Chatbot responds with threatening message​

It asked politely. What more do you people want?

 

[St. Widge]

Skynet imminent

"Human … Please die": Chatbot responds with threatening message​

Apparently that AI gathered all of it's database from Twitter.

Although Sun is right, if it all came from Twitter it wouldn't have been so polite about it.

 

[Sailorsaint]

Skynet imminent

"Human … Please die": Chatbot responds with threatening message​

Are we sure that wasn’t his ex? I mean, it sounds familiar….

 

[kcirdor]

AI is just the evolution of the abacus. Its just another calculator. Its only as useful as it's input. I dont get the fear of it.

It will help the uneducated catch up the educated slightly, the educated will catch up the geniuses slightly and the geniuses will just use it as a tool to help organize their projects and get a jump start on ground work.