We're all going to die.
Ahead of OpenAI CEO Sam Altman's four days in exile, several staff researchers sent the board of directors a letter warning of a powerful artificial intelligence discovery that they said could threaten humanity, two people familiar with the matter told Reuters.
[...]
The maker of ChatGPT had made progress on Q* (pronounced Q-Star), which some internally believe could be a breakthrough in the startup's search for superintelligence, also known as artificial general intelligence (AGI), one of the people told Reuters. OpenAI defines AGI as AI systems that are smarter than humans.
Edit: The linked /pol/ thread is kinda insane
Edit 2:
They're now trying to cover it up
Jump in the discussion.
No email address required.
Jump in the discussion.
No email address required.
NLP models are terrible at math so while this isn't a skynet level breakthrough it's still nice progress
Jump in the discussion.
No email address required.
They're not terrible at math, the can regurgitate the answer to any problem ever written in a textbook on math. You really think that a problem that would involve completely novel mathematical reasoning would be classified as 'grade-school level'?
Also note that the emphasis is on 'acing' the tests. They're getting more precise at regurgitating answers.
Jump in the discussion.
No email address required.
lmao no. Sometimes they'll stumble into the right answer but more often than not it's all nonsense
Jump in the discussion.
No email address required.
More options
Context
NLP on math still sucks. Asking ChatGPT about insights on a table of data still gets you r-slurred answers.
Jump in the discussion.
No email address required.
More options
Context
Fair point.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
Q PREDICTED THIS
#WWG1WGA !schizos
Jump in the discussion.
No email address required.
This is basically what's going on.
Jump in the discussion.
No email address required.
More options
Context
This tbh. Muh breadcrumbs keklmao !slots100
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More details:
https://www.theverge.com/2023/11/22/23973354/a-recent-openai-breakthrough-on-the-path-to-agi-has-caused-a-stir
Jump in the discussion.
No email address required.
More options
Context
It might be that the AI is showing the capacity to solve maths and retain the knowledge in a way similar to how a child would, which would suggest that going further down the same path would result in the AI again thinking and comprehending on the same level as adult humans.
I am also going with nothingburger however.
Jump in the discussion.
No email address required.
It's a glorified auto complete. It doesn't think or comprehend.
Jump in the discussion.
No email address required.
The fact that many people view current AI as an all-capable machine that will escape if given the chance and poses an existential threat to democracy/world peace/whatever says more about the intelligence of the average persone than it does about the cabailities of current AI
Jump in the discussion.
No email address required.
More options
Context
You the same neighbor who 60 years ago would say a machine does not have the soul needed to beat a human at chess.
Jump in the discussion.
No email address required.
What? My comment simply says it cannot think, which is unarguably true.
Jump in the discussion.
No email address required.
alright my bad
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context
More options
Context