What I would have never have guessed. I’m shocked I say. /s
I’m pretty sure this sort of thing was tried with the prior era of neural nets too. When the field hits a ceiling they grasp at the make-AI-teach-itself straw. It’s the Hail Mary pass. What if we keep stacking AIs on top of each other. Maybe they’ll somehow break out of their own limitations.
There’s a cadence. Once in a while a breakthrough happens. The tech is incorporated into the world. There are variations of the tech, but all have the same fundamental ceiling.
The AI Effect takes place. People forget about AI for a while. Time passes. A breakthrough paper is published. AI is upon the world once again.
Only this time with LLMs, it’s seemingly passed the Turing Test so people think it’s close to the fictional AGI. Not just recognizing handwriting, speech, or images. Or putting an annoying animated character on your desktop. This time it’s being freakishly good at predicting what the next words should be based on known sum total of human knowledge. Making it be creative isn’t it this time. That’s the ceiling.
deleted by creator
not a reflection on the quality of OP’s submission, but man… like every day now I wish we had an active “noshitsherlock” sub for headlines like these
It’s not being said for the benefit of those who already know.
Someone should DLSS this meme
Turing Test is fundamentally based on fooling humans. Not sure it’s smart to pin humankind’s future on what a birthday magician can do.
found that AI agents could solve the engineering problems necessary to do AI research but lacked the judgment and creativity to produce original research at the caliber of papers accepted by a top machine-learning conference.
I mean that describes a majority of engineers. No small feat
If the standard of ML talks at conferences I’ve attended is anything to go by then a top machine learning conference is functionally a daycare for the most annoying people you’ve ever met
Tbh a lot of ai conference papers would absolutely be writable from most engineers. A lot of papers are : we trained a model and got these results. They didn’t publish the weight though, so you would be unable to check the paper.
The fingerprinting they’re doing now is meant to help with this. If they see a fingerprint, it was probably not human generated so don’t ingest it.
They are not only using human written data any more. Reinforcement learning through human feedback (RLHF) is a big part of it now, that’s the AI running through a problem multiple times with a human picking the best attempt and then using that best attempt as training data going forwards. The model collapse stuff from a few years ago was from an AI repeatedly ingesting its own input with no guidance over many training iterations.
Won’t come at all. It’s mathematically impossible.
Proof?
The output of a statistical model cannot contain more information than what it already had.
All that an AI has to do to disprove that claim is pay someone to weigh one orange through fivr. So either your claim is wrong, or LLMs as they exist now are no longer statistical models. Go ahead and move the goal post if you want, my complaint was about your overly broad claim.
I think their point is the LLM didn’t contain the information on the weight of that orange, it had to outsource it.
I picked the way in which an AI model (as the exist now) could most trivially produce a piece of novel (and deliberately trivial) information.
I wasn’t worried about whether the credit could fairly be attributed to it. We’re talking about whether they can produce information with which to grow their knowledge. Moral credit for the growth of their knowledge is outside the scope of the discussion.
So you think that AI increasing the capability of new AI just involves training a new model on the output of the old model? Or are you pretending to be a science?
deleted by creator
“That” is not what evolution did.
deleted by creator
I want to believe this but every biology major I talked to is like “that’s the middle-school explanation of evolution” and mention the actual process had more things going on.
It’s entirely possible to improve current AI using only information available to us right now. Once that well runs dry, current AI is in theory capable of running experiments and training on their results if we give it a harness to do that. This gives it access to new information. Could it succeed doing this? Unclear, but it is capable of trying. How do you think we discover AI improvements? Divine inspiration? No, we follow a relatively simple research loop.
you’re drowning in slop my man
You assume too much. I don’t even use AI, I just keep up with the research. I haven’t been wrong so far.
Everything is false until it’s true? There’s governments and corporations around the world right now racing to make that happen. What’s the point of the article? If it was so easy it would have been delivered already
A year ago they were similarly bad at writing code, often created unit tests that tested nothing, etc.
If the models are trained in what they’re doing wrong, that can accelerate their progress toward doing it right.
They would need to be trained for open ended creative tasks, which is just hard in the current reinforcement learning paradigm.
I think they’ll find infinite ways to fuck up. The guardrails will never be high enough, or strong enough.
They still don’t get it right all the time, they just stacked a few together to filter out the obviously wrong stuff.
