The big AI models are running out of training data (and it turns out most of the training data was produced by fools and the intentionally obtuse), so this might mark the end of rapid model advancement
Oh look, businesses didn’t plan for what to do after the low hanging fruit is gone. Shocker.
While synthetic data is a thing, you’ve really gotta wonder how often you can train a model on basically empty calories before the hallucination rate starts going up.
I, for one, hope the theftbots die.
I was reading an article about how ChatGPT will sometimes go on existential rants and I figure it’s probably because so much of the training data is now generated by LLMs and posted on the internet. probably a glut of people posting “I asked chatGPT what it was like to be a robot” and things of that nature.
No, it’s not. Maybe strictly for LLMs, but they were never the endpoint. They’re more like a Frontal Lobe emulator, the rest of the “brain” still needs to be built. Conceptually, Intelligence is largely about interactions between Context and Data. We have plenty of written Data. In order to create Intelligence from that Data we’ll need to expand the Context for that Data into other sensory systems; Which we are beginning to see in the combo LLM/Video/Audio models. Companies like Boston Dynamics are already working with and collecting Audio/Video/Kinesthetic Data in the Spatial Context. Eventually researchers are going to realize (if they haven’t already) that there’s massive amounts of untapped Data being unrecorded in virtual experiences. Though I’m sure some of the delivery/ remote driver companies are already contemplating how to record their Telepresence Data to refine their models. If capitalism doesn’t implode on itself before we reach that point, the future of gig work will probably be Virtual Turks where, via VR, you’ll step into the body of a robot when it’s faced with a difficult task, complete the task, and then that recorded experience will be used to train future models. It’s sad, because under socialism there’s an incredible potential for building a society where AI/Robots and humanity live in symbiosis akin to something like The Culture, but it’s just gonna be another cyber dystopia panopticon.
Intelligence is largely about interactions between Context and Data
me
intelligence
Eventually researchers are going to realize (if they haven’t already) that there’s massive amounts of untapped Data being unrecorded in virtual experiences.
They already have. A lot of robots are already training using simulated environments, and nvidia is developing frameworks to help accelerate this. Also this is how things like alpha go were trained, with self-play, and these reinforcement learning algorithms will probably be extended for LLMs.
Also like you said there’s a lot of still untapped data in audio / video and that’s starting to be incorporated into the models.
It’s been a few years since I’ve used mturk, but there were very few VR based jobs when I last used it. Has that changed?
Ah sorry, I was just being a smartass, no idea how much VR is on mturk now. To be clear I think you’ve got an accurately bleak view of the future of this stuff
I found a YouTube link in your comment. Here are links to the same video on alternative frontends that protect your privacy:
“our artificial intelligence has read every book in the world and is still dumb as shit”
This is accurate, though I am actually going to explain why. These big model companies (Google, ClosedAI, etc) parasitize the open-weights/open-source community that actually makes good Loras, fine tunes, and research papers. Consumer hardware simply hasn’t gotten good and cheap enough for very good fine tune training, and thats why this is all slowly petering out. In a couple of generations of consumer GPUs, which will be when we get consumer GPUs geared towards AI (re: super high VRAM counts of like 70gb+ for an affordable sub 700 usd cost), we might see another leap forward in this tech. Though I will say that this mostly pertains to LLMs, generative AI models like Stable Diffusion have a lot of tricks up their sleeves that can still be explored. Most of recent research and tweaking has been based around building a structure for the AI to build on, to sort of guide it rather than letting it take random stabs at things, in order to improve outputs. Some people have been doing things like hard coding color theory, framing a photograph, etc, and interpreting human language to trigger that hard code.
We’ve had statistical models like these since the 50s. Consumer hardware has always been the big materialist bottleneck, this is all powered by small research teams and hobbyist nerds. You can throw a ton of money at it and have a giant research team, but the performance you squeeze out of adding 400b more parameters to your 13b model or having a gigantic locked-down datacenter is going to be diminishing.
Also, synthetic data can be useful, people are hating on it in this thread but its a great way to reinforce good habits in the AI and interpret garbled code and speech that would otherwise confuse the AI. I sometimes feel like people just see something about ‘AI bad’ and upvote it and don’t try to understand it, where it is useful and where it is not, and so on.
I fail to see how synthetic data is good if it makes AI used to justify job cuts, “better”.
That’s where I’m at. Sure, there might be moderately-beneficial use-cases, maybe; but it doesn’t change the fact that there’s no such thing as an ethically-trained model, and there’s still no such thing as a model that wasn’t created based on rampant theft by capitalists, so I consider anything that comes of it fruit of the poison tree.
AI bad until the base that comprises it radically changes, across the board.
Sure, there might be moderately-beneficial use-cases, maybe; but it doesn’t change the fact that there’s no such thing as an ethically-trained model, and there’s still no such thing as a model that wasn’t created based on rampant theft by capitalists, so I consider anything that comes of it fruit of the poison tree.
I mean thats just the case with everything really. Theres a lot of very good use cases that are mostly to do with data manipulation, but the coolest ones are translating. I think we’re approaching a point where small models are providing very accurate translations and are even translating tone and intent properly, which is far superior to simple dictionary translation methods. I think its very possible that new phones could be outfitted with tensor cores and you could have a real-time universal translator in your hand, though it’ll likely only add ‘subtitles’ irl for you. AI voice-word recognition has also been very good and can be miniaturized. This is the use case I’m most excited for, personally, as a communist. Currently translating in a foreign country requires a lot of typing (if you dont have a perfect grasp of language) and it removes a very human element I feel to conversation. If everyone could locally run a subtitle-translation generation app it’d be amazing for all of humanity.
Theres of course plenty of manufacturing use cases as well, but China is spearheading on that, though there is some work being done in the US as well in the few industries that remain.
Synthetic data is basically a fancy way of saying ‘I’m properly formatting data and reinforcing the ai’s good outputs’. Rearranging words, fixing / adding tags, that sort of thing. This is generated with various tools that usually have an LLM or VLM plugged in, though some are as simple as a regex script.
Better hardware isn’t going to change anything except scale if the underlying approach stays the same. LLMs are not intelligent, they’re just guessing a bunch of words that are statistically most likely to satisfy the user’s request based on their training data. They don’t actually understand what they’re saying.
