I’m waiting for the day when desperate LLM companies start paying people to post real human content, only for those people to just ask ChatGPT to do it.
Amazon mechanical turk, Outlier.ai, scale ai, are all gig platforms that pay people to produce real human content for LLM training.
At first they were paying people to actually produce content, like worked solutions to maths problems or translations, then they pivoted more to rating LLM output.
And as with every enshittification cycle, wages and work rapidly dried up, so people responded by asking LLMs for the answer just to meet deadlines and get by.
more than likely they are scraping other AI bot posts, there isnt that many users on most of the social media sites. its 50/50 AI/BOTS for most of them.
Big AI mostly switched to synthetic training data anyway. The books they’re digitizing are being used to gather knowledge, not writing styles or logic (mostly).
As in, when you ask ChatGPT how long some book is, it can just go check (if it’s in the database). It’s also useful if you ask about that book or about knowledge contained in that book. It’ll even reference books now (if you demand that in your prompt).
It’s not the same as earlier LLM tech which relied on scanned text to figure out how to respond to any given prompt (from a language standpoint). The “language” part of LLMs is a solved problem now (thanks to the synthetic training). At least for English 🤷
The problem is that synthetic data is not fit for that purpose. The more of it you use, the worse at dealing with the edge cases LLMs get.
Think of it like this… You feed a language model a bunch of genuine human-written content. Great. Now it can produce the most likely text in a lot of cases. Word combinations that rarely appear in written language rarely get generated, so most of its synthetic data lacks those rare - but still valid - combinations.
Train it on this synthetic data, and now more outliers and rare combinations get filed off. Rinse and repeat.
I’m waiting for the day when desperate LLM companies start paying people to post real human content, only for those people to just ask ChatGPT to do it.
Already happening for years.
Amazon mechanical turk, Outlier.ai, scale ai, are all gig platforms that pay people to produce real human content for LLM training.
At first they were paying people to actually produce content, like worked solutions to maths problems or translations, then they pivoted more to rating LLM output.
And as with every enshittification cycle, wages and work rapidly dried up, so people responded by asking LLMs for the answer just to meet deadlines and get by.
“now you only posted 4 times today Jenny, do you just want to do the minimum?”
This is already happening.
They are paying people bottom market rates to generate unique written content and people are just having AI write it for them.
Some kind of AI ouroboros situation.
(Non)Human centipede.
Reverse centaur centipede?
It’s simultaneously unethical as an employee and based to be feeding AI it’s own tail
more than likely they are scraping other AI bot posts, there isnt that many users on most of the social media sites. its 50/50 AI/BOTS for most of them.
Supposedly the people paid to rate responses are already using AI
Big AI mostly switched to synthetic training data anyway. The books they’re digitizing are being used to gather knowledge, not writing styles or logic (mostly).
As in, when you ask ChatGPT how long some book is, it can just go check (if it’s in the database). It’s also useful if you ask about that book or about knowledge contained in that book. It’ll even reference books now (if you demand that in your prompt).
It’s not the same as earlier LLM tech which relied on scanned text to figure out how to respond to any given prompt (from a language standpoint). The “language” part of LLMs is a solved problem now (thanks to the synthetic training). At least for English 🤷
The article quotes a post from ISBNdb saying the issue is model collapse from training on synthetic data.
Total collapse is a solution.
It’s also unavoidable.
Yes, what’s the issue?
That’s like saying, “they had some failure modes from the synthetic data, so they should just obviously stop trying forever.”
They’ll just fix the edge cases and move on. Like any programming task.
The problem is that synthetic data is not fit for that purpose. The more of it you use, the worse at dealing with the edge cases LLMs get.
Think of it like this… You feed a language model a bunch of genuine human-written content. Great. Now it can produce the most likely text in a lot of cases. Word combinations that rarely appear in written language rarely get generated, so most of its synthetic data lacks those rare - but still valid - combinations.
Train it on this synthetic data, and now more outliers and rare combinations get filed off. Rinse and repeat.
It’s like saving a lossy file over and over again. It degrades each time.