• 2 Posts
  • 122 Comments
Joined 2 years ago
cake
Cake day: June 19th, 2023

help-circle

  • Basically, model collapse happens when the training data no longer matches real-world data

    I’m more concerned about LLMs collaping the whole idea of “real-world”.

    I’m not a machine learning expert but I do get the basic concept of training a model and then evaluating its output against real data. But the whole thing rests on the idea that you have a model trained with relatively small samples of the real world and a big, clearly distinct “real world” to check the model’s performance.

    If LLMs have already ingested basically the entire information in the “real world” and their output is so pervasive that you can’t easily tell what’s true and what’s AI-generated slop “how do we train our models now” is not my main concern.

    As an example, take the judges who found made-up cases because lawyers used a LLM. What happens if made-up cases are referenced in several other places, including some legal textbooks used in Law Schools? Don’t they become part of the “real world”?


  • I tried reading the paper. There is a free preprint version on arxiv. This page (from the article linked by OP) also links the code they used and the data they tried compressing, in the end.

    While most of the theory is above my head, the basic intuition is that compression improves if you have some level of “understanding” or higher-level context of the data you are compressing. And LLMs are generally better at doing that than numeric algorithms.

    As an example if you recognize a sequence of letters as the first chapter of the book Moby-Dick you’ll probably transmit that information more efficiently than a compression algorithm. “The first chapter of Moby-Dick”; there … I just did it.














  • The thing is that social media have an oversized influence that makes a calm discussion of possible solutions very hard to have. When the US recognized the implications of letting a foreign power exert so much control over their people, they tried banning TikTok, or breaking it up so their US operation would be under US control.

    Facebook should also be split and its EU operation purchased by a European company, that could then spend more time implementing the other changes you mention (doom-scrolling, data protection) and less time lobbying to get all these pesky EU regulations removed.

    And yes, it does feel heartbreaking to count the US as a threat to national security, but China has never threatened to annex Greenland with military force, so what would have been paranoia and extreme anti-americanism last year is now the sensible, level-headed thing to do.