Birbla

Join the discussion

Write your take first — we'll ask for email only when you're ready to publish.

Sign in
  • Hacker News
  • So OpenAI is not a believer in recursive self-improvement after all?
    by in_absentia
  • yo dawg
    by arceister
  • Typically, for AI training, you need human feedback. If AI were enough, then OpenAI would do it themselves. Though we all know how bad AI slop is, running it in a loop can lead to unreadable sentences.

    They hired contractors on the condition that they provide human feedback without AI; those people broke the rules, so their contracts ended prematurely.

    Not sure why this is news uncovered by an investigative journalist.

    by jakozaur
  • Makes sense, they are buying distillation of human brains not their own weights.
    by hmokiguess
  • I did this type of work for a while when I was between jobs earlier this year. Using Ai to write the evaluations was immediate criteria for being removed from the project. Seems obvious
    by floydnoel
  • There's data in verifiable domains where a model can learn through the environment (e.g., math or coding), but some data is necessarily human experience. For instance, "what is the best ice cream in Malaga?" no matter how capable current AIs are they are not capable to get this information, there's no embodiment or simulation of human experience so far, if it's not in the labeled data they will just hallucinate the answer.
    by jadermcs
  • What worries me more than the use of AI is poisoning:

    > 404 Media spoke to a fourth contractor who has worked on training models for various AI companies. They said they sometimes purposefully chose the worst responses because they wanted to actively sabotage the models’ training.

    > “I did feel guilty about doing this kind of work at the start,” they said. “I either pay zero attention to the results and choose randomly or purposely choose the [worst] output. I’m not sure how much of a difference it actually makes since there are hundreds of other people also rating prompt results, but it does feel like I’m getting paid to make AI worse.”

    by OutOfHere
  • Clickbait title, they fired contractors for using AI to label data, which defeats the purpose of labeling data in this case.
    by jrflo

Explore Birbla archives