simple@lemm.ee to Technology@lemmy.worldEnglish · 12 days agoFOSS infrastructure is under attack by AI companiesthelibre.newsexternal-linkmessage-square40fedilinkarrow-up1369arrow-down12 cross-posted to: technology@beehaw.orgtechnology@lemmy.worldopensource@lemmy.mltechnology@lemmy.world
arrow-up1367arrow-down1external-linkFOSS infrastructure is under attack by AI companiesthelibre.newssimple@lemm.ee to Technology@lemmy.worldEnglish · 12 days agomessage-square40fedilink cross-posted to: technology@beehaw.orgtechnology@lemmy.worldopensource@lemmy.mltechnology@lemmy.world
minus-squareHubertManne@piefed.sociallinkfedilinkEnglisharrow-up21·12 days agoAny idea what the point of these are then? Sounds like its reporting a fake bug.
minus-squarewjs018@piefed.sociallinkfedilinkEnglisharrow-up54arrow-down1·12 days agoThe theory that the lead maintainer had (he is an actual software developer, I just dabble), is that it might be a type of reinforcement learning: Get your LLM to create what it thinks are valid bug reports/issues Monitor the outcome of those issues (closed immediately, discussion, eventual pull request) Use those outcomes to assign how “good” or “bad” that generated issue was Use that scoring as a way to feed back into the model to influence it to create more “good” issues If this is what’s happening, then it’s essentially offloading your LLM’s reinforcement learning scoring to open source maintainers.
minus-squareHubertManne@piefed.sociallinkfedilinkEnglisharrow-up28arrow-down1·12 days agoThats wild. I don’t have much hope for llm’s if things like this is how they are doing things and I would not be surprised given how well they don’t work. Too much quantity over quality in training.
Any idea what the point of these are then? Sounds like its reporting a fake bug.
The theory that the lead maintainer had (he is an actual software developer, I just dabble), is that it might be a type of reinforcement learning:
If this is what’s happening, then it’s essentially offloading your LLM’s reinforcement learning scoring to open source maintainers.
Thats wild. I don’t have much hope for llm’s if things like this is how they are doing things and I would not be surprised given how well they don’t work. Too much quantity over quality in training.