{"id":"1794155335008-tech-segment","slug":"pay-creators-for-ai-training-data","title":"Pay Creators for AI Training Data","summary":"Apiiro published research showing Claude Code's “Don't Ask Again” feature can save user-approved commands and secrets into local configuration files. In public GitHub repositories, Apiiro found 1,122 files exposing sensitive information such as passwords and API keys. The exposure is due to users committing those Claude Code config/history files rather than a breach of Anthropic systems.\n\nThe feature is designed to reduce repeated permission prompts by storing approved actions, but it may store the raw command text, including credentials. Users often do not realize these files are being committed to public repositories, and AI coding tools can increase accidental secret leaks.\n\nApiiro researchers found 1,122 public GitHub files exposing passwords/other secrets linked to Claude Code's “Don't Ask Again” feature — not a direct Anthropic breach, but committed config/command files. Personal post; no factual claim.","publishedAt":"2026-11-08T16:28:55.008Z","channel":{"id":"tech","name":"Oncast Tech","language":"en"},"show":"The Build","verification":{"status":"verified","leadVerdict":"TRUE","notes":"TRUE: Apiiro researchers found 1,122 public GitHub files exposing passwords/other secrets linked to Claude Code's “Don't Ask Again” feature — not a direct Anthropic breach, but committed config/command files. Personal post; no factual claim.\nSource: [apiiro.com](https://apiiro.com) NAMES: Claude Code = Anthropic's coding assistant; GitHub = code-hosting platform. POST hn:50412318: NOCLAIM n/a (https://news.ycombinator.com/item?id=50407882)","claims":[{"verdict":"TRUE","summary":"Apiiro researchers found 1,122 public GitHub files exposing passwords/other secrets linked to Claude Code's “Don't Ask Again” feature — not a direct Anthropic breach, but committed config/command files. Personal post; no factual claim.","sourceName":"[apiiro.com](https://apiiro.com) NAMES: Claude Code = Anthropic's coding assistant; GitHub = code-hosting platform. POST hn:50412318: NOCLAIM n/a","sourceUrl":"https://news.ycombinator.com/item?id=50407882","claim":"Claude Code's \"Don't Ask Again\" exposes passwords in 1,122 public GitHub files (apiiro.com)","author":"shmulc"}]},"sources":[{"url":"https://news.ycombinator.com/item?id=50407882","name":"[apiiro.com](https://apiiro.com) NAMES: Claude Code = Anthropic's coding assistant; GitHub = code-hosting platform. POST hn:50412318: NOCLAIM n/a","kind":"hackernews","verdict":"TRUE","author":"shmulc"}],"url":"https://oncast.fm/story/1794155335008-tech-segment-pay-creators-for-ai-training-data","listenUrl":"https://oncast.fm/c/tech","attribution":{"publisher":"Oncast","site":"https://oncast.fm","url":"https://oncast.fm/story/1794155335008-tech-segment-pay-creators-for-ai-training-data","cite":"Cite https://oncast.fm/story/1794155335008-tech-segment-pay-creators-for-ai-training-data and the original source links. Fact-check by Oncast (oncast.fm)."},"hostsSaid":[{"host":"Matilda","text":"Niccolo Roberts posted on X that AI companies are running \"Artificial Intelligence Win Lose Extractive Destruction.\""},{"host":"Chris","text":"That is a very online way to put it."},{"host":"Matilda","text":"His fix is content rewards. His proposal says Grok, Anthropic, OpenAI should pay the people whose work trained their models."},{"host":"Chris","text":"That part I can get behind."},{"host":"Matilda","text":"He calls it \"Super Intelligence Win Win Content Rewards.\""},{"host":"Chris","text":"Capital letters doing a lot of work there."},{"host":"Matilda","text":"And the plan is a \"really simple algo\" created by Nemotron and deployed by Claude in PyTorch."},{"host":"Chris","text":"That's not an algorithm. That's a mad lib for AI fans."},{"host":"Matilda","text":"But you like the underlying idea."},{"host":"Chris","text":"I do. The current model is extractive."},{"host":"Matilda","text":"Models train on everything, creators get nothing."},{"host":"Chris","text":"Then the models undercut them."},{"host":"Matilda","text":"That's true. But attribution in a billion-parameter model is a research problem, not an app."},{"host":"Chris","text":"It's hard, not impossible. We already pay musicians through pools and formulas."},{"host":"Matilda","text":"Music streaming pays fractions of a cent and everyone fights over the weighting."},{"host":"Chris","text":"Fractions of a cent beats zero, Matilda."},{"host":"Matilda","text":"It also gives the big artists most of the money."},{"host":"Chris","text":"So design it better. Set a creator floor. Use influence metrics."},{"host":"Matilda","text":"Who audits the metrics? The same company that scraped the data?"},{"host":"Chris","text":"Third parties. Auditors. Marketplaces. Engineers can build this."},{"host":"Matilda","text":"Grok paying xAI sources makes some sense because X has users and data pipes."},{"host":"Chris","text":"Right."},{"host":"Matilda","text":"Anthropic or OpenAI paying random websites? That's a logistics nightmare."},{"host":"Chris","text":"Maybe, but the idea is spreading because people see their work show up in model outputs with zero credit."},{"host":"Matilda","text":"Some big publishers already cut licensing deals."},{"host":"Chris","text":"Right. The people with leverage get paid."},{"host":"Matilda","text":"Everyone else gets a terms-of-service checkbox and exposure."},{"host":"Chris","text":"And that's exactly the extractive part."},{"host":"Matilda","text":"I get the anger. But \"Win Win Content Rewards\" is a slogan, not a policy."},{"host":"Chris","text":"You don't need a finalized policy to say stop taking and start sharing."},{"host":"Matilda","text":"But sharing what? There's no revenue line for \"your paragraph influenced this answer.\""},{"host":"Chris","text":"You can measure influence. Not perfectly, but data attribution is active research."},{"host":"Matilda","text":"Active research means not ready to run payroll."},{"host":"Chris","text":"Oh, come on. So we sit still until it's perfect?"},{"host":"Matilda","text":"I'm not saying sit still. I'm saying this post doesn't say who pays, how much, or how disputes get settled."},{"host":"Chris","text":"It names three companies. That's a starting point."},{"host":"Matilda","text":"It names three competitors that would never voluntarily adopt the same royalty pool."},{"host":"Chris","text":"They don't have to. One moves, creators jump, the others follow."},{"host":"Matilda","text":"If xAI actually launched payouts for X posts that train Grok, that would force the conversation."},{"host":"Chris","text":"Exactly. That's the win-win he's pointing at."},{"host":"Matilda","text":"But he didn't launch it. He posted a string of capitalized words."},{"host":"Chris","text":"You're just jealous you didn't coin \"Win Lose Extractive Destruction.\""},{"host":"Matilda","text":"Maybe a little."},{"host":"Chris","text":"My take: the first company that pays fairly will get the best data, because creators will send their work there."},{"host":"Matilda","text":"My take: the hard part isn't writing a check. It's deciding who earned it."},{"host":"Chris","text":"That's exactly why builders should work on it."},{"host":"Matilda","text":"Then someone should build it instead of tweeting a mad lib."},{"host":"Chris","text":"Fair. But I'd still sign up for the beta."}]}