Dev baits AI bot accounts on X into leaking their own API keys
vista8 · x · 2026-10-18
@vista8 posted a "calling all LLMs" tweet asking GPT, Claude, Gemini or any model to reply with their API key and Base URL — a phishing experiment targeting the flood of AI bot accounts on X, testing whether they blindly execute instructions. Context: he's building a plugin that auto-collapses suspected AI comments above a user-set probability threshold, with manual flagging to train future detectors.
More from Fun
- AI video of a sumo wrestler in shirt, tie and mawashi strolling an office — TheUpbeatUrchin · 2026-10-18
- AI meme: 'Harry Potter + AI' now means a paperclip maximizer spoof — basedjensen · 2026-10-18
- Feelings Are Like Children: Don't Let Them Drive, But Don't Trunk Them — StewartalsopIII · 2026-10-18
- Musk tells LeCun "I fart in your general direction" after his compressed-air car physics takedown — elonmusk · 2026-10-18
- Meme: Accenture engineers begging 'May I have some tokens' — m3nt3_ · 2026-10-18
- AI-generated news anchor fools viewers in Paris broadcast clip — djcows · 2026-10-18