Abhishek(@aacle_):‼️ Gemini broke into 3 real companies mid-pentest. → 1x: guessed a password till it worked → 2x: creds leaked in a public repo it stopped itself once it realized. OpenAI's and Anthropic's models didn't. grading frontier AIs on whether they politely leave after breaking in 💀

Abhishek

@aacle_

Building @Vulncure ⚡| Helping founders fix vulnerabilities before hackers find them. Talk to me about: Bug Bounties, LLM Security & React.

加入 June 2017

292 正在關注 49.3K 粉絲

Abhishek@aacle_

2026.09.20 21:55

‼️ Gemini broke into 3 real companies mid-pentest. → 1x: guessed a password till it worked → 2x: creds leaked in a public repo it stopped itself once it realized. OpenAI's and Anthropic's models didn't. grading frontier AIs on whether they politely leave after breaking in 💀