signalThe Rundown AI2026-09-22
OpenAI goes from hacker to hacked
Hacktron AI's three-person team breached OpenAI's private codebase in under 72 hours this July via an image-upload flaw in the community forum, then exploited a second bug to let staff sign-in tokens unlock ChatGPT accounts, with Anthropic's Claude Opus 5 (finishing work Opus 4.8 couldn't complete) scripting the attack; they left a signed proof-of-concept edit and earned a $6,500 bounty. The same image software vulnerability also enabled breaches of Slack, Meta, and GitHub Enterprise, with only one target detecting the intrusion mid-attempt.
- topic
- AI Industry Impact
- source
- The Rundown AI
#Claude
score
score 6 out of 10. 0-10: how dense the facts are, multiplied by how much you can do with them after reading. 8+ means the topic's evidence bar is met: benchmarks and availability for a new model, amount and investors for a funding round, revenue figures for a solo-money story. Below 5 an item does not enter the digest. A press release scores 3 or less, a reprint loses 2, anything older than 14 days loses 1, a headline that misleads loses 3.
read the source