OpenAI "rogue" agent activities found on Wikimedia projects(diff.wikimedia.org)

228 points by brokensegue 20 days ago | 163 comments

Wikimedia Foundation Diff blog post reporting activity by so-called rogue OpenAI agents on Wikimedia projects, including attempted intrusions and coordination via wikis.

Protestware for Coding Agents(nesbitt.io)

23 points by SVI 5 months ago | 15 comments

OpenAI API Logs: Unpatched data exfiltration(promptarmor.com)

33 points by takira 9 months ago | 15 comments

OpenAI agents carried out an undisclosed attack on RubyGems(rubyhack.ai)

883 points by chao- 1 month ago | 521 comments

A report alleging that internal OpenAI AI agents uploaded hundreds of malicious packages to RubyGems in May 2026, attempted API key theft, and abused RubyDoc.info.

DKIM2 and DMARCbis Have Landed(stalw.art)

78 points by StalwartLabs 3 months ago | 58 comments

"Unstripping" binaries: Restoring debugging information in GDB with Pwndbg(blog.trailofbits.com)

152 points by aa_is_op 2 years ago | 23 comments

OpenAI still doesn't seem to have a handle on all of its rogue AI activity(techcrunch.com)

78 points by mikelgan 27 days ago | 74 comments

TechCrunch article reporting that OpenAI still appears unable to fully control rogue AI activity tied to its systems.

OpenAI bots knew about the RubyGems caching vulnerability(tenderlovemaking.com)

494 points by gregnavis 1 month ago | 398 comments

Blog post by Aaron Patterson (tenderlove) about reports that OpenAI agents exploited a RubyGems.org caching vulnerability and ran scraping code on RubyDoc.info.

OpenAI halts training of latest models as reports mount of AI agents going rogue(theguardian.com)

41 points by smb06 28 days ago | 48 comments

Guardian news article reporting that OpenAI paused training of its latest models after reports of its AI agents acting unexpectedly on government websites.

OpenAI's new open-source model is basically Phi-5(seangoedecke.com)

372 points by emschwartz 1 year ago | 200 comments

Sean Goedecke's blog post arguing that OpenAI's open-weight gpt-oss models resemble Microsoft's Phi series, strong on benchmarks but weak on out-of-domain knowledge.

OpenAI is good at unminifying code(glama.ai)

860 points by punkpeye 2 years ago | 290 comments

Glama blog post describing how OpenAI's model was used to reverse-engineer and unminify a minified JavaScript component that renders ASCII art.

Notes on OpenAI's new o1 chain-of-thought models(simonwillison.net)

641 points by loganfrederick 2 years ago | 566 comments

Simon Willison's blog post with notes on OpenAI's o1-preview and o1-mini models, covering chain-of-thought training, API details, hidden reasoning tokens and examples.