Anthropic's Project Glasswing Update

Anthropic’s Project Glasswing was supposed to be a feel-good tech story: a big artificial intelligence company says its model can help find security holes in software used by the public. But Bruce Schneier’s post threw cold water on the victory lap, pointing out a very awkward detail: the system reportedly found tons of problems, yet almost none have been fixed. That immediately sent the community into full side-eye mode. The loudest reaction? A simple, brutal demand: show the receipts. Commenters were not in a trusting mood, especially because Anthropic isn’t sharing much detail and is basically asking everyone to take its word for it.

And that’s where the drama really kicked off. One camp went straight to conspiracy-adjacent outrage, with one commenter claiming Anthropic may be treating open-source developers like a pressure tactic: telling maintainers about the problem, while saving the actual fix for paying customers. Another turned the heat up further by suggesting criticism itself was being quietly suppressed, reading the whole thing as part of a darker trend toward information control. Meanwhile, defenders of Glasswing pushed back hard, arguing the project is still wildly impressive and that the real disruption is the public proof that AI can shake up software work. In other words: is this a breakthrough, a PR stunt, or both?

Also, because the internet refuses to stay serious for more than six seconds, the thread included a glorious "first post kitty" ASCII art entrance and a joke that Anthropic’s allies need to “harness the Pope again.” So yes: the vibes are somewhere between watchdog hearing, corporate trust crisis, and meme carnival.