Maybe, just maybe #AI would not have to pretend to be "too dangerous to release" if software were actually secure.

[cigitalgem@sigmoid.social]

Maybe, just maybe #AI would not have to pretend to be "too dangerous to release" if software were actually secure. #swsec #appsec #MLsec

[studiolo_rb@mastodon.online]

@cigitalgem proud owner of a copy

[cigitalgem@sigmoid.social]

@studiolo_rb thank you. It's still relevant even though it is almost 20 years old.

[cigitalgem@sigmoid.social]

Got this angle into the NYTimes today

(unlocked article) https://www.nytimes.com/2026/05/12/technology/anthropic-claude-mythos.html?unlocked_article_code=1.h1A.aXHQ.35ofMK-FbSLN&smid=nytcore-android-share