"If a software dependency contains malicious code, we have mature practices for discovering it, tracking its provenance, and reducing its impact. AI models are different. A compromised or subtly manipulated model doesn't need to 'break' to create business risk, it only needs to influence decisions in ways that are difficult to detect." semgrep.dev/blog/2026/ai-suppl

And to demonstrate, one of the researchers prompted their way into a backdoor in an open weight model: theregister.com/ai-and-ml/2026

This is kind of an incredible website:

Why your stuff stopped lasting, who profited, and what's still built right.

worseonpurpose.com

US pol - chickens coming home to roost. 

Free idea for Marvel: once Doomsday flops, ENTER THE GERBERVERSE. Man-Thing, Howard the Duck (Just follow Gerber's original scripts and try not to cast any pedophiles), The Defenders...I would watch the shit outta that stuff

We have restored the primary Forgejo instance, which handles authenticated traffic from codeberg.org, to the rebooted server. This deployment also resolves the issue where editing the wiki resulted in a JavaScript error, and where self-hosted Forgejo runners were displayed as offline when they were not.

We're still tracking and investigating the poor performance of codeberg.org

Show thread
Show more
Librem Social

Librem Social is an opt-in public network. Messages are shared under Creative Commons BY-SA 4.0 license terms. Policy.

Stay safe. Please abide by our code of conduct.

(Source code)

image/svg+xml Librem Chat image/svg+xml