One thing to think about.
Due to current anti-LLM policy of @Codeberg, I can't consider projects hosted there as secure as other projects anymore.
Yes, I saw the new rules and understand that they clarify that "old projects with pre-LLM history are most likely unaffected" and so on. But:
- Some developers would just use self-censorship just in case to avoid being banned. They may use LLM's to find vulnerabilities, but they'd be required to introduce fixes manually even if generated fixes are perfect and show no problems in human review. Unless they lie and hide it, of course.
- Other developers are just ideologically opposed to LLM's and would avoid using them even if it harms their users. Chance of encountering such developer on Codeberg is much higher.
Both of those gives more time to attackers until we're getting updates with vulnerabilities fixed.
This prompts further thoughts about my usage of software like @gotosocial and some other tools that moved to Codeberg before those rules were introduced.
#thoughts #software #security #CVEapocalypse #LLM #infosec #Codeberg #policies #politics #AntiAI #people