Hacker News Reader

Others have pointed out that we thought we would get security vulnerability from AI code, but it turns out AI is uncovering vulnerabilities in decades of human code.

View on HN

They aren't mutually exclusive. Wait until we find out intelligence services have been influencing the output of AI generated code for certain people/counties.

Already did. DeepSeek generates insecure code if you mention Tibet and underhanded code if you mention US DoD.

https://www.crowdstrike.com/en-us/blog/crowdstrike-researche...

Which it was trained on

Is the point that you're making that it only found the vulnerabilities because they were in the training set, and therefore AI didn't really "find" them?

If so, I fear you're making the opposite point, which is that AI made genuinely new discoveries that were not in its training set, about codebases that were in its training set. It quite literally expanded the horizon of what we know.

Neither, I'm trying to assert that AI generated code would follow the (flawed) pattern we encoded in it. I think they're great for cyber stuff

That assumes AI isn't used to review AI code. As it turns out that AI is great at finding vulnerabilities.

If it can find flaws in human generated code why not find flaws in self generated code as well?

Reply on news.ycombinator.com