Based on my understanding of programming I think they're going to need an extra couple people on the security team because of the Ai's "fixes"

Because its doing so so well now unattended...

If an AI can be used for automatic scalable defense, it can also be used offensively. It'll just be another digital arms race between blackhats and everyone else.

Easterly said that if cybercrime was a country, it would be the third biggest in the world, just behind the US and China.

People who say these things clearly have no experience. I spent an hour today trying to get one of the better programming models to parse a response. I gave it the inputs and expected outputs and it COULD not derive functional code until I told it what the implementation needed to be. If it isn't cookie-cutter problems then it just can't predict it's way through it.

The one thing I will agree with is that If you ignore the AI part and just focus on the idea of having good software that can find code vulnerabilities, that’s a good idea.

At first I thought it was the Beaverton. Had to check the URL.

This is like a landlord painting everything white. It'll hide / seem to fix some issues, but probably isn't the fix you're looking for.

Good idea 👍

AI can't even write decent unit tests. How the hell is it going to properly red-team this service?

@cm0002 #nowplaying Absolutely Right - Five Man Electrical Band (Absolutely Right: The Best of Five Man Electrical Band)