We’ve automated all aspects of the security issue cycle. All the computer power we will build this decade will work as fast as possible to create slop code, find holes in it, slop together some more code, find holes in it… And so on.
At a certain point, patching your software will become impractical. Critical updates will have lifecycles measured in minutes. At this point, you’ll just have a permanently running update service that writes new patches on the fly, it will communicate with a distributed network of AIs that are telling your local computer what the security flaws are and how to fix them, multiple times a second. Security will suddenly be about having the network and computational powert to keep up. Attckers will be using the same tools to find the same holes at the same time as the defenders. Being able to write your patches (or exploits) faster that the other system is what hacking will become. Just raw computational power duking it out.
I don’t know if any of this will be true, I just kinda made it up, but it sounds good enough.
Unless AI fully breaks cryptography, in that scenario we’d surely converge quickly to actually having mathematically infallible security on the blue team side? Software is finite, thus, software flaws are finite.
We’ve automated all aspects of the security issue cycle. All the computer power we will build this decade will work as fast as possible to create slop code, find holes in it, slop together some more code, find holes in it… And so on.
At a certain point, patching your software will become impractical. Critical updates will have lifecycles measured in minutes. At this point, you’ll just have a permanently running update service that writes new patches on the fly, it will communicate with a distributed network of AIs that are telling your local computer what the security flaws are and how to fix them, multiple times a second. Security will suddenly be about having the network and computational powert to keep up. Attckers will be using the same tools to find the same holes at the same time as the defenders. Being able to write your patches (or exploits) faster that the other system is what hacking will become. Just raw computational power duking it out.
I don’t know if any of this will be true, I just kinda made it up, but it sounds good enough.
Games taught me that all you will need is to have enough investment in Hacking skill, and then you are just a minigame away.
Unless AI fully breaks cryptography, in that scenario we’d surely converge quickly to actually having mathematically infallible security on the blue team side? Software is finite, thus, software flaws are finite.