If I go back to school in any shape or form, I'd like to study ECE. Upon being exposed to all sorts of new hardware at my job, I'm increasingly interested in how these designs come to be.

It's sort of insane how quickly we go from "Wow!!!!!! AI can find vulns" to "Meh another headline about AI finding a gazillion vulns." Today's miracle is tomorrow's mundane, eh?

Instead of writing, "I found X bugs in Y codebase using AI" (cool! congrats!), please consider this:

(A) you propose some way to find the same bugs without burning a fortune on tokens, OR

(B) you propose some way to eliminate that class of bug by construction. With more and more people reliant on coding agents, where can we enforce the invariant such that the agent can't emit the bug to begin with? (new programming languages, frameworks, the model itself?)

So, an interesting thing about having a STEM degree is that it qualifies you to take the patent bar. Upon passing, you can practice as a patent agent, and maybe consider applying for law school if you care to become a patent attorney.

It seems interesting in small doses, and I can see myself doing it when I'm old and tired of research/eng. For now I quite enjoy writing about people's inventions and tracing their lineages and thinking about whether they'd actually work.

The funny part is that patent law is about arguing that an idea is new and who came up with it first, the opposite of what this name redacted government employee suggests we should want to want:

"The best thing in the world is when no one can remember whose idea it was. Then you know you've succeeded because the greatest thing that can happen is no one can remember who did it or how it was done; everybody has taken a piece of the idea, and it's been institutionalized."

This strikes me as superhumanly selfless, and I believe it. Monkey brain wants its name on something. At least monkey brain wants to want the selfless thing, too.