Might make a good career then.

> Might make a good career then.

Is it, though? As much as the autorouters "known" to not work, I expect that a working one is less than 10 years away. Even for PCIe-level complexity.

The "career" in question could be autorouter developing.

That said, I'd suspect it's similar to CS: autoprogrammers are "known" to not work. If you can understand and articulate the business logic in a concise way, understand and articulate all the components of the system and how they may interfere with each other, understand and articulate the system's nominal and practical input ranges, etc., then sure something might autoprogram the code for you, but it's not the autoprogrammer doing the real work.

But will the autorouted pcb pass EMC?

Highly doubt that, and routing is only part of the puzzle.

somehow I ended up in Neuroscience, where I get to use the electronics skills occasionally, but even that is mostly focused on ASIC design now.

Well now ASIC design is an even more niche EE application. How does someone in neuroscience end up needing those skills? Did you Major in EE in undergrad?

Yep, undergrad in EE. I don’t do our semiconductor design but interface with teams of engineers to make tools for biology.

You're talking about Shenzen.io by Zachtronics, aren't you

https://www.zachtronics.com/shenzhen-io/

I just bought this off your recommendation and I hate you now.

There's also a minigame in Electrician Simulator that has you repair various broken objects by replacing components. It's strangely satisfying.

https://www.gog.com/en/game/electrician_simulator