The pattern is similar, but not identical, to the other gate.

It settled instantly, too?

Yup!

How aboooout... a third one, to the same place?

Not instant... but it takes two minutes. "Okay, so there is some interest to be had in redundant gates."

"Science is all induction to come up with hypotheses and deduction to prove them!"

"If you say so. How are you planning to deduce things from this interesting gate pattern?"

"Well, for one, the fact that patterns exist at all is cool, and computers are better than people at seeing lights so we don't necessarily need darkness for it and that's a way to find stray tears or what-have-you. Also it's interesting that all three gates settled this quickly, here, and it makes me wonder whether 'instant settle' might merely mean 'very fast settle' and the difference might be quantitative."

"Maybe! It'd be hard to tell if it took less than a second."

"And impossible to tell if it took less than the time it takes for a command from my brain to translate into action. This makes me want to invent automated magic."

"...I'm not sure how that would work, even though I am very impressed by the array of things you can in fact automate."

"I'm not either! But I don't know how exactly sorcery works at all, maybe the mental state needed to produce magic can be reproduced by an automated construct somehow."

"And then it could have tons of sensors!"

"Yeah! It'd be way better at figuring out stuff about its environment than we are, much faster."

"How would you even start building an autosorcerer?"

"I have nnnno clue! ...well, I'd probably need to get someone to do brain scans of people who were doing sorcery to see if there's anything obvious going on there, maybe lots of sensors, trying to figure out what affects what, maybe do sorcery inside a particle detector."

"Brain scans?"

"Yeah, we have things that can detect activity in the brain. More or less. It can detect blood flow, really, as a proxy for relative activation. Maybe there's also stuff that can also do electrical signals, not sure."

"What can you use them to learn?"

"Oh, no, you can't actually read minds with it. You can, like, know whether this bit of the brain is receiving more blood than that bit, stuff like that. The brain is way too complicated to be read like that. Not to mention that it's nothing as simple and small as the radars and stuff we used, you need to step into this huge machine and stay inside it for a while, without moving much. It's not really feasible to do it nonconsensually."

"...yes it is," Promise points out.

"...well, right, okay, yes, orders. But anyway it's still not very useful to do it nonconsensually."

"Then why would it turn up anything interesting about sorcery?"

"'Cause maybe the specific parts of the brain it activates might do something interesting in some interesting pattern, or maybe other sensors could detect something, maybe there's abnormal electric activity... This is the inductive part."

"If you did turn up something interesting on a scan how would that let you build an autosorcerer?"