We’ve been productively using AI for decades now – just not the AI you think of when you hear the term. Fuzzy logic, expert systems, basic automatic translation… Those are all things that were researched as artificial intelligence. We’ve been using neural nets (aka the current hotness) to recognize hand-written zip codes since the 90s.

Of course that’s an expert definition of artificial intelligence. You might expect something different. But saying that AI isn’t AI unless it’s sentient is like saying that space travel doesn’t count if it doesn’t go faster than light. It’d be cool if we had that but the steps we’re actually taking are significant.

Even if the current wave of AI is massively overhyped, as usual.

And Paul Fisher really just wanted to make a cool pen that can reliably write upside down. Congress and The Party agreed that the pen was cool and bought a couple hundred each.

I’d argue that accurate color perception isn’t necessary if one makes an assumption about the average age of the riders. Given that bright hair in humans is either blond or whitened by age (excepting albinos, which are rare), all of the riders having bright hair means that they’re either blond or old. Assuming that there are few large groups of senior riders, Legolas could come to his conclusion based on brightness alone.

Unfortunately I don’t know enough about optics to say whether this makes any difference.

I’d say that Microsoft’s Fluent design language is even worse. Material at least tries to use rounded shapes and animations; Fluent has been pretending that monocolored rectangles are interesting since 2010. And it has been consistently wrong.