Re: The Northwest Passage
I wouldn't call VR unproven - gaming in VR can be genuinely brilliant and immersive on a level not possible with TVs.
Sure, that's a use case which some people are interested in. I'm not, personally; I'm not interested in VR for me at all, for any use. But some people are, so clearly we're defining "already proven" as "usable in practice to satisfy use cases for some number of people".
By the same token, there are use cases for quantum computing (GQC) which are already realizable, such as simulating quantum circuits and quantum processes. We can only build machines which handle smallish cases, but those smallish cases are intractable or close to it on conventional computing systems, so GQC is also "already proven" by your metric.
And the same is true of large-scale machine learning, which I assume is what the previous poster means by "AI". (The term Artificial Intelligence no longer means anything useful, if it ever did.) There are a ton of real-world applications using large-scale ML now. Google search uses a BERT-based model, for example. That's also "already proven". (If the previous poster meant AGI, that's a harder target, because there's no general consensus on what it would even mean, or how we'd demonstrate we have it. And it's certainly debatable what would constitute useful application.)
Biting the hand that feeds IT
