It's like a giant shared illusion.

We know we don't learn from computer games. (Some very very few exceptions)

We know VR is just not good enough.

As developers or people who plan and pay for projects we know the money to develop this properly is prohibitively expensive and has not been spent. I'd guess $1 million to do it in a manner that works? More?

This is not like the dream of going to Mars where the shared delusion allows us to get GPS and Satellite internet and weather predictions and amazing things along the way.

This is just a roundabout cycle, someone pretends to make a game were people learn, it gets junked because it doesn't work, and repeat. We add words like VR, AR, AI to hide the cycle a little but it's not half obvious.

This isn't evolution or prototyping or MVP or experimenting, it's a known circle.