I think people did at least attempt to produce creative-play environments following that vision for a number of years. I'd personally credit a good part of my own education to them, so at least some of us even consider the effort partially successful!

Here's a short 1984 InfoWorld article surveying the booming state of science software at the time, including virtual dissection (Operation: Frog), principles of optics (Reflections), and even sensor accessories that could be plugged in for a sort of virtual science lab (the AtariLab series; later there was also the Science Toolkit series). I'd probably also add in: Logo, The Incredible Machine, Rocky's Boots, Lemonade Stand, Microbes, SimCity, Transport Tycoon, Balance of Power, etc., as at least partial successes, in various ways, and to varying degrees of game-ness.

I do often see claims that seem to discount all that, though. Am I an outlier in considering this era of software to be in large part a successful realization of the creative-play / educational gaming / constructivist learning vision? Or is the complaint just that momentum died by the mid-90s and we're now stagnant or even regressing?

Sometimes I think if we weren't so helplessly in love with video games' boundless potential we might make better video games. That's what I think.

I figure that, at the time, it would be impossible to accurately predict the consequences of videogame playing habits so far ahead into the future - even if Bushnell is often considered a visionary. The part of this 'videogames as creativity-inspiring entertainment' issue in which I'm interested the most is the ability to generate memories that go beyond the borders of real and virtual - to the point where some aspects of videogames are assimilated in our young minds, later becoming the mixed memories of real experiences. Because of that openness to learning, our young minds tend to be highly receptive to artificially induced digital contents – even if the organic is known to reject the artificial.

In my opinion, Bushnell was referring not so much to the category of "educational" games per se, but what can be learned from non-educational games instead. Not the learning process of school education or even pre-school education (with use of symbolic toys, for instance) or using the tools and methods of audiovisual tuition.

From a more comprehensive perspective, the answer remains unanswered: what are the true results of the interaction with virtual objects and realities - namely in this particular case where the interactions are made during the ages when the mind is most receptive - given our natural tendency/gift to fantasize endlessly and limitlessly? What did these simple games of our childhood instill on our innate ability to think and imagine?

I think that creativity and interaction are often wrongly conflated in contemporary mainstream gaming, and that a 'creative game design' does not necessarily mean 'creative game experience'.
In some defined sense, Tetris is still much more 'creative' than Fallout 3. Game designers put effort into the immersiveness of the experience, which is a commendable goal, but different from giving space to creative action. Often the very structure that produces immersive environments somewhat puts limits to creativeness. This is probably caused by market necessities.
Still, exceptions appear. Little Big Planet for example.
Videogames are a mature cultural product, and as such they should make extra efforts to take different shapes for different kind of experiences

In my opinion, Bushnell was referring not so much to the category of "educational" games per se, but what can be learned from non-educational games instead

Or even, perhaps, that what "educational games" means is "what can be learned from non-educational games."

Agreed.