Soviets landed on the moon the same way China did, several times, starting in 1970:
https://en.wikipedia.org/wiki/Lunokhod_programme. Their space program was one of the few things where they actually had the balls to innovate. Their space shuttle, Buran, for example, flew only once, but it was a fully autonomous flight, something the United States did not manage to pull off until X-37, 20 years later. And then Russia just pissed it all away, and they're still launching Soviet rockets because that's the only thing they know how to make. I'm sure they still have the blueprints to all the cool stuff though.
So even though the Chinese had all the public info (and stole what wasn't public), it still took them 43 years, billions of yuan, and several attempts to do what the USSR has done in 1970. As hard a feat as this is, putting people on there is so hard, not even the USSR was able to pull it off. And I'm pretty sure they knew pretty much everything there was to know at the high level, through spies and such, and had similar capabilities in most regards. What they did not have is what we don't have right now: the institutional knowledge. They "whys" and "why nots". The nitty-gritty. This is why we can't fly to the moon today - with all the people gone, we're only in a slightly better position than the USSR was back then. Perhaps even in a worse position, because back then it was fashionable to spend money on space programs, and now it's not.
There's a reason why the United States is the only country able to successfully put semi-autonomous landers on Mars. That reason is, someone kept the program alive instead of ticking a checkbox and letting the engineers go. Had they been let go, we'd lose this capability. Elon Musk wrote about this. In his view, technically complex capabilities aren't something you get to keep by default, and not something you can just start doing on a whim. You have to work hard and make stuff just to keep the institutional knowledge, and harder to push the state of the art forward.
There are dozens of such systems you interact with daily. Indeed, the very computer I type this on has no fewer than a dozen things in it that you wouldn't be able to make even if you had complete blueprints not just for the things themselves, but for the entire set of machinery required to make them, and for the machinery to make those machines, and so on.
The field I work in (deep learning / computer vision) is at times akin to black magic. You can give two people the same problem and one of them (that is, me
) will solve it much better than the other. It's not an exaggeration: that's literally how I feed my family right now. All the info is publicly available in papers on ArXiv. This is one of the most reproducible fields of computer science, by the way, even most of the code is publicly available under friendly licenses on GitHub, and you can reproduce the results on academic data, which you can also download.
But there are so many ways to skin the cat when it comes to
practical problems that you benefit quite heavily from the right set of intuitions, something Cliff has an abundance of after 10+ years of working at the leading edge of his field (as well as from his previous jobs).
Now, it is possible that they'll just come up with another set of intuitions. But the process takes quite a long time, and you get a different result. This is not something you can really copy. The only real benefit you get when doing this today is you know for a fact that things you're trying to figure out at a deep level are in fact possible. This is not something the trailblazers have in their mind. There's more trial and error.