[godelski]

As others have mentioned about different levels of research having different levels of secrecy let me tell you about some of the research I've done.

I work in HPC and people are getting into machine learning now. Lots of papers are getting published that do not contain enough information to reproduce the work and source code is held tight. That's not uncommon is a lot of the sciences. Results are public, but how to get the results are not exactly. I've also been in the engineering and physics fields. Papers there are really not reproducible. They may detail a high level overview of how to do the experiment but they leave out all the "secret sauce". This gives them an edge.

So with the other post that made the front page, just the other day, it detailed how someone got stopped at the border for lying on the visa and that that person was sending pictures and very low level detailed explanations about the lab and its setup. This is essentially the "secret sauce" we're talking about here. If these researchers want to turn their results into a company (say you invent a new drug) you can only do this if you don't tell everyone every detail (which you can't reasonably do in a 10 page journal paper).

This all is especially true for medical work and semiconductor work. The US government is funding many of these labs because they actually want them to do the low level research, make a product, and market it. You can think of it as angel investing (except you don't have to give up shares of the company). But they want it to be a US company that builds and makes the product because that helps the US economy. If the US does the "angel investing" and then the product is created in China by a Chinese company, then that investing was not nearly as useful.