Progress 1900-1941. Mostly based on practical 'tinkering' with a little science and engineering at just the college level. With a few exceptions, didn't need research contributions. Some Exceptions: Antibiotics, anesthetics, chemistry and chemical engineering. Also Bell Labs had started working on transistors based in part part on understanding of quantum mechanics. WWII interrupted the work, but soon after the war Bell Labs had the transistor.
Progress 1941-1960. Heavily driven by big spending, lots of drive, for US national security, especially WWII and the Cold War. E.g., US DoD spending helped create Silicon Valley.
Progress 1960-1974. Throttled by US efforts in Viet Nam, the oil shock, and inflation from Viet Nam spending.
Progress 1974-present. Moore's law, the Internet, some pharmaceuticals, some in materials.
Otherwise progress was throttled. And here I will have to attempt to describe a cause, my idea, apparently not mentioned so far.
There's plenty of good pure and applied research on the shelves of the research libraries. There are plenty of well trained people in the results of the pure and applied research and how to do more such research, in science, engineering, technology, etc.
But for this material on the shelves of the libraries and these trained workers to contribute to the US economy and standard of living as Thiel discusses, we need a lot of other people to go along, approve, fund, act as CEOs, Members of Boards of Directors, etc. Then, these other people upchuck at what the well trained people know, feel threatened, inadequate, intimidated. They fear that going along with the corresponding projects would be imprudent for them as managers and/or financially irresponsible.
Or, through 1940, work was mostly in a management hierarchy where the supervisor knew more and the subordinates were there to add muscle and/or person-hours to the work of the supervisor. Now that situation has changed: Mostly the people who really understand the technical material and how to apply it are just the worker bees, and the rest of the organizational hierarchy is trembling because they do not understand, in the old model, enough about the technical material really to manage the work.
There are some big exceptions: The US DoD is able to get leading edge research into systems into the field. E.g., GPS, the SR-71, indeed, the H bomb. And so can the pharmaceutical industry. And what has been done in the lithography for microelectronics is astounding.
But, my view of information technology in Silicon Valley, it can't: That is, send them a project plan with a technical paper, and Sand Hill Road will just toss it into the trash; really. NSF, NIH, DARPA, ONR, etc. are fully able to review and evaluate technical material for projects, but nearly no one on Sand Hill Road can. Indeed, nearly no venture partners have the qualifications to serve as problem sponsors for DoD, DoE, NSF, NIH, etc., tenure track faculty in STEM fields in research universities, direct research, get research grants, review research papers, publish in good journals, serve on the editorial boards of good journals, etc.
Elsewhere in the US economy, middle managers do not want to bet their careers on technical projects they do not understand from their highly technical subordinates.
Basically the need is to get people with good technical backgrounds in positions where they have the authority to request, evaluate (or have evaluated), approve, fund, and oversee projects that exploit the pure/applied STEM work on the library shelves.
So, for the high level technical material we need to exploit, we need middle management well enough educated with such material to be able effectively to manage the exploitation. At least the middle managers must be able to direct competent evaluations of high end technical material.