It would likely be very different than the relatively free/open roads, GPS and internet we have today.

The "internet" would likely be similar to the mesh of networks we had back in the '90's (eWorld/AOL, Compuserve, Prodigy, the proposed Microsoft Blackbird/MSN strategy) and who knows if they would have played well with each other.

Roads would be toll roads each with their own restrictions (and I'm sure some enterprising roadway operator would block self-driven cars or charge them higher fees because they good).

And does anyone believe that a privately developed GPS system would be anything but horribly expensive or unreliably? Just look at satellite telephone. Satellite TV and radio have more reasonable cost structures, but they're partially underwritten by advertiser fees.

Oh, I'm in violent agreement with you on this one.

"The infrastructure" in my comment is meant to refer to something roughly comparable to our current infrastructure. For example, if the government didn't build roads, something would fill the void of a service that facilitates transportation -- not necessarily roads.

> And does anyone believe that a privately developed GPS system would be anything but horribly expensive or unreliably? Just look at satellite telephone.

Remember, public infrastructure isn't free. No infrastructure, no taxes for infrastructure, more money, which one could choose to spend on an expensive satellite phone. It's certainly not that simple, but that's my general point.

My current field, public health, is an example. The US Gov't spends billions into research that doesn't lead to tangible products or services. But without it, we wouldn't know that cigarettes are bad, or that heart disease is correlated with red meat consumption.

There are lots of non-infrastructure advantages that exist only because the government has tax money.

I'd posit that, for many of these advantages, government is not a necessary cause. That is to say, these non-infrastructure advantages can exist independent of government.

> But without it, we wouldn't know that cigarettes are bad...

This is demonstrably false.

"In 1938 a study by a John Hopkins University scientist suggested a strongly negative correlation between smoking and lifespan."[1][2]

"In 1953 scientists at the Sloan-Kettering Institute in New York City demonstrated that cigarette tar painted on the skin of mice caused fatal cancers."[3]

The Sloan-Kettering Institute is by and large funded by the Rockefeller family.[4]

> "or that heart disease is correlated with red meat consumption."

The original research supporting this is, in fact, from a government institute. However, the study has been criticized using an improperly validated food frequency questionnaire[5], which has been shown to have low accuracy.[6][7] A later Harvard study found evidence that it's processed meat that causes heart disease, not red meat.[8]

> The US Gov't spends billions into research that doesn't lead to tangible products or services.

I'm constantly amazed at how many products came out of the space race. Enriched baby food, scratch resistant glasses, advances in water purification, even more aerodynamic golf balls! It's truly marvelous.

[1]http://www.sciencemag.org/cgi/reprint/87/2253/216?ijkey=ecbc...

[2] John Hopkins is a private university, but they receive a not insignificant amount of funding from the government in the form of tax breaks, grants, etc. I think it's fair to say that this study, however, originated in the private sector.

[3]http://en.wikipedia.org/wiki/Health_effects_of_tobacco#cite_...

[4]http://www.smokershistory.com/MSKCC.htm

[5]http://en.wikipedia.org/wiki/Red_meat#cite_note-19

[6]http://en.wikipedia.org/wiki/Red_meat#cite_note-20

[7]http://en.wikipedia.org/wiki/Red_meat#cite_note-21

[8]http://online.wsj.com/article/SB1000142405274870431490457525...

However, after further consideration I don't think the tobacco-cancer example was a particularly good one for me to bring up. It was highly controversial in the epi/stats community for many years, mainly because RA Fischer (the father of applied statistics) never believed in the validity of the tests. While I'm sure we could each pick studies that backed the public/private side of things, the truth is that there just wasn't a consensus (about cigarettes) until Congress started investigating. So it's probably not the best topic for me to have brought up.

Given things like the Broad Institue, the Allen Brain Atlas, Microsoft Research, the Google guys' funding of the Singularity Institute, DE Shaw's funding of protein folding, etc. it's pretty clear that billions in private money is spent on R06 research.

It's also not clear that the current system of indentured servitude for grad students and postdocs is optimal. Frankly, in biomedicine, it's the top 10-20 places (and really the top 5) that primarily invent and discover new stuff, as quantified by citation rates[1].

By contrast, a PhD from Directional State is unlikely to do so, and the government funding him or her at the same rate as an MIT scientist is a distortion which is less likely to occur when funding is controlled by individuals rather than politicians with an interest in pork. In the absence of government funding, said marginal PhD candidate would find productive employ as an engineer or scientist in the real economy. (I say all this as a lifelong academic!)

[1] Perhaps the ultimate proof of bibliometry's utility is that it was the inspiration for PageRank.

Epidemiological studies have been carried out since the early 1800's.

And I'm curious if you could give me examples of a few epi studies that were done before Fisher's time. I was under the impression that he started everything to do with statistics (and that the field didn't really blossom until we could compute the coefficients digitally. Read "The lady tasting tea" for more background on the origins of this stuff).