Part II So, what's going on here? I start with five general points: (1) Status of Venture Capital. In a word, now and for the past 10 years or so, there is

Part II So, what's going on here? I start with five general points: (1) Status of Venture Capital. In a word, now and for the past 10 years or so, there is just one appropriate word for venture capital -- it sucks. To be more clear, the returns to the limited partners from the VC 'asset class' have been not just poor but commonly negative. It may be that there are fewer than 20 VC firms in the US that have a good track record for their limited partners. From a graph posted some months ago by VC Mark Suster, over the past 10 years the number of venture partners has fallen from a few thousand to a few hundred, or some such. The VC layoff rate compares with the UAW Detroit layoff rate. (2) What's Available in Practice. It's easy to guess that, being on the VC side of the table, it's not easy to find really promising projects. That is, VCs that say that they need and want more and better proposals likely have a good point. For me, I'd rather be on my side of the table so that if I want a better project all I have to do is think of one; on the VC side of the table I'd have to wait for such a project from entrepreneurs -- bummer. (3) What's Available in Principle Still, whatever the daily reality is for VCs as they read their e-mail, in principle the entrepreneurs and their VCs should be printing money in huge piles. Why? There is one, just one, key to printing money: Create economic value. For that key, there is one, just one, main approach: Increase economic productivity, that is, turn out more, much more, value per person-hour worked. For more productivity there is essentially just one good approach -- automate the work. For automation, there is now exactly one, unique in all of history, grand champion technique -- computing, especially aided by the Internet and readily available infrastructure software. So, find some much needed activity and automate it. (4) What's Available with Some Effort. Whatever work X the entrepreneurs and their VCs are doing, there's more, much more, that is clearly available, with a fantastic track record, far better than the VC track record, that clearly totally blows the doors off X. We get the examples of how to do such powerful technology from the all-time, unique world-class, grand champion of everything technical, the US DoD who, thankfully for US national security, has been wildly successful for about 70 years now. War story: Around 1940, the American Chemical Society, seeing war coming, sent a representative to the US War Department offering to be a liaison with the US community of chemistry. Soon he got back a letter of thanks but saying that the War Department already has a chemist. About five years later, there had been a, uh, change of attitude: Eisenhower said "Never again will US science be permitted to operate independent of the US military.". Vannevar Bush, James Conant, etc. arranged that US research universities would get "an offer they couldn't refuse": Take DoD money for research or cease to be a competitive research university. Result: Since then the top few dozen US research universities have gotten about 60% of their annual budgets from research grants and have dominated academic research in the world, and the US DoD is right there, sure they are not falling behind in fundamental research and getting a good stream of well selected and trained researchers. Congress and the US DoD take research and technology very, very seriously. So, comparing with entrepreneurs and their VCs, I will mention two differences: (A) In the system set up by the DoD, given a problem, it is common to attack it by doing some original research, e.g., that can pass peer review. Often the work -- in physics, electronic engineering, mechanical engineering, chemistry -- has been highly mathematical. This research, where to do it, how to do it, the standards, and the culture, are rooted in the best US research universities but also extend to many laboratories and companies working on US national security. (B) Given some research, the results can quickly receive quite competent and serious evaluation, often from presentations just on paper. With that evaluation passed, the project can move quickly through prototype software, breadboard hardware, demonstration systems, fully militarized design and prototype development, full testing, and large scale production and deployment. Given a good evaluation of the research, the rest has an excellent track record of success. E.g,, at one time I was programming in the group that did the orbit determination for the Navy's version of GPS, and there I learned the story of how that system had been proposed to the Navy: Not much more than some physicists scratching on the back of an envelope. And the final system worked as planned, and a test receiver on the roof routinely 'navigated' its position to less than one foot. Yes, the later Air Force GPS is more accurate. But the Navy's system was just crucial for the US SSBN fleet and its missile targeting for years. So, in strong contrast, for the VCs, for (A), they just will NOT take research seriously. They just will NOT do it. Indeed, we might look at how the US research universities and the US DoD evaluate research and the qualifications they use: A researcher who publishes some good research can be invited to be a reviewer of a journal or of proposals for funded research. A good reviewer can be invited to be a journal editor. A good editor can be invited to be a journal editor in chief. Etc. Then that system just will not permit itself to be seen as unable to evaluate original research, no matter how advanced, no matter what the prerequisites, no matter how involved. Period. Write a paper with title "P = NP" and submit it to any of at least a dozen journals, and within a year will get back a letter with "The first serious error is at line n on page m" or "Your check for $1 million from Clay Mathematics is in the mail.". Sure, the path to reviewing papers is usually a long and challenging Ph.D. followed by a few years of successful research. Editors in chief typically have plenty of gray hairs. This system for evaluating research is highly competent but challenging. Then, for the VCs, in strong contrast, apparently one could count with shoes on all the VCs who could even direct a competent evaluation of some research; this is not a surprise since one could count with shoes on all the VCs with a relevant Ph.D. and a solid research record. Next, for (B), funding based just on research just on paper, the VCs won't do it. At most all a proposal will get back are laughs. Send a VC a paper with some research, and the VC may break a leg slapping his thigh in wild laughter. The US DoD has been funding projects based on research just on paper to knock balls out of the park and into orbit with high reliability from early in WWII to the present, and VCs won't even come to the ballpark. So, net, we have to suspect that much more ROI would be available to the VCs "with some effort", that is, by taking research seriously. (5) Fundamental Issue. If we ask that each entrepreneur found "another Google" and be worth $200 billion, we start to have doubts. E.g., the whole US economic pie is not that big, and we wonder if computing, etc. is powerful enough to grow that pie that big soon. So, without a big change in pie size, the number of $200 billion successes is small, only a few each decade. I believe that by taking research seriously we can multiply this few by several quickly. So, a fundamental issue is that we are still limited to not very many $200 billion successes per decade. But apparently for my claim of "multiply this few by several quickly", there is hardly a VC or limited partner in the country who would believe any such thing and fewer still, say, zero, who would bet more than 10 cents on it. They would be afraid they were just being swindled, and they don't want to be swindled out of even a 10 cent cup of coffee. Again, VCs and their limited partners seem to be suck solidly in evaluating projects just by traditional financial accounting but using some surrogate measures. In particular, they look at 'traction' but will NOT look at research.