Hacker News new | ask | show | jobs
by graycat 4387 days ago
Here's what to do:

(1) Problem. Pick a "big ass" problem where the first good or a much better solution will clearly be a 'must have', not merely a 'nice to have' and will be very valuable, in the sense of making money, the green kind. Note the "clearly". If pick a problem that does not satisfy "clearly", then try again. I didn't say that the success would be easy to do.

Part of the reason VC has been "digging in the wrong place" is that it tries to make money with small problems, where the solution is just 'nice to have' instead of 'must have'.

The ideal example of a 'big ass' problem, although not in information technology, would be a safe, effective, cheap one pill cure for any cancer. "Breakthrough"? Yup. "Crazy"? Sure; NIH has been throwing billions a year at that. But it would be a 'must have'; no question about 'product-market fit'. A 'must have' doesn't have any questions about 'product-market fit'. If looking at projects where the results have some doubt about 'must have' and likely need to work on 'product-market fit', then are looking at the wrong projects, are "digging in the wrong place".

Another fad for teenage girls does not qualify as a 'must have'. Yes, such a fad might make a $1 billion, but it is too difficult to evaluate. Instead, I mentioned "clearly" be a 'must have'. That is, it is crucial to be able to get a good evaluation early on, and nearly never can do that with fad products, as, indeed, the Andreessen remarks so clearly indicate.

(2) Solution. Find the needed solution. Often the solution will not be easy to find, but this need not concern VCs since they need look only at deals where a solution has already been found. If can't find the needed solution, then return to (1) and try again. A solution that is difficult to find can give a technological barrier to entry.

Q. Whatever you are working on, someone else has already done that.

A. Let's consider that: Take all the people who have done that. Then, except for rare ties, there must have been a first such person, and for them no one else had already done that. So, the claim is false. Sorry 'bout that. Instead, there really are things that are significantly and genuinely new.

Q. Then someone else can just reproduce the solution.

A. Some solutions are quite difficult to find, may be based on original work and/or advanced prerequisites understood by only a few people, implemented in software, secure inside a server farm, and, thus, provide a technological barrier to entry.

(3) Prove that really have a solution as described. When have such a proof early on, usually it is just on paper, especially for solutions in information technology. Is this possible? Yes, for nearly all of science and engineering on practical problems, it's routine and standard. It's how we build a new airplane engine, a new airplane, a new rocket, a new satellite, a new application of carbon fiber, etc.

(4) Evaluate the paper solution. Can this be done? In nearly all of applied science and engineering, such evaluations are standard and routine; it is insisted that such evaluations be done. No one puts up a new satellite with 'agile' techniques iteratively 'searching for product-market fit'. There's no counting on "magic". Fads of teenage girls are just irrelevant.

(5) Develop the solution.

(6) Deliver, sell the solution, the first good or much better "must have" solution.

Can these six steps be done? Sure; history shows that there's no doubt. Thankfully for US national security, the best examples are funded and exploited by the US DoD.

A great example was the SR-71: Problem: Get pictures of the USSR. Solution: An airplane that can fly at Mach 3+ at 80,000+ feet for 2000+ miles without refueling. Proposed essentially just on paper by Kelly Johnson and the Lockheed Skunk Works. The project passed evaluation and went ahead as planned.

Other examples: GPS and the earlier version by the US Navy -- proposed on the back of an envelop by some physicists at the JHU/APL, A seafloor, passive sonar array. The F-117 stealth plane; flew across Saddam's anti-aircraft artillery without a scratch; another Lockheed success. The Keyhole satellite, that is, a Hubble but aimed at earth.

Outside of the US DoD, the LHC; found the Higgs as planned. The WMAP satellite -- found the details in the 3 degree K background radiation as planned and got enough resolution to permit an estimate of the big bang 13.7 billion or so years ago.

These projects were evaluated essentially just on paper and worked as planned.

Generally such science and engineering has a much better batting average than VC. E.g., in each case, the SR-71, GPS, the LHC, WMAP, there was a problem that needed a 'must have' solution, such a solution, evaluation just on paper, and success. No "magic". No getting 3000 projects, looking carefully at 200, picking 20, and having 10 fail.

What should VC do? Sure: Stop "digging in the wrong place" and, instead, follow steps (1) - (6).

Will VCs get many such project proposals? Well, likely now they get more than they are willing or able to evaluate, but with interest in and success with such projects, they stand to get more such proposals. Considering how few VC projects pay off now, VCs don't need to get more than a few dozen good proposals each year. Digging in the right place doesn't require moving much dirt to get to a lot of gold.