Korva is a scroll of the internet’s best ideas.
Warnings and base rates
In any hard field, experts give the same warnings. Young people can't build chips, nobody should start another semiconductor company, the base rate is too low. Almost all of it is totally right, except when it isn't.
The partner's saying
A saying passed to one investor sums up the logic: if it doesn't work, it will be for all of the reasons your partner gave, and if it does work, it will be because those reasons didn't matter. So the question becomes whether there is a path where they don't matter, and whether anyone can see it.
Naivete as an asset
Not knowing the full difficulty helps too. Cerebras took naivete: the investor knew little about physics or the supply chain behind a chip like that, and the effort was worth venture capital whether it worked or not. Build something people have tried for 50 years and failed at, when there is now a reason and an application for it.
Why can a high failure rate still pay?
A venture investor's loss is capped at the amount invested, while a win has no cap. A single company that grows to many times the size of a fund can cover a long list of failures. Funds are built around that skewed payoff. This is why a field with a high failure rate can still attract investment.
Does not knowing the difficulty help?
The economist Albert Hirschman described the Hiding Hand principle in his 1967 book Development Projects Observed. He examined World Bank-funded projects and found that planners underestimated both the difficulties and their own ability to solve them. The unexpected problems pushed teams to find creative fixes, and many projects succeeded because of it. Hirschman saw the underestimate as an essential part of how ambitious projects got started.
Why did earlier wafer-scale attempts fail?
Trilogy Systems, founded by the computer architect Gene Amdahl in 1980, set out to build a wafer-scale computer. It raised about 230 million dollars, a record for a startup at the time. Defects in the silicon, packaging problems, and a damaging factory flood hindered the effort. The company abandoned wafer-scale integration in 1984.
More from Invest Like the Best with Patrick O'Shaughnessy
