It generally takes about 4-5 years of being in tech before people seem to recognize this, if at all, and even then there is deep resistance to accepting it, even among VCs. There’s always some VC trying to buck the trend and solve for median 3x-10x return instead of outlier 20x-100x return (eg the old “black swan vs moneyball” debate ~2009-10). It never works, and I only understood why after like a decade. Bill Janeway’s book Doing Capitalism in the Innovation Economy covers the gist but incompletely. VC is the rare investment class where past investor success is in fact predictive of future success, unlike “fooled by randomness” Wall Street. VCs let this go to their heads like they have a magic touch, but it’s actually a statistical artifact of a) solving for black swans b) non-ergodicity in tech evolution and investment opportunity and c) SV being small enough, the “winners” of a generation have talent-market clearing evaporative cooling, and “cornering” effects. I think of this as “gravity altering.” All these are downstream of Moore’s Law as the root cause btw.

The result of this is that SV has a distinct political economy that favors blitzscaling. What Bezos captured as the “get big fast” effect. It’s not just get big fast, but big enough to alter the gravity of the industry. The last time this was possible for the economy at large was railroad stocks in the late 19th century, when the economy was small enough, and railroads big enough a piece of it, they could alter macroeconomic gravity.

What SV has failed to recognize is that the “world” is a much bigger thing than SV’s economy. Even though they dominate the indices, they don’t have gravity-altering levels of economic mass at full-economy scale. So blitzscaling doesn’t actually work. This has nothing to do with the morality of govt vs private. It’s sheer impedance mismatch.

Warp speed is not an example of blitzscaling I think. The IP there was sitting on shelves and scaling was not as hard, and was unrelated to Moore’s law. That was closer to retooling car factories to produce tanks in WW2. And WW2 “arsenal of democracy” btw was not imo a blitzscaling example either. That was more a partial command economy that paid off because the allies won.

SV tech economy is a very anomalous beast and only exists fundamentally because of Moore’s Law. It’s as close to a monocausal historical chapter as we’ve ever seen. The model does not generalize to where Moore’s Law effects are weak. You can see how badly SV stumbles when it ventures where Moore’s Law isn’t the main factor. They’re in deep denial about this. Yes software and AI can and will eat the world, but they won’t trigger a Moore’s law for everything. Atoms fundamentally don’t evolve at the speed of bits.