An observation about big gains in technology is that it is the result of pushing a rate limit to the extreme.

Most people think that the limit has a reason to exist, so they will play around it, scraping with many others marginal innovations and leaving no outsized winners.

Very few wonder what would happen if you push that rate by 10X or 100X. I think it’s because it takes two realizations. The first is that the limit is rarely physics based, therefore it can be improved with lots of engineering, and, that the 2nd order effects are extremely hard to predict, so it’s not obvious what it unlocks.

The biggest recent innovations have this signature. The people that worked on them don’t have good predictions about the 2nd order effects, so they figured out the scaling and worried about the business opportunity later.

Tesla pushed Wh/area to unlock the best performing cars. SpaceX pushed tonnage to orbit to unlock satellite communication. OpenAI pushed neural net parameter count to unlock chatgpt. Neuralink is pushing bits/s between brain and computers to unlock …

These teams focus on dramatically increasing rate limits and then capture the huge opportunity that arises uncontested.