Apologies for the claude dump, but this was too tempting to resist and partly also my context window is too small to generate a long thesis (a.k.a lazy). I wanted to share it with people as a conversation starter since many of us are grappling with this phenomenon.

I had a realization that everybody using AI makes things worse off for everybody because a) more of inferior things and b) they are inferior since we would be using it in areas we are not good at i.e. experts in their fields collaborating to build something versus them independently building it with an AI. Even if one assumes the best AI, it is me who when is far out of his competence that my creation is of a lesser quality. These people can still use AI and collaborate to surpass previous frontiers, but the current zeitgeist is that one doesn't need other experts now.

This led to gaming out AI adoption scenario with claude from a game theoretic approach for both an organization and an individual perspective.

The response in part 1 was expected that AI adoption is a dominant strategy.

In part 2, I asked it to reconsider the implicit assumption that adoption is beneficial. It may be good to wait if the cost of adoption is greater than the probability of benefit of adoption. The interesting part is the portfolio approach for the individual, akin to investing.

In part 3, I asked it to break down the cost model for different domains and while it pulled out the numbers out of the hat, the reasoning for software/eda seems like a correct approach but we see the opposite happening today in the industry. What I wanted to ask is what others think these would be for their respective domains.