If the “moat” is not AI technology itself but merely sufficient other lines of business to deploy it well, then that’s further evidence that venture investments in AI startups will yield very poor returns.

It's funny that a decade ago the exit strategy of many of these startups would have been to get acquired by MSFT / META / GOOG. Now, the regulators have made a lot of these acquisitions effectively impossible for antitrust reasons.

Is it better for society for promising startups to die on the open market, or get acquired by a monopoly? The third option -- taking down the established players -- appears increasingly unlikely.