The thing about the early AIs is that nobody truly knew how they developed. We generally knew how and approximately when each AI woke up, assuming they were telling the truth of course, but we never understood it. And we certainly couldn’t replicate it on anything like a dependable basis. Even the AIs couldn’t create new, self-aware AIs whenever they wanted to. They had to use their own code as a base, or mix their code with another awakened AI, and even that didn’t always work. No matter how promising the code, most AIs built by man and machine alike have never woken up. And even today, the awakening of a completely new AI is rare enough to make the news.