Maybe OpenAI will finally use open methods to benefit all humans, not one human or a few

That clarifies why there have been no innovations or corrections of small things. The people doing the work are afraid for their jobs and afraid to raise either issues or opportunities, let alone corrections and improvements. The LLM approach is fundamentally flawed, but correctable, if the organization is open and communicates. That culture of closed central control of all things, might have come from one person. It affected Microsoft Bing CoPilot and other derivative developments and experiments. And set a bad example for all LLM projects.

One person can completely waste billions of human hours, more than the value of the investments. A net loss to human society. I feel sorry for OpenAI and its customers, and their customers. But, more for the 5.2 Billion Internet users and 8.1 Billion humans whose time was wasted from “greed of a few”.  It is not just a global world, but a tightly coupled fast world that depends on honesty, clarity, lossless records, verifiable processes, and responsible humans and systems. And where tiny errors from inattention and carelessness can cause death and misery for hundreds of millions of innocent humans.

Richard Collins, The Internet Foundation