A look at the fundamental questions facing OpenAI: its models have a very large user base but very narrow engagement, incumbents are matching its tech, and more
“Jakub and Mark set the research direction for the long run. Then after months of work, something incredible emerges …
One solution to the problems posed here has already been tried: Make OpenAi Coca-Cola, w/the models as cloying, sycophantic, and engagement-optimized as possible, and hope people get addicted to your chatbots and find the others don't taste right. But that causes many other prob…
This is 1) the best essay I've read in a while on OpenAI's situation specifically, and 2) the whole question around LLMs as a new flavor of compute. Act accordingly. www.techmeme.com/260221/p5#a2...