I don't think this makes a rigorous or compelling case on better startups being formed during above-average rate periods, at all. There doesn't seem to be enoug

I don't think this makes a rigorous or compelling case on better startups being formed during above-average rate periods, at all. There doesn't seem to be enough data in the article to show correlation, even less causation, between high-rate environments and the emergence of better startups. The rise of the great wave of startups like Apple, Microsoft, etc. probably had very little to do with interest rates. Also: > Second, complementing this new energy model is a shift away from Moore’s Law and CPUs to the proliferation of GPUs. This would support scaling Moore’s Law through parallelism, which favors applications of machine learning and AI. As a result, the marginal cost of compute will go to zero. And here I am looking at the sticker prices of cloud GPU computer and going: really. I won't deny there's an era of AI products and apps coming with the emergence of OpenAI, Huggingface, etc but I'm not sure the societal or economic value will outweigh the costs, or that marginal costs will go to zero. I mean, yea, training AI models will go the way of traditional software, where marginal cost of distribution tends to zero: you build/train once (or a constant number of times), and you distribute infinitely. But the ground-level reality seems that many startups will foot higher compute bills than ever to get started.