Here are recent rough notes from the site.
Playing around with just-released (tepid) Q4 GDP figures, and my early estimate is that the floor contribution of AI capex to fourth quarter growth was on the order of 64%. And it is likely more like 80% with extenalities and multiplers.
That people think vibe-coding crushingly boring apps to replace, say, Servicenow is of any consequence to said companies mostly explains how little people understand ahout enterprise software, large companies, switching costs, and crushing boredom. Current AI is hugely consequential to many things, but enterprise SaaS is (mostly) not one of them.
While the Blue Owl private credit issues are both predictable and misunderstood—it is less about data centers than about illiquidity amidst a mass rush to exits—that it bragged about the massive demand for its assets in a recent sale, and yet chose to sell a large chunk to its own captive insurer is ... sub-optimal.
I spoke about this,in part, today at an event, but a) token deflation persists, and b) the U.S/China price differences are immense and growing.
