Here are recent rough notes from the site.
Fed governor Christopher Waller argued recently in a talk that the upward-skew in BLS jobs data means that last year was likely only the third time since 1940 that the US lost jobs outside a recession.
Inevitable, but worth noting that state energy regulators are finally revisting rate-base-driven systems-utilities earning a premium on deployed capital—and their role in increasing (consumer) energy costs. The combination has been a flywheel for runaway capex spending, made worse by by AI data centers.
https://www.utilitydive.com/news/new-jersey-regulators-begin-rethinking-electric-utility-business-model/812799/
I have argued repeatedly that the AI-driven anti-SaaS arguments, like the attached, are silly and incoherent.
1/ If SaaS is going to be replaced by SaaS, then SaaS isn’t going away
2/ People don’t buy enterprise software because it’s complicated. They buy it so they don’t have to do it, and so they have someone to shoot at when it doesn’t work.
There is a confused "leak" from OpenAI about cutting capex spend over the next 4 years, from $1.4 to $600b. But the piece doesn't make clear if they are the same spend buckets, but someone there is trying hard to leak a spending cut.
A tidbit in the piece: Inference costs up 4× in 2025; adjusted gross margin down from 40% → 33%. Price deflation and competition are hammering it.
https://www.reuters.com/technology/openai-sees-compute-spend-around-600-billion-by-2030-cnbc-reports-2026-02-20/