Daily Alpha · Substack
· Premarket Alpha · by Buzzberg Research
Noahpinion's full-text roundup was the substantive Substack item: AI regulation risk is rising, SaaS revenue accelerated with coding agents, and growth forecasts run above expert scenarios.
Themes on this desk
AI regulation
A bipartisan majority now worries about existential AI risk and thinks AI is moving too fast; Obama is urging Democrats to make AI safety a tentpole issue.
AI as complement to software
Stripe Economics data show publicly traded SaaS revenue grew faster after coding agents emerged, and SaaS stocks recovered the ground lost in early 2026.
AI growth realism
A 15% AI growth bet is far above the 5.3% most optimistic expert scenario, with diffusion, J-curve and physical-robot bottlenecks.
Noahpinion: AI is becoming a permanent central force, with public opinion turning against it
Noah Smith argues AI will be a permanent central force in the world economy like industrialization was from 1870-1970, and that a bipartisan majority of Americans now worries about existential AI risk and thinks AI is moving too fast, with Barack Obama urging Democrats to make AI safety a tentpole issue; a strange coalition (Trump administration, investors, libertarians/techno-optimists, China hawks, and progressives who denied AI works) is opposing the slowdown but losing in public opinion.
Rising bipartisan support for slowing AI implies growing regulatory/policy risk for AI-exposed equities and capex plans, even as the current administration resists.
Watch Whether US AI-safety legislation or executive action advances, and whether AI-related polling majorities persist or fade.
Source →Noahpinion: SaaS revenue grew faster after coding agents, refuting the 'SaaSpocalypse'
Smith cites Ernie Tedeschi/Stripe Economics, which tracked an index of publicly traded SaaS companies and found revenue grew faster since coding agents emerged, with SaaS stocks recovering all the ground lost in the early-2026 repricing (nearly $1T wiped out by early February, perhaps $2T including subsequent weeks). Salesforce is cited as the expected victim, but companies still pay for software rather than rolling their own with Claude Code, and SaaS firms likely improved their own productivity with AI.
AI is proving a complement rather than a substitute for incumbent software, undermining the bear case that coding agents destroy SaaS business models and revenue.
Watch Whether SaaS revenue growth and stock recovery persist, or whether new AI-centric business models begin disrupting incumbents.
Source →Noahpinion: 15% AI growth bet is far above AI experts' own most optimistic 5.3% scenario
Smith highlights a public bet pitting AI researchers forecasting 15% economic growth against skeptical economists; AI experts' most optimistic scenario averaged 5.3%, and China hit 15% growth only once since rapid growth began, in 1984, when it was very poor. Alex Imas and Ben Moll list limits on near-term AI growth including slow diffusion, the J-curve productivity dip, difficulty automating physical robots, political barriers, production reorganization, the relational sector, bottlenecked chips, Baumol's cost disease, and AI disasters.
If the extreme AI-growth forecasts are wrong, the most aggressive AI capex and valuation assumptions are vulnerable; the enumerated bottlenecks (chips, physical robots) are specific constraints to monitor.
Watch Whether realized productivity/growth data accelerate toward double digits, or whether diffusion and J-curve effects keep growth in the economists' range.
Source →