THE SWELL #20
SAIL Weekly Digest June 1–5 | Issue #20
This week SAIL authors sized up a boom that keeps outrunning its referees. Agents took over more of the software stack and video models lined up as the next frontier, while the economics — no bubble yet, but no bottom-line payoff either — refused to resolve. Policymakers, governors, and the Pope all reached for the same lever: making AI answer to someone.
The Week in AI
NSA tapped to lead new frontier AI benchmarking — President Trump signed an executive order establishing a voluntary evaluation and early-access framework for frontier models, placing the National Security Agency in charge of the benchmarking process amid reports that officials are discussing government equity stakes in major AI firms.
The Great American AI Act hits DC — Reps. Obernolte and Trahan released a 269-page bipartisan discussion draft featuring a $100M-a-year Center for AI Standards and Innovation and mandatory third-party audits, but the bill faces immediate pushback over a clause that would block state-level AI regulations for three years.
Anthropic warns of recursive self-improvement — Citing data that Claude now generates more than 80% of the code merged into its own production systems, Anthropic urged global leaders to retain the option to temporarily pause development as AI-driven AI engineering accelerates.
This Week from SAIL Authors
The Boom’s Balance Sheet
We checked. Again. Still no bubble. — A customer-led boom with a few fraying edges. — Exponential View. Read more
Why AI isn’t showing up on your bottom line — A framework for understanding your firm’s AI transformation, and why the spend outruns the gains. — Exponential View. Read more
Agents and the Model Frontier
The Age of Async Agents — Cognition’s Walden Yan & OpenInspect’s Cole Murray — 80% Devin commits, spec-to-PR workflows, full VMs, agent memory, and PMs shipping code. — Latent Space. Read more
Why Video Agent models are next — Ethan He, xAI Grok Imagine — A first deep dive with the lead who built Grok Imagine in three months, and the case for video agents over world models. — Latent Space. Read more
Open and closed models are on different exponentials — Where marginally higher intelligence drives value, and where it doesn’t. — Interconnects. Read more
Policy in Practice
Adventures in Vibecoding Policy — Can policy microsites save America? Shipping policy ideas like software. — ChinaTalk. Read more
Arizona’s Abundance Playbook — A view from the governor’s office on building the abundance agenda at the state level. — ChinaTalk. Read more
Minds, Machines, and Meaning
Render Unto Caesar, Not Unto Claude — The Pope’s AI takes, explained. — ChinaTalk. Read more
Does AI make you dumb? And why our forecasts suck #576 — The cognitive cost of leaning on AI — plus jobs in high demand, ByteDance, and anti-drone nets. — Exponential View. Read more
Full Library
Access the complete, searchable archive of SAIL Media in our Sitemap.
We checked. Again. Still no bubble. — Azeem Azhar | Exponential View
Render Unto Caesar, Not Unto Claude — Aqib Zakaria | ChinaTalk
Why Video Agent models are next — Ethan He, xAI Grok Imagine — Swyx | Latent Space
Does AI make you dumb? And why our forecasts suck #576 — Azeem Azhar | Exponential View
Open and closed models are on different exponentials — Nathan Lambert | Interconnects
Adventures in Vibecoding Policy — Jordan Schneider | ChinaTalk
The Age of Async Agents — Cognition’s Walden Yan & OpenInspect’s Cole Murray — Swyx | Latent Space
Arizona’s Abundance Playbook — Jordan Schneider, Aqib Zakaria & Lily Ottinger | ChinaTalk
Why AI isn’t showing up on your bottom line — Azeem Azhar | Exponential View
