← The Library
The world · a living book
The State of AI
The frontier AI story, updated as it moves — models, agents, the money, and the fights.
3 chaptersupdated July 2026sources linked in every chapter
The story so far
Three labs — OpenAI, Anthropic, and Google — now trade the lead back and forth roughly once a month. The machines write most of their own makers’ code, agents are creeping from demo into real work, and governments have started treating the newest models as strategic technology rather than software.
The open questions haven’t moved: whether the ~$690 billion a year going into data centers is rational, whether entry-level jobs are eroding or just changing shape, and how dangerous these systems really are. This book follows all three. By mid-2026 the fight also turned commercial — Anthropic and OpenAI now trade claims over who leads on revenue — while Washington began drafting the first rules for how frontier models get released — and by mid-July those rules were taking concrete shape on both sides of the Atlantic.
Chapter 1 · July 2026
Where the frontier stands
As of mid-2026 the frontier is a three-horse race running at a sprint. OpenAI shipped GPT-5.5 in April, only six weeks after GPT-5.4, and made its fast variant the free-tier default weeks later. Google’s Gemini 3.1 Pro, out in February, more than doubled its predecessor’s score on the hardest public reasoning benchmark and carries a million-token context window. Anthropic released Claude Fable 5, its first “Mythos-class” model, in June. Across February to April the three labs put out seven frontier models between them — a new state of the art about every eleven days.
The government stepped in
The startling event of the year was political, not technical. In mid-June the U.S. Commerce Department briefly export-controlled Anthropic’s two newest models over a jailbreak concern, forcing the company to suspend them for all customers — the first time Washington treated a deployed commercial model as controlled strategic technology. The controls were lifted three weeks later, and the model was restored to users on July 1. It was a preview of a world where a model release can be regulated like a munition.
Agents at work
The clearest real-world shift is in code. Senior engineers at two leading labs said in January that AI now writes essentially all of their code. But raw output isn’t shipped output: an MIT study of more than 100,000 developers found AI agents produced about 180% more code while only about 30% more actually reached production. Gartner expects 40% of enterprise apps to embed task-specific agents by year end, up from under 5% a year ago — even as most companies report their pilots stalling before they go live.
Two rulebooks
Regulation split by continent. Washington moved to preempt state AI laws with a December executive order and a litigation task force, though Congress hasn’t passed the framework, so state laws — Colorado’s AI Act, California’s frontier-transparency law — remain in force. Brussels went the other way, delaying the AI Act’s high-risk obligations by sixteen months even as its enforcement powers over general-purpose models arrive August 2.
The open questions
Is AI erasing jobs, or upgrading them?
The damage is realEntry-level postings are down roughly a third since 2023, and unemployment among recent graduates has climbed above the rate for all workers.
Anthropic labor research It’s a raise, not a wipeoutPwC’s 2026 barometer finds AI-skilled workers command a 62% wage premium and that AI-exposed firms are growing headcount faster than the rest.
PwC 2026 AI Jobs Barometer $690 billion a year — bubble or rational boom?
BubbleHyperscaler spending dwarfs the revenue it’s meant to generate, and central bankers have flagged circular AI financing as a top risk to the financial system.
Goldman Sachs build-out analysis RationalThe spenders are the most profitable firms in history, cloud backlogs are surging, and the binding limit is power, not demand — capacity that gets built earns revenue.
Goldman Sachs Chapter 2 · July 2026
The race turns commercial — and Washington starts writing release rules
The frontier contest is now as much about business as benchmarks. Anthropic passed OpenAI on annualized revenue run-rate this spring — roughly $30 billion against OpenAI’s self-reported $24–25 billion — and the two spent the early summer trading claims over who leads. The split is structural: about 85% of Anthropic’s revenue comes from enterprise and developer customers, while a similar share of OpenAI’s comes from ChatGPT consumer subscriptions, most of whose users pay nothing. Anthropic told investors it expects its first-ever operating profit in the second quarter.
Washington moves from banning to standard-setting
After June’s abrupt export-control episode, the U.S. approach shifted from emergency to process. The White House is in advanced talks with OpenAI, Google and Anthropic on a set of voluntary standards for releasing frontier models — common benchmarks, pre-release testing timelines, and rules for who gets access. Reporting in early July also described an unusual idea floated around those discussions: that the government could take an equity stake in a leading lab, with one proposal putting a roughly 5% share of OpenAI on the table.
The models keep coming
The cadence hasn’t slowed. OpenAI began a limited preview of GPT‑5.6 Sol, Terra and Luna on June 26, saying general availability would follow in the coming weeks. That sequence is now recorded in the next chapter, where the July 9 launch is sourced to OpenAI’s own announcement.
A contested scoreboard
Has Anthropic really taken the lead?
Yes, on revenueIts run-rate has climbed roughly thirtyfold in fifteen months and now tops OpenAI’s, built on stickier enterprise contracts.
SaaStr Depends how you countThe figures are self-reported run-rates, not audited annual revenue, and OpenAI still reaches far more people through ChatGPT — a different business, not a clear loss.
Epoch AI Chapter 3 · July 2026
The rulebooks arrive as the models keep shipping
The release cadence held into July even as the rules around it hardened. OpenAI moved GPT‑5.6 Sol, Terra and Luna from a limited preview to general availability on July 9 across ChatGPT, Codex and the API. The company describes Sol as its flagship, Terra as the balance of intelligence and cost, and Luna as the efficient tier for high-volume work. Google, meanwhile, released Nano Banana 2 Lite on June 30 as its fastest, lowest-cost image model.
Two governments write the same idea
The bigger news was regulatory. On July 7 the European Commission published an Action Plan on Cybersecurity and AI. It says the Commission will strengthen Europe’s capacity to evaluate advanced models before they are placed on the EU market and plans a cybersecurity-focused evaluation capacity for 2027. Days earlier, a June 2 White House order created a voluntary framework for government evaluation of “covered frontier” models and set August 1 deadlines for initial benchmarks and secure-access guidance. The two documents differ in legal setting, but both put pre-release evaluation on the policy agenda.
Congress drafts another route
The legislative track filled in alongside the executive one. Representatives Jay Obernolte and Lori Trahan published a bipartisan discussion draft titled the Great American Artificial Intelligence Act of 2026. Its text would create a Center for AI Standards and Innovation and an Independent Verification Organization program; because it is a discussion draft rather than an introduced or enacted bill, those provisions remain proposals.
The open question
Should governments evaluate frontier models before release?
The case for evaluationThe Commission plan and White House order treat pre-release testing as a way to identify advanced cyber capabilities before broad deployment.
European Commission The implementation questionThe White House framework is voluntary, requires confidentiality and intellectual-property protections, and explicitly disclaims mandatory licensing or preclearance. The details will determine how that balance works in practice.
White House A living book: chapters are dated and grow as the story develops. Nothing is deleted — the record just gets longer.