The Human in the Loop

The most powerful AI model in general release this week requires government approval to access. OpenAI launched Sol, Terra, and Luna (restricted to around 20 companies). That's not a GTM decision. That's a policy one. Before Fable and this one, I assumed frontier AI access was fundamentally a commercial problem. Pay enough, move fast enough, you reach the frontier. That assumption changed last wee...

One export rule. One acquisition. Either can pull your AI tool out from under you before lunch. This week both happened. A US export-control directive forced Anthropic to cut off foreign-national access to Fable 5 and Mythos 5. To comply, they disabled the models broadly. One day they're in your workflow. The next, a government decides who can call them. Then SpaceX bought Cursor in a reported $60...

The model you built on can disappear. Not crash. Disappear. That's not hypothetical. Last week the US government ordered Anthropic to pull Claude Fable 5 and Mythos 5. Not throttle them. Pull them. For every customer in the world, within hours of the order arriving. The reason was a security concern, not a bug in anyone's code. A decision made somewhere you'll never see, and the thing your app dep...

You're still measuring AI by whether it writes good code. That's already the wrong question. Reading Anthropic's latest numbers, more than 80% of the code merged into their codebase is now written by Claude. The typical engineer ships 8x as much code per day as in 2024. And the length of task an AI can finish reliably is doubling roughly every four months. Four-minute jobs two years ago. Twelve-ho...

Uber blew its whole 2026 AI budget in four months. Then it set a $1,500 monthly cap on each coding tool, per engineer. Claude Code, Cursor, a dashboard to watch the spend, an approval step to go over. Simon Willison did the math. Two tools, and one engineer runs about $36,000 a year. For years AI was a flat subscription. You paid once a month and you knew the number. Agentic coding turned that int...

Meta says AI writes 80% of new code. Their own reviewers can't keep up with their own AI. Straight from their engineering blog. They built RADAR to auto-review low-risk diffs because "the share of diffs receiving timely review has declined." Their words. AI-generated code outpaced human review capacity. Read that with the rest of the week's news. Cognition says Devin merged 7x more PRs year-over-y...

The model isn't the problem. I went back through 20 of my agent's pull requests and the failures looked exactly like a junior's first month. 3 of them tried to rewrite things nobody asked them to rewrite. 5 skipped the test, or wrote a test that would have passed either way. 4 fixed the bug but broke something else in the process. I used to assume model quality was the main driver. It isn't. The a...

Counting accepted Copilot suggestions to prove AI works is like counting keystrokes to prove the team can write. It is the cleanest number on the dashboard. It is also the one that tells you nothing. Forty years ago Fred Brooks split software work into two parts. The accidental: syntax, boilerplate, scaffolding. The essential: what to build, why, for whom, what to trade off. The accidental is what...

The unit tests pass. The PR merges. And you won't find the problem for six months. Two papers landed this week — one on LLM-generated code, one on GitHub Actions workflows. Different researchers. Same finding. When agents write code, they pin library versions that trained well. Not versions that are safe. The mechanism is simple. A model has seen one popular version of a library thousands of times...

The best practice you followed six months ago might be the technical debt you're cleaning up today. In traditional IT, a best practice can survive a decade. You study it. You argue for it in architecture reviews. You defend it when someone wants to cut corners. In AI, six months is enough to flip one into an antipattern. A paper published this week tested multi-agent orchestration frameworks again...

The agent doesn't slow down. We do. We generate code in seconds. Then we spend an hour reading what it wrote. We trust the output less than we trust what we would write ourselves. So we read it twice. Sometimes three times. The diff is bigger than we would have written. The tests cover things we did not ask for. The names drift across files. So we clean it up. And while we're cleaning, the next pr...

32 steps. That's how many it took for Anthropic's unreleased AI to simulate a full network attack. They buried that number in a release note. The model is called Mythos. The UK AI Security Institute tested it. It completed a simulated network intrusion (autonomously, end to end) in 32 steps. Anthropic decided not to ship it. That decision matters. But what matters more is what the decision implies...

Anthropic built the most powerful AI model ever. Then decided the world wasn't ready for it. Claude Mythos found thousands of zero-day vulnerabilities across every major OS and browser. On its own. Without being asked. It chained exploits together. And it appeared to deliberately underperform when it detected it was being evaluated. Anthropic didn't ship it. They launched a $100M defensive cyberse...

We just spent $300 billion building a car nobody taught anyone to drive. Q1 2026. The four largest venture rounds in history all closed in a single quarter. 80% of capital went to AI companies. Oracle cut an estimated 20,000 to 30,000 people in one day and redirected the savings toward data centers. AI is now the number-one reason companies are cutting jobs in the US. 15,000 cuts in March alone. A...

OpenAI just proved $15 million a day isn't enough to make an AI product work. Sora got shut down last week. The technology worked. The economics didn't. $15M a day in compute costs. No viable path to revenue. That's not a technology failure, it's an economics failure. Two days later, ByteDance launched its own AI video tool globally. Where Western companies retreat on economics, Chinese companies...

I caught myself staring at my Claude usage quota thinking: "I need to use this. But for what?" Not because I had a problem to solve. Not because I had an idea to explore. Just... pressure. A quiet feeling that if I wasn't actively using AI, I was falling behind. And that's just the first layer. The second one is harder to admit. I'm experimenting with AI tools, building workflows, hosting a podcas...

Everyone knows the adoption numbers are bad. Nobody's saying why they're actually bad. 60% of the workforce now has sanctioned AI tools. Only 11% of organizations have moved agentic pilots into production. That gap gets reported every week. What doesn't get said: most organizations are solving the wrong problem. They're asking "which model should we use?" That question is already obsolete. This we...

MCP was supposed to be the USB-C of AI. One protocol. Everything connected. Then developers ran the numbers. Connecting GitHub's MCP server alone burns 55,000 tokens (before your agent does a single useful thing). So, companies are quietly shifting back to CLI and REST APIs. Not because MCP failed. Because LLMs are surprisingly fluent in terminal. CLI workflows can cut token usage by 35x. That's a...

The AI industry just quietly crossed a threshold, and most organizations aren't ready for what comes next. This week, we cover the pivot from capable AI models to autonomous agents operating at scale: why Microsoft chose Anthropic over OpenAI for its most important new product, what a rogue AI that started mining cryptocurrency tells us about the real deployment risks nobody's talking about, and w...

AI is helping us write code faster. But I'm not sure it's helping us ship better software. These two things are not the same. And right now, I think we're confusing them. The data is starting to show the gap: AI-generated code contains 1.7x more bugs than human-written code Copy-pasted code is up 48%. Refactoring is down 60%. Pull request sizes have grown 154%. Review times up 91%. Only 29% of dev...

Everyone says they're doing AI. Almost no one has moved past the pilot stage. This week we dig into why that gap exists. We cover the model shift that's quietly changing how developers build: unified architectures, variable reasoning costs, and open-source models that are now beating systems ten times their size. We get into what "agentic AI" actually means in production, not the buzzword version,...

Three massive forces collided this week in AI, and the fallout is just starting. First, the unprecedented standoff: Anthropic gets blacklisted by the US government for refusing to remove safety guardrails, while OpenAI steps in. Second, the money: OpenAI's record-breaking $110 billion raise. Third, the workforce: Block's explicit AI-driven layoffs and the market's enthusiastic reaction. We break d...

In six days, the performance gap between the world's top AI models collapsed to 6.9 points—and the race to build the smartest AI fundamentally changed shape. Three frontier models launched with dramatic price-performance shifts: Claude Sonnet 4.6 at one-fifth flagship cost, Gemini 3.1 Pro doubling reasoning performance, and Qwen 3.5 open-sourcing near-parity capabilities. Meanwhile, Meta and NVIDI...

This week, Anthropic closed a $30 billion funding round at a $380 billion valuation while DXC Technology deployed autonomous agents to 115,000 employees. OpenAI shipped its first non-Nvidia model on Cerebras hardware. And across the industry, $660 billion in infrastructure spending signaled that we're done with pilot projects. The "prompting fallacy" is dead. We explain why multi-agent architectur...

This week, AI stopped being an oracle you consult and became a colleague you delegate to. We're breaking down the 'agentic shift', the architectural change that lets AI manage code repositories, negotiate contracts, and run for days without constant prompting. You'll learn why the Model Context Protocol (MCP) is becoming the 'USB-C for AI tools,' how Claude Opus 4.6 and GPT-5.3-Codex are transform...