Everything we've published.
10 articles across 6 sections. Sorted newest first.
Anthropic bets the next platform war is won in the agent runtime, not the model
MCP adoption crossed a threshold this quarter and the Claude Agent SDK quietly became the default scaffold for production agents. The model is no longer the moat — the runtime is.
2 min read
The verifier is eating the generator: why test-time compute keeps winning on arXiv
A cluster of recent papers points the same direction: a smaller model that checks its own work beats a bigger model that doesn't. The economics of inference are quietly being rewritten.
2 min read
Follow the money: AI funding is quietly rotating out of models and into the application layer
The mega-rounds are still going to foundation labs, but the velocity has moved. Vertical, workflow-native AI companies are raising faster and at saner valuations. The application layer is where 2026 is being decided.
2 min read
Your agent isn't unreliable — your evals are. Notes on the real production bottleneck
The teams shipping agents that actually work in production share one habit, and it isn't prompt engineering. It's treating evaluation as a first-class engineering discipline with the same rigor as tests.
2 min read
The small-model thesis finally pays off: where on-device AI actually makes sense in 2026
Small models stopped being a compromise and started being a strategy. The economics, latency, and privacy math now favor running locally for a surprising share of real workloads.
2 min read
Claude Opus 4.7 lands quietly, raises the floor for agentic coding
Anthropic released Claude Opus 4.7 with little fanfare. The benchmarks are predictable. The vibes, less so.
1 min read
Postgres 18 made async I/O the default and nobody noticed
The most consequential database release of the year shipped without a keynote. Async I/O is now on by default and your read-heavy workloads just got 40% faster.
1 min read
OpenAI's enterprise pivot is more interesting than its product pivot
OpenAI hired its third Chief Revenue Officer in 18 months. The product team is being reorganized around verticals. The plot has thickened.
1 min read
Async traits in Rust are finally stable. We are going to regret this.
Eight years after the first RFC, async fn in traits is stable. The ecosystem is about to learn what the lang team has been worried about.
2 min read
Edge runtimes are out. Fluid compute is in. The post-mortem nobody asked for.
Vercel quietly stopped recommending Edge Functions. The platforms aren't telling you why. We'll explain it.
2 min read