2026

13 posts

OpenAI is testing pay-per-outcome: what changes when evals become billing infrastructure

ChatGPT, Claude and Grok went down on the same day: correlated failure is now the default…

Record on Terminal-Bench, catastrophic on a private eval: Fable 5.1 and the case for your…

Why AI gateways should be built in Go: a language audit of every major gateway

The supply side is exploding: 136 providers, 502 models, and 2,084 client apps in a…

When AI works: Monday is its busiest day, weekends run at a third, and only 1 in 4…

The agentic long tail: 68% of AI runs are one and done, but the deepest chained 160,297…

The context window is the new battleground: we now feed AI 19 tokens for every 1 it…

LLM Observability in Production: The Metrics That Actually Matter

What the gateway saw in April 2026: agents live on Anthropic, open-source models got…

New: spend alerts for LLM traffic, webhooks when budgets get hit

Label your API keys: the cost-attribution trick most teams miss

Closing the loop: how to turn user feedback into a routing signal

2025

3 posts