Great article from Fortune on routing's challenges, Not Diamond, and why every company is trying to set up intelligent model routing now. We've been working on this problem since 2023 and there are a *lot* of dead ends where routing will cost you *more* money. This quote from @cclark from OpenRouter (who's also been working on this since 2023) captures it well: “I know a lot of companies are trying this – we’ve seen some sort of launch from just about every company under the sun over the past few weeks. But I think those companies and their customers will discover this is a harder problem to get right than they anticipate.” Model routing is both a deceptively difficult problem and a constantly moving target. It's not by chance that Cursor didn't publish any performance benchmarks when they launched their router—this is a company that is extremely proficient at benchmarking. In long-horizon coding agents, the interdependent variables and downstream consequences of model selection makes model routing incredibly challenging, but also incredibly rewarding if done right. As all inference spend shifts to agentic workloads, it's more important than ever to have effective, reliable, and secure routing infrastructure for every AI product and consumer, both at the gateway level and at the intelligent recommendations layer. Check out the full article by @sharongoldman here: