Retrieval Is Eating Search: A Look Inside structured outputs
Notes from the teams shipping fine-tuning to real users.
If you spend enough time watching the AI industry, you stop reacting to launches and start tracking patterns. Retrieval Is Eating Search is one of those patterns.
Teams that win with tool-first agents tend to share a habit: they write the evals before they write the prompts. Everything else follows from that.
Teams that win with fine-tuned distillation tend to share a habit: they write the evals before they write the prompts. Everything else follows from that.
What Anthropic actually shipped with Llama 4 is less a single capability and more a cluster of small, compounding improvements — the kind that only show up when you put a real workflow on top.
What Google DeepMind actually shipped with Gemini 3 Pro is less a single capability and more a cluster of small, compounding improvements — the kind that only show up when you put a real workflow on top.
Ramp has been quietly running research synthesis through Zed for months. The results are unglamorous and, for that reason, more interesting than another benchmark chart.
Eval harnesses, once an afterthought, are becoming the most important piece of code in many AI projects. HubSpot's team treats theirs the way an SRE team treats a runbook.
The cost curve matters here. Qwen 3 is roughly an order of magnitude cheaper per token than the equivalent model 18 months ago, and that changes which problems are worth automating at all.
None of this guarantees a clean story. xAI could ship a model next month that rearranges the assumptions in this piece. But the direction of travel, for now, is clear enough to plan around.