#llm
Every post tagged "llm".
20 posts · showing 11–20
-
What "Agents Got Better" Would Actually Look Like in Your System
Waiting for agents to improve without deciding what improvement looks like makes the waiting unfalsifiable. Write the threshold down first, and every release becomes a measurement instead of a vibe.
-
Why Benchmarks Keep Rising While Deployments Stay Hard
Both reports are accurate. Benchmarks test well-posed, self-contained, checkable problems — and almost none of deployment difficulty is any of those things.
-
Why Turn 20 Costs More Than Turn 2
A six-turn run whose final context is 12,000 tokens sends 42,000. That gap is agent economics — and it makes an early oversized tool result the most expensive thing in your run.
-
Local Models for Agent Work: Where They Actually Fit
"Can a local model replace the frontier one?" is the wrong question. Which steps can move is the right one — and the comparison people skip is against a small hosted model, where most of the saving already lives.
-
Tool Calls or Structured Output? Two Mechanisms, Two Jobs
Both emit schema-conforming JSON, which makes them look interchangeable. One expects a result back and one ends the turn — and that single difference decides every other question.
-
Why Two Identical Runs Give Different Answers
Same input, different result. Four things cause it, only one is a bug — and the most common culprit is that the input wasn't actually identical.
-
Migrating Model Versions Without Breaking Your Agent
The new model is better at nearly everything and different at everything, and tool selection is what shifts first. Replay a frozen eval set against both — and put the cost-per-task row in the comparison.
-
Model Routing: Using the Cheap Model for the Cheap Steps
Classification, extraction, and formatting move to a cheaper tier safely. Tool selection and planning don't — and the failure looks like a tool-description bug, so you'll debug the wrong thing for a week.
-
The LLM Judge That Agreed With Everything
A suite scored 94% and 93% on two versions — but the judge would have passed almost anything. Four biases quietly manufacture those numbers, and one prompt change fixes more than the rest combined.
-
Your Agent Isn't Thinking, It's Looping — Read One Turn of the Trace
There's no reasoning engine and no memory — just a while-loop, a stateless model, and an array that grows. Unroll one turn of it and most agent mysteries stop being mysteries.