86% of What Coding Agents Do Is Just Reading — Not Solving | Alexander Whedon of Subquadratic

In transformer models, the cost of attention grows with the square of the context length. So doubling context makes the compute roughly four times more expensive.

It goes beyond the title with direct discussion of context, like, model, including: So doubling context makes the compute roughly four times more expensive.

1. 01In transformer models, the cost of attention grows with the square of the context length
2. 02So doubling context makes the compute roughly four times more expensive
3. 03When we've done benchmarking the frontier models with Superbench Pro, we found that 86% of the steps were actually read steps, just trying to do that context engineering before exe