> Markdown alternate of https://steerscale.com/blog

Blog

# Notes on AI in production, distributed systems and reliability.

Written by Barnabás Szőke from production work. Experiments, failure analysis and methodology — no listicles.

[All](https://steerscale.com/blog) [AI Systems](https://steerscale.com/blog#ai-systems)[AI Economics](https://steerscale.com/blog#ai-economics)[Reliability](https://steerscale.com/blog#reliability)[Evaluation](https://steerscale.com/blog#evaluation)[Distributed Systems](https://steerscale.com/blog#distributed-systems)[Performance](https://steerscale.com/blog#performance)

RELIABILITY · AI ECONOMICS 12 Sep 2026 9 min read

## [The hidden cost of retries in agentic AI systems](https://steerscale.com/blog/the-hidden-cost-of-retries-in-agentic-ai-systems)

How provider, application, tool and workflow retries combine to increase both latency and cost — and how to put a budget on them.

Barnabás Szőke · Senior Software Architect

\[Figure: Retry amplification: one request, twelve downstream calls\]

AI ECONOMICS 21 Aug 2026 7 min read

## Stop measuring AI cost per token

Why cost per successful workflow is a more useful production metric, and what it exposes that provider invoices hide.

Barnabás Szőke · Senior Software Architect

\[Figure: Cost per token versus cost per successful workflow\]

EVALUATION 30 Jul 2026 11 min read

## How to decide whether a cheaper model can replace your current model

A practical process using representative workloads, evaluation thresholds, latency and economics — with the gates that should block a release.

Barnabás Szőke · Senior Software Architect

\[Figure: Release gate: candidate passes quality, cost and latency thresholds\]

## Running one of these problems in production?

Retry behaviour, inference economics and model evaluation are part of the [AI Systems Audit](https://steerscale.com/services/ai-systems-audit).

[See the audit](https://steerscale.com/services/ai-systems-audit)