Skip to content

Black Friday: up to 61% off our packs until November 30.

See the packs

Blog

Blog & Insights

Practical articles on building and running AI products, agents and workflows.

All articles

    • Cloud & Infrastructure

    Inkling 2 Small Self-Hosting vs Tinker API: The Resident-vs-Active Break-Even Test

    Inkling 2 Small activates only 17B parameters per token, but self-hosting means keeping all 594B in memory. This guide sets out a step-by-step method for comparing hosted Tinker inference with running the open weights on your own GPUs, using cost per task rather than cost per token.

    • Cloud & Infrastructure

    Kubernetes Probes: When to Wait, Route Traffic, or Restart

    Startup, readiness, and liveness probes answer different operational questions. Use a decision matrix, explicit endpoint contracts, and a proposed staging failure-test plan to choose when an application should wait, leave traffic rotation, or restart.

    • Cloud & Infrastructure

    Benchmark Scores Need Protocols: How Evaluation Cards and Every Eval Ever Make AI Results Comparable

    UK AISI and EvalEval's September 22 release points to a more useful way to read AI benchmarks: join the score to the protocol before comparing models. This guide explains how Evaluation Cards, Every Eval Ever, and public inference-scaling records can support traceable comparison without pretending that schema validation makes experiments equivalent.

    • Cloud & Infrastructure

    WeatherNext 3 and the Forecast Freshness Test for Operational Decision Systems

    WeatherNext 3 brings hourly, higher resolution AI weather forecasts into surfaces that operators already use, but better weather data does not automatically create better decisions. This article introduces Optijara's Forecast Freshness and Decision Test for deciding when AI forecasts are fresh, local, reliable, and safe enough to influence operational workflows.

Next step

Have something to build?

Tell us what you need. We reply with a clear scope and a fixed quote.