The Best Model Is a Routing Decision
The best model isn't the smartest — it's whichever should get the next unit of work. Notes from NVIDIA on routing, local models, and the AI factory.
Co-founder and CTO of Speedscale, expert in Agentic AI and cloud data warehousing. • 9 posts published
The best model isn't the smartest — it's whichever should get the next unit of work. Notes from NVIDIA on routing, local models, and the AI factory.
The trace was sampled out. I found the bug anyway — by filtering recorded traffic on the customer's email instead of a trace ID. Here's how to follow one request across four services with no trace IDs and no OpenTelemetry.
Metrics, logs, and traces were built for humans and cheap storage. AI inverts both assumptions, and the next maturity level is a deterministic replay sandbox.
Logs, metrics, and traces are a lossy compression of production. Five things you can do with a traffic data lake that observability can't.
UI synthetics only tell you something is broken. Traffic replay per microservice isolates failures before any human walks up. Zero scripts required.
SaaS AI fails when agents need continuous access to your codebase and internal APIs. Here's why BYOC is the only deployment model that works at scale.
Teams spend six figures on observability but test with synthetic data. Close the gap between what you know about production and what you validate pre-release.
Speedscale launches proxymock as an OpenClaw skill on ClawHub, bringing traffic replay and production context to Claude for improved reliability.
Speedscale is a Representative Vendor in the Gartner Market Guide for API and MCP Testing Tools. See how traffic replay modernizes testing.
Choose the desktop proxymock or the hosted cloud trial to get started.