Posts
Independent essays on AI-native work patterns, agent infrastructure, and what actually works when running agents as real tools.
-
The Agentic DevSecOps Stack Is Taking Shape
Software delivery is reorganizing around specialized coding, security, release, and operations agents. The missing layer is the evidence and governance connecting them.
-
Grok 4.6 Is on Amazon Bedrock. Here Is How to Wire Grok Build to It
SpaceXAI's Grok 4.6 is now on Amazon Bedrock. Connect Grok Build directly with an AWS profile, a short-lived credential helper, and three switchable model routes.
-
Benchmark saturation is real. The price collapse is an illusion.
Industry claims models are plateauing and prices are crashing. One is true but misframed. The other is a measurement trap.
-
Dictation and Meeting Notes, Fully Local: Rebuilding Wispr Flow and Granola Without the Cloud
I built VoiceVault, an open-source macOS app that replicates the Wispr Flow dictation experience and the Granola meeting-notes experience with everything — audio, transcription, and summarization — running on-device.
-
Building Voice Agents with LiveKit, Cartesia, and AWS
A phone agent becomes useful only when its media, model, action, security, retention, and teardown boundaries work together.
-
Exploring the Voice AI Stack
A practical map of the voice AI stack, from real-time transport and speech models to orchestration, telephony, and operations.
-
Building an Autonomous AI News Agent on Step Functions and AgentCore
A daily pipeline that fetches 24 RSS feeds, dedups across sources, categorizes into nine fixed buckets, tracks story threads across days, and writes a digest with no human in the loop. What broke, what the real timeout ceiling turned out to be, and why the shape changed twice before it worked.
-
MCP Became Stateless: Migrating a Real AgentCore Gateway
MCP 2026-07-28 removes protocol sessions from remote calls. Here is what changed in a working AgentCore WebSearch gateway, what stayed untouched, and how to migrate without breaking older clients.