(Writing)
Deep dives & technical insights.
Exploring the intersections of AI infrastructure, machine learning ethics, and the future of digital provenance.
- 6 deep dives
- ~100 min total
- Field notes
(All deep dives)
06-
Before You Build an AI Agent
An agent is one layer of a production AI system. Here is the stack underneath it and the engineering that decides whether it survives production.
-
How Your Agents Communicate
The five multi-agent communication patterns, the four orchestration topologies, and the questions that reveal which one you actually built. Every build ran — the layer that decides whether they hold together is the one the room never tested.
-
Meet Eve
Vercel's filesystem-first agent framework, where every capability is derived from where a file lives. Durable by default, composable by design, and built to make production AI agents feel as simple as building a web application.
-
Agents Harnessing
AI cannot replace thinking. It can only amplify the thinking you have done or the lack of thinking you have done
-
The $20 AI Blueprint: Ship Real Products This Weekend
Microsoft's AI costs now exceed its payroll. You don't need their budget — you need the right workflow.
-
The Race to Watermark Reality
OpenAI borrowed Google's secret weapon to detect AI images. Here's exactly how C2PA and SynthID work and where they still break. Proving an image is real is becoming the hardest problem in AI.
(The list)
Get deep dives every week.
AI infrastructure, machine learning ethics, and digital provenance notes from my latest research and builds.