Signals

// signal vs noise

Writing from the build, not about it. First person, operator-meets-engineer: the business outcome next to the technical fact. No SEO filler, no hype.

Read the Signals →

What a reader gets

  • Building with AI, in practice. Agents, MCP servers, and local LLMs: how they actually ship, what broke on the way, and what the fix cost. The source material is real projects: when a post cites code, it points at one of 165+ repos on GitHub (RandomSynergy17).
  • The method, with receipts. The studio runs on one loop (build once, prove cheaply, graduate what works), and posts here show it operating on real work: studio products, plus advising teams like Levels AV on AI integration and innovation.
  • Three technology eras, one perspective. I worked the last two cycles from the inside: content management, distribution, and web multimedia 20 years ago; hospitality tech (gamification, CRM, shift management) ~10 years ago. Now AI. Posts here read this era against the two before it.
  • Studio notes and UAE context. The view from the ground: I run funding research and go-to-market for Abu Dhabi, UAE, and international programmes (Fortune 500 clients and small teams alike), and some of what that teaches ends up here.

Cadence

Honest answer: there is no content calendar. A post ships when the work produces signal: something learned, measured, or broken in an instructive way. If it’s quiet here, the building was louder than the writing.

Latest

  • Brain-first: the memory every agent checks before it answers

    Agent sessions start from zero and re-derive the same decisions endlessly. GBrain is the self-hosted memory server every agent I run checks before it answers, and writes back to after it works.

  • Three eras, one method

    Why RandomSynergy isn’t an AI studio, and what two prior technology cycles taught me about this one.

  • One week, 1,189 tests: building The Foundry cockpit

    My portfolio intelligence kept freezing in static handover docs, so I built a system that keeps it live: profiled, scored, and routed to a decision. It took one week and 1,189 tests.

  • Deterministic first, LLM second

    The reliable way to ship an LLM feature is to make the model the last layer, not the first. Here’s how MeetKlay’s diagnosis is built.

  • Playwright can’t tap: giving an agent a real iOS device

    CI runs Playwright, which renders Chromium, not the WebKit an iPhone actually uses. iPhone-only bugs never fail where you’d catch them, so I built an engine that drives the real device.

  • Generation belongs in an MCP server

    An agent shouldn’t break flow to open a design tool when it needs an image. Generation should be a service any agent can call, which is what I built RNSNB to be.

Subscribe for new posts →