Briefings on what is actually happening in AI, written for people who run businesses rather than labs: which models lead and what they cost, what the open-weight rankings say, where AI has genuinely moved science, what the adoption surveys find, and which rules are now in force. Every figure carries a footnote to a primary source. Below them, the R&D and open experiments we run ourselves.
Dated to the events they cover, so some are written after the fact. Where something changed since, the briefing says so in a dated update rather than quietly rewriting history.
Beyond client work, we partner with founders and teams we believe in to build and ship real products, where we push the techniques we then bring back to customers. Two are live today.
“Put intelligence into action.” Aeronaut is an autonomous AI system for unmanned drones, processing video, LiDAR, and thermal data on-board, with no ground link or GPS, so missions keep running in signal-denied or jammed environments. It runs as software across airframes, with mesh-networked swarm coordination, for security, inspection, and surveillance.
“Take back control of your personal data.” Soryx is a GDPR-native privacy platform that finds, monitors, and removes your exposed personal data across the surface web, dark web, and public databases, with automated removals and continuous monitoring. EU-hosted, with field-level encryption and audit logging.
An open competition to explore how unusual, long-tail, or under-represented data improves AI systems. Teams across Ireland are building entries through January–February 2026 with a €5,000 prize pool. We're documenting what works, what doesn't, and what surprises us.
We run open events where teams build AI systems under time pressure and present in public. Demo Days, NomadChat sprints, and Young AI Dublin meetups all feed the same loop, short cycles, honest critique, and people learning by shipping. We collect notes from each one and roll lessons into how we train clients.
Nomad is part of Anthropic's Claude Ambassador programme and Cursor's SMB Ambassador programme. They aren't research outputs, they're a tighter line to what changes in production, fed back into how we evaluate models, configure tools, and train teams.
The briefings report other people’s results, with the sources listed, and label our own opinion as opinion. We don’t run benchmarks of our own yet; when we do, they will be published here with the method. Subscribe to the newsletter to hear when something new lands, or read the blog for the working notes.
Book a call with the founders. Thirty minutes to understand your business, your team, and where AI could actually help. No deck, no pitch.
Talk to us →
