Field notes

Rob Burke

Latest side quest: Oct 2026

AI agents, side quests, and 25 years of synthetic characters

The Story To Date…

A 30+-year journey from synthetic characters to the AI characters I build now, year by year.

I’ve been building brains for synthetic characters since before we called them “agentic.” Here’s the short version, oldest first.

  • 1995–1999 · Queen’s University. Fifty of us built a student game, and I wrote the autonomous AI for its characters. That game is what sent me to MIT.
  • 1999–2001 · MIT Media Lab, Synthetic Characters group. Grad school under Bruce Blumberg. My thesis was on temporal representations: giving a virtual creature a sense of time.
  • 2002–2004 · MIT Media Lab Europe, Dublin. Research Associate in the MindGames group; architect and lead developer of Symphony, for biometric interfaces (EEG, ECG and other). Reading a person’s state and responding to it.
  • 2004–2007 · Microsoft, Dublin. Technical evangelist (remember when that was a thing?!).
  • 2007–2009 · Carrington Technologies, Toronto. My own small consultancy: financial institutions, legal practices, and Microsoft’s National Productivity Centre in Ireland.
  • 2009–2017 · Infusion (now part of Avanade), Toronto. Team Lead, then Account Manager. Emerging tech for financial services and beyond.
  • 2017–present · EPAM. Client Partner and Head of Public Sector Accounts in Canada. The day job is building teams of people and AI agents and then actually building with them.
  • 2026 · The field came back around. Ziggy, our family’s AI, runs a five-layer memory architecture with a temporal layer.

Things I made

  • Legend of the Greasepole (1998) – 50 Queen’s students built it. AI-powered and inexplicable. Later resurrected by another AI.
  • It’s about Time (2001) – my MIT thesis: temporal representations for synthetic characters.
  • Symphony (2002–2004) – signal processing and visualization framework behind 10+ projects; one went to NASA.
  • Still Life (2003) – physiotherapy turned performance art, performed at the UN in Geneva.
  • Mind Balance (2004) – control a video game with your brainwaves. Presented at Microsoft Research Cambridge in 2006.
  • Ireland’s Call for Guitar Hero (2007) – the “oh, and one last thing…” from Rob’s Last Stand.
  • Star Trek and Chill (2016) – HoloLens + biometrics recreate Star Trek TNG’s mind-control game.
  • Galton’s Goalie (2026) – computer vision over Plinko math, inspired by Mark Rober’s CrunchLabs.
  • Withings → Garmin Sync (2026) – my first AWS Lambda and I still use it every morning.
  • Minecraft Build Bot (2026) – L-systems, Lorenz attractors, and a G1 Ark.
  • Tom (2026) – learns my photo-editing taste: 2,500 frames in, 24 bangers out.
  • Garmin Edge 1000, modern graphs (2026) – a bike computer from the drawer, hung off a USB and brought back from the dead.
  • Japan by Rail (2026) – field notebook souvenir from 18 days in Japan, with the trip recipe on GitHub.

The same question runs through all of it, from a virtual terrier in 1999 to the characters I build now: what makes a synthetic character feel present? What does it remember, what does it expect, and what should it do next? The tools are unrecognizably better, but the questions are uncannily aligned to the ones Bruce challenged us to ask.

There’s one more question stacked on top now, which I think is the most pressing, and certainly the one I care the most about: when the AI can “do the doing,” what are the people for? Their judgment relocates – into specifying the work, validating it, teaching it, and being across the table when it matters.