
Putting the best AI into the hands of the best creators to tell otherwise impossible stories.
I’m a Staff Research Scientist at Google DeepMind, studying how humans and machines create together. I launched the Veo and Gemini Omni video models and the Flow filmmaking tool, and was Technical Director on ANCESTRA, which premiered at Tribeca.
Writing #
Start here
- Building an Artificial Improvisor 2016
- The Cherry Pick Ratio 2024
- What is it like to be an AGI? 2022
- How Competition Defined Canadian Improvisation 2018
Recent
Why Literacy Alone Won't Solve Canada's AI Problem
Why don’t Canadians use AI?
What role should we hire an AI for?
Worn In: Birthday Asks for 38 and 39
The Case Against Taste
Unlock More Creativity
Creative Tools On Demand
The Danger of Nothing
Examining the Cherry Pick Ratio
Teaching The Disruption
Theatre & Improv #
Canadian Comedy Award-winning improvisor with Rapid Fire Theatre, Improbotics, and Die Nasty. Created the first AI to perform live improv.
Productions: Improvised TED Talks (with Julian Faid) • The Professors & Plays by Bots (with Joe Vanderhelm) • Improbotics & HumanMachine (with Piotr Mirowski)
Research #

2026
Launched Gemini Omni, a model that creates and edits video from any mix of text, images, audio and video. Flow grew into a unified creative workspace combining video generation, image creation, and audio, powered by Veo 3.1, Imagen 4, and Lyria 3. Wrote about delegating creative drudgework to AI agents and the role of human attention in the creative process.

2025
Technical Director for “ANCESTRA”, which premiered at the 2025 Tribeca Film Festival. Published a Google Keyword blog post, “ANCESTRA”: Behind the scenes of our experimental film and accompanying behind-the-scenes video on the making of. Launched Flow: an AI filmmaking tool built with and for creatives.

2024
Launched Google Veo for video generation + VideoFX consumer tool + led our collaboration with professional filmmakers (including collaboration on filmmaking with Donald Glover and his creative studio, Gilga). Published on measuring divergent creativity in humans and LLMs and an evaluation of LLMs’ humour alignment with comedians.

2023
Launched Dramatron for professionals co-writing screenplays and scripts with language models (code, paper, patent) and published on cultural general intelligence and communicative capital.

2022
Published a guide to designing and evaluating human-centred interactive machine learning, research in emergent communication in populations and at scale, and BIG-bench for evaluating the capabilities of language models.

2021
Launched StoryCentaur for using large language models as creative writing tools (code, paper), published research on collaborative storytelling with humans and AIs, tele-immersive improvisational performance, generating diverse and realistic laughter for interactive art, and symbolic behaviour in artificial intelligence.

2020
Published research on shaping narrative arcs via information-theory and collaborative dialogue, on imitating interactive intelligence, and on using AI to perform improv in any language.

2019
Published Ph.D. thesis Humour-in-the-loop: Improvised Theatre with Interactive Machine Learning Systems and launched TalkGenerator.com for automatically generating engaging presentation slide decks.

2018
Launched Improbotics, a theatre lab pioneering performance alongside generative AI.

2017

2016
Built and performed with an Artificial Improvisor – a chatbot and projected puppet to perform alongside.