Once outcasts, three Canadian researchers now lead the deep learning efforts at Google, Facebook and IBM
Welcome to the AI Conspiracy: The ‘Canadian Mafia’ Behind Tech's Latest Craze — In the late '90s, Tomi Poutanen, a precocious computer whiz from Finland, hoped to do his dissertation …
Context & Ripple Effects
In the late 1990s the neural-network researchers profiled here were academic outsiders; this Re/code piece marks the moment they became establishment, with three Canadians heading deep learning at Google, Facebook and IBM simultaneously. The related coverage shows how fast that reversal compounded: Google put $3.4M into the Montreal Institute for Learning Algorithms and opened its own Montreal deep learning group within a year, then joined Ottawa and 29 other firms on a $128M Toronto-based AI institute under the pan-Canadian strategy.
The state's role became explicit in later interviews with PM Justin Trudeau about how the government helped assemble the research community, and the talent pull ran both directions — DeepMind later hired prominent Canadian computer scientists for an Edmonton lab, its first outside the UK. The lineage matters beyond Canada: eight former Google scientists went on to co-author "Attention Is All You Need", the paper that set up today's LLMs.
First-order effects
- Google, Facebook and IBM each anchor their deep learning efforts on one of these researchers, making individual academics the scarce asset the largest tech companies compete over.
Second-order effects
- Canada responds to the brain-drain risk with public-private money — the $128M Toronto institute backed by Google and 29 firms, plus Google's Montreal investment — while DeepMind plants its first non-UK lab in Edmonton to recruit from the same pool.
Third-order effects
- If the pattern holds, frontier AI consolidates around a small academic lineage absorbed into corporate labs — the same cohort's students produce "Attention Is All You Need," and national governments treat retaining a handful of researchers as industrial policy.
The trend: Frontier AI is institutionalizing: once-marginalized academic researchers become the core assets of corporate and national AI strategies.