Summary
Jeremy Howard, founding CEO of Answer.ai and co-founder of Fast.ai, opens Day 2 of the AI Engineer Melbourne 2026 conference with a keynote that reframes AI not as a disruptive threat but as the latest chapter in a decades-long human pursuit: building richer, more expressive connections between people and computers. Howard — whose ULMFiT paper is widely credited as foundational to modern large language models — delivers a talk that is equal parts psychological research, technical history, and professional philosophy.
Drawing on self-determination theory (SDT), a body of research spanning 30 years and hundreds of randomized controlled trials, Howard argues that human flourishing requires eudaimonia — effortful mastery and full actualization of one’s capacities — rather than passive ease. He traces a lineage of builders who pursued this same vision through computing: Ken Iverson’s APL notation (which enabled entirely new classes of mathematical proofs through expressive syntax), Bret Victor’s interactive programming environments, and Chris Lattner’s progression from LLVM through Clang, Swift Playgrounds, and Mojo.
Howard’s central argument is that AI tools, when designed with craft in mind, continue this tradition rather than ending it. The talk serves as a grounding corrective to both uncritical AI hype and reflexive fear, urging AI engineers to choose work that builds genuine mastery. It is particularly valuable for practitioners questioning professional identity in an era of rapid automation, and for anyone interested in the philosophical lineage connecting symbolic computing to today’s language model ecosystem.
📺 Source: AI Engineer · Published June 04, 2026
🏷️ Format: Keynote Launch
