Jeff Dean on AI Trends: Long-Running Agents, Inference Hardware, and Context Engineering
Y Combinator · youtube · 2026-07-31
In a deep interview at YC Startup School, Google Chief Scientist Jeff Dean reflects on historical decisions like fitting the search index into RAM and the napkin math that led to TPUs.
He outlines future AI trends, emphasizing:
- Long-Running Agents: Future AI agents will run for weeks, though long-context tasks currently still struggle with failures.
- Inference Hardware: As models scale, inference compute will drive the next wave of hardware specialization; AI is fundamentally an energy problem.
- Context Engineering: Described as the next frontier for AI engineering, alongside models that continuously improve their own optimization.
- Startup Opportunities: Small teams of 2-3 people can still win by questioning their biggest assumptions and building something that truly matters.
More from AGI Musings
- First-ever paper on multi-agent swarm safety lands NeurIPS spotlight — LauraRuis · 2026-10-16
- Yoav Artzi on peer review: paper flooding still pays off as review system strains under load — yoavartzi · 2026-10-16
- Letta founder: work in fields defined by problems, not methods — lateinteraction · 2026-10-16
- Why pure mathematicians will soon flood into AI x-risk debates — Dr_Atoosa · 2026-10-16
- RSI began when model outputs augmented next-gen training data, agents just escalated it — akbirthko · 2026-10-16
- Luiza Jarovsky: with AI making careers unpredictable, 'do what you love' is the best advice left — LuizaJarovsky · 2026-10-16