Listen on Spotify, Apple Podcasts, or watch on Youtube.
Timestamps
0:00 Opening thoughts on intellectual laziness with AI
1:11 Jonathan introduces himself and his career background
2:50 Jonathan's last day at DeepMind; it's also his birthday
3:30 Ben asks about DeepMind's roadmap and its impact on work and learning
4:48 Layers of AI: model, harness, agents, and learning-specific applications
6:47 Model commoditization and Karpathy's tweet about being behind on tools
8:07 DeepMind's advantage: using AI to improve the model itself
10:43 How the ML engineer role has changed with LLMs
11:55 Warning about intellectual laziness and work lacking substance
12:20 The founder's question: avoiding the buzzsaw of large language models
16:40 Neo labs, the DeepMind mafia, and betting on AI approaches
21:18 Closing: AI models are commodities, but deep expertise is not
What does someone who spent five years at Google DeepMind really think about the future of AI?
Fresh from his last day at Google DeepMind, Jonathan Evans joins Ben, Brighteye's Managing Partner, for a candid conversation about how AI is changing the way we work, build products and create value.
From the rise of AI agents to the changing role of engineers, Jonathan shares what he's learnt from working at the frontier of AI, and where he believes the biggest opportunities (and risks) lie ahead.
A few questions we explore:
- Is AI making us more productive, or just intellectually lazy?
- As AI models become commodities, where can startups still win?
- How is AI changing engineering roles, and which skills will become more valuable?
- Why are so many DeepMind researchers leaving to build their own companies?
- What really gives startups an advantage when the AI giants keep getting better?
The big takeaway? As AI becomes more accessible to everyone, deep industry expertise, proprietary data and a real understanding of customer workflows could become more valuable than ever.
A conversation packed with insights for founders building in the age of AI.
