Software Tomorrow
summary
Built a web series from ~100 hours filmed and 60 hours of interviews with 28 engineering leaders at eight companies and extracts five concrete fixes: invest in specs, reduce tool sprawl, design for decoupling, codify knowledge sharing, and run continuous feedback. Claims an agentic stack (LLM + tools, memory, planning) already boosts code-task performance on HumanEval-style problems so developers shift toward specifying and guiding AI agents while keeping final ownership and last-mile fixes.
AI-written summary · about
chapters (15)
- 0:00🚀 Introduction: Jonathan Harel discusses his background and the future of software
- 3:00The story of building a web series on software development and AI
- 7:00Exploring the frustrations in software development and how AI can help
- 11:00The metaphor of bicycles: How software development is like riding a fixie bike
- 13:20Dan's idea to rewrite code from scratch and the decision-making process
- 17:00Code as an organic entity: Why software requires constant maintenance and care
- 20:11The balance between building new features and maintaining old code
- 23:14Lessons learned from producing the web series and interviews with R&D leaders
- 25:50The importance of specifications in software development
- 28:30Tooling paradox: The challenges of too many tools in the development process
- 32:56The need for decoupling in software architecture
- 36:10Predictions for the future of software development and new skills for developers
- 37:50Shifting developer roles: From coding to problem-solving and guidance
- 38:20The future of IDEs and the evolution of developer tools with AI
- 40:17AI's impact on the software development lifecycle: Planning, testing, coding, and deployment
related talks
42mThe Future of Software Development
53mThe Human Side of AI: Unlocking Opportunity in the New Bottlenecks of Software Development
30mBuilding the systems that build the software
14mFull-Stack AI Engineering from AI-Native Coding to Knowledge, Models, Infrastructure & OSS
46mPanel Discussion: The Agentic Software Factory That Enterprises Actually Need
50mAgile Intelligence: Co-Engineering the Future, Not Just Automating It
24mBeyond the Commit: A Fireside Chat on AI and Developer Productivity
27mHuman-first leadership, AI-powered
