Listen to the original at Lenny’s Podcast ↗lennysnewsletter.com
By Lenny Rachitsky · with Dhanji R. Prasanna · lennysnewsletter.com · @lennysan on X · YouTube · LinkedIn
Dhanji R. Prasanna is Chief Technology Officer at Block, overseeing a team of over 3,500 people and leading the company's push to become one of the most AI-native large enterprises. The conversation covers Block's open-source AI agent Goose, the measured productivity gains from AI across teams, the org restructuring from GM to functional structure, and broader lessons on engineering culture, product failure, and leadership.
01Key learnings
- Track AI impact with self-reported time savings plus validation metrics such as PR throughput and feature delivery, which Block reports as roughly 8-10 hours saved per week for AI-forward engineers and 20-25% of manual hours across the company.
- Expect the current AI value baseline to keep rising; adopt tools continuously and re-evaluate where they add value as capabilities improve.
- Non-technical staff building their own small internal tools with agents can compress weeks of waiting on engineering queues into hours, a productivity gain engineering leaders often overlook.
- Restructuring from GM-led business units to a functional engineering organization under one head of engineering made it possible to share platforms, align technical strategy, and build depth; per Conway's Law, org structure shapes what you ship.
- Have leaders use the AI tools daily themselves, as Block's executive team does, so they understand real strengths and limits before asking teams to change workflows.
- Hire for a learning mindset and openness to AI tools rather than existing AI expertise, and consider AI-assisted interview exercises.
- Question base assumptions, such as whether a process or tool needs to exist at all, before automating or building software, and avoid building in-house what distracts from your core purpose.
- Start small with experiments that can grow, and keep a learning-first mindset in which you judge experiments by what they teach and whether you gave them your best shot.
“This is the worst it will ever be. This is now the baseline.”Dhanji R. Prasanna · Lenny’s Podcast · 00:00:27
“All these LLMs are sitting idle overnight and on weekends, while humans aren't there. There's no need for that.”Dhanji R. Prasanna · Lenny’s Podcast · 01:01:12
“A lot of engineers think that code quality is important to building a successful product. The two have nothing to do with each other.”Dhanji R. Prasanna · Lenny’s Podcast · 01:01:53
02Frameworks mentioned
Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.