Oct 26, 2025 · 1h 26m · lennys-podcast
How Block is becoming the most AI-native enterprise in the world | Dhanji R. Prasanna
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Block CTO Dhanji R. Prasanna reveals how Block became an AI-native enterprise through the development of their open-source agent Goose and a radical functional reorganization. He shares practical insights on autonomous developer workflows, Conway's Law, and hard-earned engineering lessons prioritizing customer utility over pristine code.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Lenny holds 24.9% of the talking time here. How this is scored →
Lenny as informed peer 4.3 Guest teaching 6.1 Guest disagreement 1.6 Lenny pushing back 1.1
speaking balance: gold is Lenny, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:02:10 Code quality has nothing to do with product success
Dhanji bluntly dismisses one of engineering's sacred cows, asserting that code cleanliness has no bearing on building a winning product.
Hardest push from Lenny ▶ 35:45 Pushback on rewriting codebases from scratchLenny directly challenges Dhanji's suggestion of deleting and rewriting codebases by citing the classic software engineering doctrine against ground-up rewrites.
Biggest teaching moment ▶ 34:45 Autonomous multi-hour agents replacing vibe codingDhanji educates Lenny on why short-turnaround vibe coding is obsolete, showing how agents can run multiple overnight experiments in parallel.
Lenny holds their own ▶ 11:10 Synthesizing functional restructuring principlesLenny demonstrates organizational design expertise by crisply breaking down Block's transition from GM silos into centralized functional engineering.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Lenny as informed peer | Guest teaching | Guest disagreement | Lenny pushing back | Why |
|---|---|---|---|---|---|---|
| The AI Manifesto and Taking the CTO Role | 4 | 5 | 2 | 1 | Lenny opens by inquiring about Dhanji's famous memo to Jack Dorsey and establishes background context on Dhanji's rise from senior engineer to CTO. Dhanji recounts how he urged Block leadership to centralize AI efforts. | |
| Restructuring Block: From GM Model to Functional Tech Organization | 5 | 6 | 1 | 1 | Lenny synthesizes the architectural takeaways around functional leadership versus business unit silos. Dhanji explains the transition away from independent GM units to unified technical reporting. | |
| Conway's Law and Unifying Technical Standards | 4 | 6 | 3 | 1 | Dhanji notes that Conway's Law made the initial structural reorg painful rather than seamless, correcting any assumption that everyone was immediately on board. Lenny asks how this translates to day-to-day engineering workflows. | |
| Measuring Realized AI Gains and Non-Technical Software Creation | 5 | 7 | 2 | 2 | Lenny frames the debate between AI hype and practical adoption. Dhanji provides specific numbers around hours saved and schools the audience on how non-technical departments gain the most leverage from custom automated tools. | |
| Introducing Goose and the Model Context Protocol | 4 | 7 | 1 | 1 | Lenny asks Dhanji to define Goose and explain its architectural foundation. Dhanji explains the mechanics of the Model Context Protocol (MCP) and how it grants LLMs actionable execution capabilities. | |
| Pluggable LLMs and Broad Industry Adoption of Goose | 4 | 6 | 1 | 1 | Lenny clarifies how Goose interfaces across foundational models and local setups. Dhanji outlines how Goose dynamically writes its own MCPs to orchestrate tools across disparate enterprise systems. | |
| Open Source Philosophy and the Origin of the Name Goose | 3 | 5 | 1 | 1 | Lenny expresses surprise that Block open-sourced Goose rather than commercializing it as a standalone startup. Dhanji articulates Block's foundational philosophy of building public open-source protocols. | |
| The Future of Development: Overnight Autonomy and Disposable Codebases | 4 | 8 | 3 | 2 | Dhanji presents a radical view of future software development where code is entirely disposable and whole applications are deleted and rewritten overnight by autonomous agents. Lenny highlights the contrast with traditional software engineering rules against full rewrites. | |
| Human Taste, Portfolio Judgment, and Avoiding AI Slop | 4 | 6 | 2 | 1 | Lenny asks for concrete examples of where human intervention remains indispensable. Dhanji details how humans are needed for high-level taste and questioning underlying operational assumptions rather than automating flawed processes. | |
| Build vs. Buy and Focusing on Core Competencies | 5 | 6 | 2 | 1 | Lenny raises the classic build versus buy debate in the AI era. Dhanji cautions against building internal tools that distract from core business empowerment, illustrating with past Cash Card edge cases. | |
| Sponsor Segment: Persona | 4 | 6 | 1 | 1 | Following an ad read, Lenny asks about headcount planning and hiring criteria in an AI-augmented company. Dhanji clarifies that headcount needs shift because of functional architecture rather than AI tool adoption alone. | |
| Who Benefits Most from AI and the Reorganization Hot Take | 5 | 7 | 2 | 2 | Lenny explores which engineering seniority tier benefits most from AI tooling. Dhanji argues that non-technical staff achieve the most profound leverage and reiterates that organizational reorg outperforms AI tooling for pure productivity. | |
| Optimizing Engineering Systems and Deleting Redundant Processes | 5 | 6 | 1 | 1 | Lenny cites Elon Musk's optimization principles about questioning process requirements. Dhanji provides a personal workflow example of using Goose and AppleScript to automate personal expense receipts. | |
| Goose Distribution and Desktop Extensibility | 3 | 6 | 1 | 1 | Lenny asks how accessible Goose is for general users and how it contrasts with existing developer tools. Dhanji explains its desktop Electron distribution and extensible protocol nature. | |
| Counterintuitive Wisdom: Code Quality vs. Product Success | 5 | 8 | 3 | 1 | Dhanji shares a counterintuitive lesson that code quality has virtually zero correlation with product market success, citing YouTube's early architecture compared to Google Video. Lenny validates the insight with examples of hyper-growth chaos. | |
| Cash App's Early Growth, Controlled Chaos, and the Mad Scientist Role | 4 | 6 | 1 | 1 | Dhanji explains how Cash App scaled through deliberate controlled chaos and creative autonomy for engineers. Lenny brings up Dhanji's former internal title as Block's official Mad Scientist. | |
| Starting Small and the Original Bitcoin Hackathon | 5 | 6 | 1 | 1 | Lenny asks for core leadership tenets. Dhanji emphasizes starting small, recalling the three-person hackathon team with Jack Dorsey that launched Block's first Bitcoin integration. | |
| Fail Corner: Google Wave, Google Plus, Secret, and Startups | 4 | 6 | 1 | 1 | In fail corner, Dhanji reflects on lessons from high-profile unsuccessful projects including Google Wave, Google Plus, and Secret. Lenny quips that those failed products probably had immaculate architecture. | |
| Staying Grounded in Core Values Amid Industry Hype | 4 | 5 | 1 | 1 | Dhanji offers grounded perspective for operators overwhelmed by AI industry developments, advocating for anchoring in user utility. Lenny synthesizes the core takeaways around focus. | |
| Lightning Round: Books, Culture, Hardware, and Life Mottos | 4 | 4 | 1 | 1 | In the lightning round, Dhanji shares literary recommendations including Russian classics and Tennyson poetry, praises the open hardware of the Steam Deck OLED, and shares his philosophy on career autonomy. |
Statements from this episode (30)
Block consolidated engineering and design into single functional organizations
“We S we changed the organization. So all engineers report into one single team. Now all designers report in one single team and their single head of engineering, single head of design, et cetera.”
Block's business units previously operated in complete technical silos
“We had a lot of momentum in each of these silos, be it cash app, be it after pay. Be it square or even title or music streaming service. And no one was really talking to each other. No one was really aligned on technical strategy on what we even wanted to be f…”
Block standardized engineering tools, levels, and mobility company-wide
“We at least speak the same language. We're all have access to the same tools. We share the same policies. So like a certain level of senior engineer means the same thing across the whole company. People can move from one team to another into an area of need.”
Block uses AI tools to autonomously generate bug patches overnight
“So we have these tools that run 24 seven or run in the CI pipeline, and they're analyzing vulnerabilities. They're looking at even bugs filed on tickets and trying to build patches While engineers are asleep. So they come in and the next day and look at it.”
Block engineers using AI agent Goose report saving 8-10 hours weekly
“But even now we find engineering teams that are very, very AI forward that you're using goose every day are reporting about eight to 10 hours saved per week.”
Block expects AI to save 20-25% of manual operational hours company-wide
“And we feel across the whole company, we're probably trending towards 20 to 25% of manual hours saved.”
AI falls short on architecture and race conditions compared to senior engineers
“So when you have some very senior engineers and they're thinking about things like architecture and design and race conditions, orchestration, things like this. That's still an area where AI isn't quite there.”
AI code gains are strong in greenfield projects, lagging in legacy code
“If you're building a totally new green fields code base, or you're building an app for a new platform, then we we're seeing those pretty aggressive gains, but in, you know, very complex code bases that already exist, those gains are not quite there yet.”
Block's open-source Goose agent can author its own MCP integrations
“And Goose can write its own MCPs, so it's pretty, ah, bootstrappable as well.”
Databricks and competitors are actively using Block's open-source Goose agent
“And we have a lot of companies using Goose pretty actively. I don't want to name too many names, but from our competitors to our sort of close partners, a lot of them are using Goose pretty regularly on their teams. I know Databricks talks about it a lot, but …”
Block engineer built an ambient AI setup that autonomously drafts pull requests
“And so he built this system where it's essentially just watching everything he does all the time. And he'll be talking to a colleague on Slack or an email, and they'll be discussing some feature that they think is useful to implement. And then a few hours late…”
AI orchestration eliminates the need to wait for vendor features
“So these are things that I think we were waiting for the calendar vendor to build as features into calendar, but we don't need to do that anymore because AI is able to orchestrate this for us.”
Block's AI agent Goose sees median sessions of five minutes
“Our average, our median session length is five minutes and on average seven.”
Prasanna: AI will make teams rewrite entire apps instead of refactoring
“And I think that you're gonna see instead of us, for example, refactoring a, an app to have a different UI or to evolve into its new version, we're just gonna rewrite that app from scratch.”
Block's experimental Goose agent autonomously implements 60% of requested features overnight
“And it's successful, I would say, on, like, 60% of those things if the features are sort of well enough described, and it struggles on the remaining 40 where you have to kind of intervene and massage it, yeah.”
In-house software cost savings rarely justify the lost engineering focus
“The savings and costs that there might be in replacing a vendor tool by something you build in-house is probably not worth it in the mental bandwidth that you've lost and the amount of the team's sort of technical focus that's being taken away.”
Block built Cash Card's core functionality in roughly a week
“In cash card, for example, we built the entire functionality of cash card. I would say pretty much in a weekend or maybe a week of sort of integration and work. And then it took a really long time to iron out all these edge cases”
Prasanna: AI hasn't fundamentally reduced the headcount needed for scale apps
“I don't think that things have Progress far enough that it's really impacted in a fundamental way how you would, how many people you would need to sort of build an app of the scale of cash app, for example.”
Block's headcount shifts stem from functional reorganization, not AI tools
“I think what's changed for us is much different and it has nothing to do with AI. It's that what we talked about earlier is moving from our GM structure to a functional structure.”
Deep problem understanding matters far more than AI-native coding skills
“I still think that things that we interviewed for in the past, a critical mindset, the ability to really understand deeply the technical nature of a problem. Is still much more important than whether you're a fully AI native programmer.”
AI tools will blur lines across legal, risk, engineering, and design
“And I think that speaks to how these roles are going to evolve in the future. The lines are going to be blurred between whether you're in legal or in risk or in engineering and design even.”
AI-enabled internal software creation induces greater software demand, not less
“Although I think that it's a little bit like the analogy of if you build a bigger highway, you'll just get more cars on the road. So I think the fact that everyone's building software means that there's more software to be built, more coordination to happen, a…”
Block cut test runs by 50% using a selective testing tool
“So we have this really cool tool that analyzes our test suites and selects the right tests to run for changes that were made. So we cut down basically 50% of test runs this way, which is pretty great, and like we're not warming the planet as much with all thes…”
Jack Dorsey and Block's executive team use Goose AI daily
“Jack uses goose. I use goose, our executive team all have used goose and use it regularly and use other other AI programming tools and assistance as well. And we do it every single day.”
Prasanna: Code quality has nothing to do with product success
“A lot of engineers think that code quality is important to building a successful product. The two have nothing to do with each other.”
Google Wave had 70-80 engineers before securing any external users
“I worked at Google on this product called Google wave, which was trying to be everything to everyone. And. You know, we were 70, 80 engineers building this thing before it even really had any users outside Google.”
Block built Cash App after observing unexpected peer-to-peer Square usage
“You know, I remember even in the early days, people used it to pay each other as a peer to peer money transfer system. And that's why we built Cash App and that was really successful on the back of that.”
Prasanna: Classic literature and philosophy are more valuable than business books
“You shouldn't read books that are about like your daily work or your professional life. I read fiction. I read the classics. I read poetry, philosophy, history. These are the books I really enjoy. And I think it expands your mind and gives you creative ideas a…”
Valve proved that locking down hardware user experiences is unnecessary
“And in this era where like, we're constantly told by big tech companies that we need to lock everything down. We need to lock down the user experience and customizability in order to have things work for people. I think Valve showed that's totally unnecessary …”