Dhanji R. Prasanna

Leads one of the most AI-native large enterprises in the world. Under his leadership, Block deployed an internal open-source AI agent called Goose that saves employees an average of 8-10 hours per week. Previously built Google Wave and Google Plus. Their core argument: AI isn't slowing hiring — the biggest productivity gains come from non-engineers using AI, not from replacing engineers. The evidence is specific: Block's internal AI agent Goose saves employees an average of 8-10 hours per week. Furthermore, an engineer's Goose instance watches Slack conversations and autonomously builds features and opens PRs. Non-technical employees in legal, risk, and ops are showing the most productivity impact from AI tools. In Dhanji R. Prasanna's own words: "What's been surprising and really amazing, the non-technical people using AI agents and programming tools to build things, the people that are able to embrace it to optimize for their particular workday and their particular set of tasks are really showing the most impact from these tools." (On which employees benefit most from AI tools — surprisingly, not engineers.)

Led Block's transformation to one of the most AI-native large companies, previously built Google Wave and worked on Google+ Their core argument: Block moved FROM GM to functional to enable AI transformation -- Conway's Law means your org structure determines what you build. The evidence is specific: Square and Cash App had mirrored corporate structures including separate compliance, communications, marketing teams, and even separate offices. Furthermore, block's internal open-source agent Goose saves employees 8-10 hours per week -- enabled by shared platform that GM walls would have prevented. In the functional model, Block thinks about areas of optimization and modularity rather than headcount per feature. In Dhanji R. Prasanna's own words: "We went from a GM structure to a functional org structure, which was I think the key to making our transformation into being more of an AI-native company." (Why Block moved from GM to functional.)

Wrote the 'AI manifesto' that convinced Jack Dorsey to go all-in on AI; was promoted from part-time contributor to CTO. Under his leadership, Block became one of the most AI-native large enterprises in the world, building the open-source agent Goose that is now saving employees 8-10 hours per week. Their core argument: Go functional — engineers report to engineering leaders, not GMs — and measure real productivity. The evidence is specific: Goose, Block's internal open-source AI agent, monitors Slack conversations and proactively opens PRs for features being discussed — an engineer finds PRs waiting that Goose built overnight. Furthermore, 8-10 hours saved per week for AI-forward engineering teams, and that number is described as 'the baseline' that will only improve. Non-technical employees using Goose are showing the most impact — contradicting the assumption that AI tools primarily benefit engineers. In Dhanji R. Prasanna's own words: "This is what Jobs did when he came back to Apple as well. He reorganized Apple to be functional, and it wasn't like we were following a playbook. We discovered this as we were investigating what it's going to take to make these teams more tech-focused and to bring our DNA back to our roots, which really was putting engineering and design first, which is what technology first means to me." (Explaining the Apple parallel to Block's functional reorganization.)

Dhanji runs engineering for one of the most aggressively AI-native large companies in the world — Block ships its own internal agent, Goose, that saves employees 8-10 hours per week on average. So his argument for human taste as a moat isn't a designer defending turf; it's a CTO who has seen what happens when you let LLMs free-run. His view: in the current transition phase, agents are roughly 60% successful on well-described tasks and go off-script on the remaining 40%. The failure mode he keeps circling is drift — agents that eagerly build the wrong thing, InfoSec teams twisting themselves into knots automating processes that shouldn't exist at all. Taste is what prevents this, and it extends from the negative case (don't ship slop) to the positive one (build something people actually love). In Dhanji's own words: "I do think we're going to need a lot of human taste to anchor these AIs so they don't go off script to be honest. And that's really where our design lead and our design teams are pushing us to think, and that's a differentiator that I think will push us beyond this era of AI slop that everyone's talking about." (Explaining why Block leaning harder into agents makes taste more load-bearing, not less.)

Prasanna offers the most aggressive contrarian position in the archive. His core move is to argue that the anti-rewrite rule was never a law of software; it was a consequence of economics. Historically, rewriting was expensive because humans were expensive and slow, so 'don't rewrite' was a rational heuristic. But AI agents collapse both the cost of generation and re-derivation to near zero. He's clear this is directional rather than currently possible, but he is pushing his teams at Block to live inside that assumption — what would our world look like if every single release, rm -rf deleted the entire app and rebuilt it from scratch? Specifications and tests become the durable artifact; the code becomes ephemeral output, regenerated each cycle. Undocumented tacit knowledge in a legacy codebase becomes a liability, not a defense. In Dhanji's own words: "I think that you're going to see instead of us, for example, refactoring an app to have a different UI or to evolve into its new version, we're just going to rewrite that app from scratch. And one of the things I'm really pushing our teams to think about is what would our world look like if every single release, RM minus RF deleted the entire app and rebuilt it from scratch?" (Articulating his contrarian position that AI inverts the classical cost calculus of rewrites.)