Lenny's Podcast

The episode contains genuine operational gems - using Claude Code as a management instrument across repos, the Bad/Sad quality framework, pairwise programming lunches to fight agent-induced loneliness, and JIT monthly planning - but these insights are buried in extensive personal narrative (bank teller origin story, grandma, high school), generic growth-mindset advice, and meandering tangents that dilute the density considerably.

“I actually have a cloud code remote session that I enlist in all of our repos. And so this way I have full visibility into the work that everybody's doing. And this instance also has access to all our Slack channels”

“I started this. Hey, let's have a concept of uh, what's bad versus what's sad. And bad is a very bad irrecoverable error. And sad is something that's kind of like a pain point recoverable. But it's interesting when you stack up SADs, it could generally go to bad.”

There are genuinely fresh practitioner observations - using routines to asynchronously spawn agents that deliver morning PRs, spec-driven code review as evolved TDD, and the Bad/Sad UX taxonomy - but the episode leans heavily on recycled frameworks: growth mindset, high-agency-plus-accountability, latent demand, and the fear-as-compass metaphor that circulates everywhere.

“I call it JIT planning now. Like just in time planning. So it is like around like because, yeah, I think six months was too long.”

“along with hi, because, like, we really. It's about like, hey, here's a problem... we say, with high agency, it's also high accountability.”

Fiona Fung is an exceptional practitioner guest: she founded Facebook Marketplace (now $100B+ GMV), shipped Orion AR glasses, ran a 500-person Meta org, and now directly oversees the teams building Claude Code and Cowork - she has done the thing at massive scale across multiple paradigm shifts and speaks from lived operational experience throughout.

“Facebook Marketplace team, which she took from idea to launch today. Facebook Marketplace generates over $100 billion in GMV every year”

“the last time I shipped production software at Meta was probably 2017... I remember that first week on um, cloud, I'm like, at first I almost again went, did my usual, let me go meet all the engineers”

There are solid concrete data points - the 8x code output metric, $100B Marketplace GMV, the Chile/LTE Marketplace war story, crash-rate as a 'Bad' metric example, swear-word dashboard - but the episode frequently retreats into abstraction when more detail would be most valuable, particularly around verification approaches, hiring criteria, and planning specifics.

“Anthropic engineers on average ship eight times as much code per quarter as they did compared to 21 to 20, 25”

“on cli, could be crash rates. Like a crash is pretty bad, you lost work. Um, and for example, a sav might be, hey, is it flickering?”

Lenny makes genuine connections ('Basically, it's the evolution of test driven development'), asks the under-covered question about what engineers lose in this new world, and follows up on planning effectively at the end; however, he allows very long personal tangents to run unchecked (bank teller story, grandma narrative), doesn't challenge the 8x metric's validity, and misses chances to push on how verification actually works in practice.

“Basically, it's the evolution of test driven development.”

“I'm curious what is lost in this new world of software engineering?”