Love this driving analogy… Leadership sits in the car. Your engineers are driving. You pull the handbrake only when the direction is actually wrong. Then you communicate clearly and let them get back on course. The ideal leadership style imo.

Funny thing about AI in hiring: it made fakers easier to spot. Ask someone to walk you through the hardest thing they ever built. If it sounds scripted or they're reading off a side screen, you know. The "can you write JavaScript" era is over.

AI is helping teams move faster, but the pressure to constantly keep up can quietly turn productivity into burnout. As AI changes how we work, leaders need to rethink how they measure performance, and how they protect the balance between output and sustainability.

I'm starting a series of free workshops for folks wanting to learn Argentine Spanish. My end goal here is to eventually accept donations for these workshops that will 100% go to provide English school scholarships to underprivileged children. cal.com/argentine-spanish-fr…

AI-generated code that nobody fully understands is getting merged into production. It looks fine, but then it breaks later, and nobody knows why. AI is crazy useful but it still needs to earn its way into the codebase through proper guardrails.

What's the new joke for "that'll take 2 weeks" in software estimations? "I can one-shot it"? Do you even do estimates any more? (slightly rhetorical but I'm curious)

Funny thing about posting is I have around 450 followers. But an engineer on my team referenced a post, and someone I hadn't spoken to in years reached out to reconnect over managing engineers with AI. I was reluctant to start posting more content but this is why I started. If I can positively influence even a small amount of people, that’s a great outcome.

Google open-sourced AX, their agentic orchestrator. first impression: it feels like kubernetes for agent executions. check it out - github.com/google/ax

The goal isn't to resist AI or use it recklessly. It's to build the systems that let teams trust it. Combine AI-first workflows with clear policies, automated testing, and human oversight where it matters, and you get speed without chaos.

I went full digital nomad in 2017 thinking remote work would be easy. The first year taught me otherwise. Time zones, no routine, and working somewhere you barely speak the language forces you to build better habits. Turns out those habits became useful leading global teams.

Many engineers over-engineer too soon. They'll make one function the most performant thing on earth because it's fun, not because it's needed. But the other extreme is shipping slop. So make sure you make small checks along the way, don’t go overboard from day one.

AI is changing software engineering from solving every low-level complexity by hand to focusing more on what we can actually create. The real opportunity isn't building faster just to ship more. It's using that newfound speed to build better systems, better products, and ultimately better outcomes for people.

Best leadership question I've heard: What's the one decision you can make that makes a thousand decisions? That's the whole job now. I'm pulling out of the day-to-day code so I can think 6 and 12 months out instead of fixing what my team is already handling.

Made a fun visualization of our @calcom engineering team's PR interactions.

The hardest feedback conversations become easier when you lead with facts instead of assumptions. Give feedback early and often, use data to remove subjectivity, and then create space to understand the context behind the numbers, because honest conversations work best when they're grounded in reality.

Today we had degraded service for ~38 minutes due to a bad query counting inefficiently against a table with tens of millions of records in it. Sincere apologies for the disruption. Our internal processes for reviewing and testing these queries clearly need work. We are on it.