Georgios Diamantopoulos @georgiosd

@everestchris6 Reel

On January 15th, 2009, Captain Chesley Sullenberger glided a plane carrying 155 people onto the Hudson with both engines dead. Everyone walked away. Asked later how, he said: "For 42 years, I've been making small, regular deposits in this bank of experience, education and training. And on January 15 the balance was sufficient so that I could make a very large withdrawal." That banked balance has a name: reserve. The capacity you've built for a demand you can't see coming. Your body keeps the same kind of account, and most of us are flying on a balance we've never looked at. The withdrawal comes with no warning: the back that "goes" lifting a box, the knee that won't take the stairs, the fall at 70 you don't get up from. You don't pick the day, or which capacity it asks for. The only thing you control is the balance you built before it shows up. Here's the uncomfortable part. It's the one account with no overdraft protection, and no v2. Last year at Future Frontend I spent almost the whole talk on the size of that withdrawal. Arteries stiffening within hours of sitting. The discs. The fact that developers sit more of the day than any profession measured, around 90% of it. All true, and people came up afterwards a little shaken. Then they asked the question that actually matters: ok, so what do I do? There's a line, usually pinned on Einstein, that you don't really understand something until you can explain it simply. Last year I could explain the problem in my sleep. The fix I had, but not yet boiled down to something you could act on: it was the last five minutes of a talk that spent forty on the problem, 80% diagnosis and 20% what-to-do. A year of living it later, I can say it plainly. So this year's version flips the ratio. Mostly the deposits now: how you build the balance on the maintenance budget, the boring daily way, before the bill lands. Closer to half diagnosis, half prescription. It's called Building Resilience, and reserve is most of what resilience really means. I'm giving it at Future Frontend in Helsinki on June 8. While I'm over there I've kept a couple of days free to run the hands-on version with engineering teams: the actual routines, the mobility work, building movement into a workday instead of bolting it on afterwards. If your team is in the area and that's useful, reach out. I'm taking a few of these on this year regardless of geography, so if you're nowhere near Finland, drop me a line anyway. You make the deposits on ordinary days, against a withdrawal you can't predict. Today is an ordinary day.

Original video: youtube.com/watch?v=zdUNGQ…

What do you call someone with zero experience who sets out to make the greatest piece of pancake art in history? Delusional, probably. That was my read too, until I saw what he actually did about it. He didn't shortcut a thing. Four months. 200 pancakes, freeze-dried, fitted together, sealed in resin. He taught himself freeze-dry chemistry to stop them warping and built custom molds when none existed. He was the artist and the engineer. He went through every part of it and never tried to skip the hard half. (link below!) That used to be unremarkable; it's what making something great has always cost. Today it looks almost alien, because we've fused two appetites that were never meant to meet. The first is speed. We order something at midnight and it's on the porch by 9am, and somewhere in there waiting itself turned into a defect. The second is scale. Every founder story that reaches us ends in a billion-dollar exit, so making something small and good quietly stopped feeling like it counts. Want it now, want it huge. Modern entrepreneurship stirred the two into one pitch: Skip the apprenticeship. Ship the empire. The fantasy underneath it is that the art and the engineering come apart. That you can own the vision and rent the making. Nowhere is that pitch louder right now than in software. Describe the app, ship the app, never learn how it works. You get the empire, a thing that runs, and you skip the one part that was ever going to make you a builder: understanding it well enough to explain or fix it when it breaks. It runs beautifully, right up until it breaks. Then someone asks you what went wrong, and you have no answer. Craftsmanship is the orphan in that house: both halves, same hands, over a long time. A culture built on speed and scale quietly starves it. His pour failed, by the way. Bubbles everywhere, four months gone. He's still more of a craftsman than anyone who shipped an empire they couldn't have built by hand. You don't have to suffer four months to make something worth keeping. But notice what the artist never did: he never let the art and the engineering come apart. He made the whole thing himself — the vision and the freeze-dryer both — and that's exactly the part the pitch now tells you to hand off. So vibe-code it. Ship it fast. Nobody's asking you to go back to writing every line by hand. But before it goes out, do one of two things: understand it well enough to explain and fix it when it breaks, or put it in front of someone who can. You can skip both; it'll just come back to bite you the first time it falls over and nobody in the room can say why. Getting expert eyes on it is the floor. Becoming someone who could have built it yourself was always the point.

@johnvirality Rizz

@georgesttock Workflow

@lucaspatiri_ UGC

@johnvirality ARCADS

@blvckledge demand

@itsalexvacca design

"More is not better, only better is better." — Scott Sonnon said that about fitness. It's the cleanest critique of AI coding I've heard. His prerequisite was the load-bearing half: you can only assess "better" if you're clear on the goal AND know how to measure it. Without both, "more" is what's left. What he was attacking: pop fitness selling personal records — more weight, more reps, faster times. The stuff that fits on a board. Recovery, mobility, the body you'll have in five years — the stuff that actually matters — never makes the board. The personal record comes this Saturday. The torn meniscus comes later. Same trick, different gym. The AI-coding pitch sells "more" — features shipped, build speed, funding raised — because speed is what the pitch can measure. Define "better" and most of the "more" stops counting. Not because anybody's lying. The math just changes when you add a denominator. The question isn't whether the AI can ship. It's whether you've defined what "shipped well" means, before it does. Three things "better" looks like at the keyboard, before any user sees the app. Your phone wakes you up. Production's broken. Can you debug it half-asleep when you see "Claude API Error: Overloaded"? If not, don't merge it. Diffs you've actually read. Architecture that survives the next feature. You don't have to slow down, but you do have to know what "better" means before you ask AI for more of it. That's what I call Human-Driven Development — constraining AI before it writes, not after.

@codyschneider AEODAshboard

@FynCas V2

@itsalexvacca ADS

@vitaliidodonov @IAmPascio MRR

@chrisboettcher9 No sir

@ReactAdvanced @TheAlexLichter @FlorianRappl @plahteenlahti @etodanik @puppo92 @PopovaViolina @gotbahn @KrasimirTsonev @souporserious @DanteDeRuwe @viniciosneves @TkDodo @krisvelkov @HoudaMouttalib @ericbureltech @pawelsaw @GregBrimble @bradwestfall Wish I could be there in person but thank God for streaming!

💻Learn from the engineers shaping the frameworks and tools you use every day. Can't wait to see you there💚reactadvanced.com

@dhh A bit surprised you don’t have a standing desk. Is that a conscious choice?