Saw this thread from @paoloanzn doing the rounds and it nails something I’ve been chewing on.
vibecoder asks claude code to build a chat app, gets a working prototype in 20 minutes, immediately tweets “just killed slack and discord”…
brother you don’t even know what a distributed system is. you don’t know what database replication means. you have no idea how websocket connections behave at scale or what happens when 50k people are online at once and someone’s message needs to show up in 200ms across 3 continents
slack has engineers making $300k+ who have spent a decade solving problems you don’t even know exist yet. race conditions, eventual consistency, message ordering, presence systems, file storage at scale, search indexing across billions of messages
your app works on localhost with 2 connections. that’s not the same thing as “killing slack” that’s a college homework assignment
the prototype is maybe 0.5% of what makes these products actually work in production. the remaining 99.5% is infrastructure, reliability, edge cases, and years of iteration on problems that only surface when real humans use your thing at scale
and the worst part is the confidence. “yeah its not perfect but ai one-shotted it, just need to adjust a few things and deploy” - the few things you need to adjust IS the entire product. thats like pouring a foundation and saying you basically built a skyscraper, just need to adjust a few things
ai is genuinely incredible for building tools and prototypes. i use it every day. but there’s this weird thing happening where people who have never shipped anything to real users at scale now think the hard part of software is writing the first 200 lines of code
it never was bro
Paolo is right and the framing matters.
I now use these tools every day. The acceleration is real and I’m not interested in pretending otherwise. For personal projects, throwaway scripts, internal tools, prototypes, learning exercises: ship away. The AI is your unfair advantage and you should use it.
But acceleration is not substitution. AI is a development accelerator. It is not a replacement for software engineering nous, the kind that comes from shipping things to real users and watching them break in ways you didn’t predict. A model can give you the appearance of a solution in twenty minutes. Whether that solution survives contact with scale, reliability, security and a paying customer base is a completely separate question, and it’s the question the hype skips.
The localhost demo with two clients connected is a homework assignment. It is not a competitor.
What’s strange about this moment is the confidence. The “few things you need to adjust” really are the entire product. The websockets that don’t fall over at 50,000 concurrent users, the message ordering that survives a network partition, the search index across billions of messages, the presence system that doesn’t lie. None of that gets one-shotted. It gets earned, badly, over years.
The limiting factor in software has shifted, but it hasn’t shifted to where the loudest people think it has. It’s moved from “can you write the code” to “do you understand what the code actually has to do.” That second question is harder than it’s ever been, because the cost of generating plausible-looking answers has collapsed to zero.
The hard part is still the hard part. It always was.
Sources:
- @paoloanzn on Threads (threads.net)