chilldev

Imagine going back in time, walking into this room, and knowing what the next 20 years would look like.

hey @grok if anyone can build an app in a weekend now, what’s actually worth spending 5 years getting good at?

lol crazy times we living in

road to $1k mrr, week 2 mrr: $233 (same as last week lol) apps: 3 followers: 415 (+3) what moved this week: nothing. stripe is quiet still posting every monday till it says $1k

deployed on halloween if it breaks at midnight i deserve it

hot take: ai didn't make coding easier it made reading code the whole job agree or am i just bad at prompting

deployed to prod prayed checked slack nothing broke checked again anyway

a vibecoder once asked: “Who is API and why are we calling him”

what kills more saas companies a) no users b) no distribution c) shipping features nobody asked for d) checking analytics instead of talking to users pick one

> wake up > 1 new signup > finally > check stripe > it's me testing checkout again

road to $1k mrr, real numbers mrr: $233 apps: 3 (two of them make $0 lol) followers: 412 fake screenshots: 0 posting this every monday till it says $1k

Link previews that move. Grassy now roasts you right in iMessage

Railway had an incident today. And honestly, I think this should be a wake-up call for everyone building software. It's a reminder of something that's easy to forget when you're moving fast: your infrastructure is part of your product, whether your users ever see it or not. You can have the best code, the cleanest architecture, and the most carefully designed application. But when a critical part of your infrastructure goes down, everything built on top of it can be affected. Managed platforms have made deploying applications ridiculously easy. A few years ago, getting a production environment running required significantly more infrastructure knowledge and operational work. Now, you can deploy an application, connect a database, configure a domain, and ship to production without managing an entire fleet of servers yourself. That's incredible. But there's a difference between making deployment easy and making an application resilient. And I think a lot of us are moving so fast that we're forgetting that. What happens when your database becomes unavailable? What happens when a background job gets interrupted halfway through execution? What happens when an external API times out, a queue stops processing, or your infrastructure provider experiences an outage? Does your application recover gracefully, or does everything start falling apart? Now, here's where things get more serious. We're entering an era where AI isn't just generating code or answering questions. We're giving AI agents access to production environments, databases, APIs, deployment pipelines, and systems that can make real changes to the world. We're moving toward software that can execute entire workflows autonomously, make decisions, modify code, and take actions without a human reviewing every single step. Think about what happens when we combine that level of autonomy with infrastructure failures, poorly designed recovery mechanisms, or agents that make the wrong decision. A bug that once affected a single request could trigger thousands of repeated actions. A failed workflow could leave data inconsistent. An autonomous agent could misdiagnose an incident and make things worse while trying to fix it. And a failure in one critical dependency could cascade through an entire system. These aren't reasons to stop building with AI. They're reasons to take engineering more seriously than ever. The more power we give our software, the more responsibility we have to make sure it fails safely. We need proper monitoring, reliable backups, idempotent operations, bounded retries, permission boundaries, human approval for high-impact actions, and systems that can stop themselves when something doesn't look right. We need to design for failure instead of assuming everything will always work. And we need to stop treating infrastructure, security, and reliability as things we'll figure out after we get our first thousand users. Because by then, the consequences of getting things wrong can be much bigger. AI will make building software faster. It will make complex systems accessible to smaller teams. It will unlock things we couldn't realistically build before. But faster development also means we can create and deploy problems faster than ever. That's the part we shouldn't ignore. Today's incident is a reminder that even the infrastructure we trust can experience failures. No provider is immune, and no architecture is magically resilient just because it's built on a managed platform. The future of software isn't just about what we can build. It's about whether we can trust what we build. And with AI making software increasingly autonomous, that distinction is going to matter a lot more than most people realize.

won't be surprised if someone turns the whole internet off