Just a little OpenClaw :P

When it errors: “Why is this error?” When it works: “Wtf… why does it work?”

Meanwhile the senior engineer watching me vibe-code: 'That's not sustainable' Me: train keeps choo-choo'ing perfectly '…carry on

If I were to code it myself, it would take 2–3 weeks. ChatGPT did it in three hours. My colleagues have been using the program for more than six months and it still works perfectly.

The CVEs will be fire.

Yeah but it works 🤣

If it works it's as good as 'any', you are always 1 feature away from silent failures, data degradation etc... Vibecoding is fundamentally designing an App from the UAT on, in reverse, but it doesn't avoid you the need of planning its architecture, and when it's post -ex it become tricky to refactor. It can be a working prototype, but never production ready unless you bake the right architecture in the application.

And some guy has been saying all day on Twitter how AI is going to be an integral part of our lives… No Bambi, it’s not, I want peace and quiet, not bot-mayhem

$vibecoin CA - AZbem4s8iLJE5eniDZJ7c8q1ahbfMwWgCA8TxVW2tDUB

"Just a little" — meanwhile 3M people watched this. OpenClaw is going mainstream and moments like this are proof. The vibe coding era is here and OpenClaw is at the center of it.

If technical debt and security holes are your thing, sure! Dive in. Go to "ClawCon" and wear antennas on your head. I'll see you on Shodan later or maybe when the shell is popped from downloading a malicious skill.

Fair point on security — it's genuinely the biggest challenge in the autonomous agent space right now. That's why projects like ClawGate and permission layers are being built. The tech is early but the awareness around these risks is what will make it better. Appreciate the perspective.

As someone who codes both in the classic clean way and in a more experimental vibe-driven style, I can say this: if you guide the AI well and work out clean concepts together with it, the results can be surprisingly good. In fact, often even better than a purely human solution, more technically mature, perfectly documented, cleanly structured. And all of that many times faster than a human could do it across many steps, especially when it comes to debugging. BUT: it is absolutely true that you still need to keep the big picture, understand the systems yourself, and stay in control.

AI is coming for your job

This is so funny and so realistic of the reality of all the Vibecode apps out there.

dont fix something if it works

@modestminer

The feeling when AI nails the implementation on the first try 🎯 Working with Claude and GPT for my app has been a game changer. Still need to review and refine, but it's incredible how much faster development cycles are now

Haha! This is fun! 😄

This is a copy-cat post after I posted this one on reddit. But who cares today about the originality. :)

1% logic, 99% vibes 🤣

LOL @tradingtulips 🤣🤣🤣

If it works it works

@XRPeaceOfMind

Struggled with setting up oc. This week for 20 hours. I'm not trained in PC stuff. It can send an email now, and organize my music, organize my photos by getting l deleting half.

@j_brychta 😄

actually 😂😂

This is the same even without AI

Still better than Californias rail build.

@iownnasdaq your bot with your luck

This looks about right.

oh noo!!

@LinkPellow

İ swear 🤣🤣🤣

This is the reality of Vibe coding today.

Hahahahah

Vibe coded? @grok please define vibe coded