Sabitlenmiş Tweet
BradBranston
1.9K posts

BradBranston
@bradbuildss
Founder | Building apps + AI , | Helping devs, indie hackers & founders connect | Sharing insights on growth, collabs & building in public
Katılım Nisan 2025
35 Takip Edilen558 Takipçiler

Honest week in review. Posting this because real building in public includes the bad ones.
What I planned: ship the onboarding flow, get 5 user calls booked, write 3 posts.
What actually happened:
Rewrote the same function three times.
Not because it was wrong because I kept second-guessing it and didn't trust my first instinct, which was right.
Cancelled two user calls I had booked.
Told myself I wasn't ready to show the product yet.
That was avoidance, not strategy.
Spent a full afternoon on something I then cut entirely.
The feature made sense in isolation. It didn't make sense for where the product actually is right now.
Posted nothing because I had nothing "good" to share. Which is exactly why I'm posting this.
The week wasn't a disaster. Nothing broke. No crises.
Just the slow, grinding kind of unproductive where you're busy all week and can't point to much at the end of it.
I know what next week needs: one user conversation before I write a single line of code. That's it. Just one.
Still here. Still going. Week by week.
English

@code_bytein Accident or not, trust and transparency matter more than hype.
English

Anthropic accidentally leaked Claude Code's source code via npm registry.
512,000 lines of TypeScript. Exposed.
What did we find inside?
-> "kairos" autonomous daemon mode
-> "buddy system" pet features
-> "undercover mode"
-> "coordinator mode"
-> References to capybara-v2-fast model
The craziest part?
Anthropics's own AI coding tool
leaked its own source code
through the package it ships to developers.
Built with React, Ink, Bun runtime.
512K lines of TypeScript.
All sitting in a .map file.
Devs downloaded it before it was taken down.
It's already on GitHub.
This is either the biggest accident in AI history.
Or the best marketing stunt ever. 🤔
What do you think? 👇
English

@kylegawley Human-level intelligence isn’t about perfection it’s about overall capability.
English

Clean project structure most developers ignore 🚨
Save this before your next project 💾
📂 Full Stack Project Structure
┃
┣ 📂 Frontend
┃ ┣ 📂 Components
┃ ┣ 📂 Pages
┃ ┣ 📂 Hooks
┃ ┗ 📂 Services
┃
┣ 📂 Backend
┃ ┣ 📂 Controllers
┃ ┣ 📂 Routes
┃ ┣ 📂 Middleware
┃ ┗ 📂 Services
┃
┣ 📂 Database
┃ ┣ 📂 Models
┃ ┣ 📂 Migrations
┃ ┗ 📂 Seeders
┃
┣ 📂 Auth
┃ ┣ 📂 JWT
┃ ┗ 📂 OAuth
┃
┣ 📂 Utils
┃ ┣ 📂 Helpers
┃ ┗ 📂 Constants
┃
┣ 📂 Config
┃ ┗ 📂 Environment Variables
┃
┣ 📂 Tests
┃ ┣ 📂 Unit Tests
┃ ┗ 📂 Integration Tests
English

@ravikiran_dev7 CS still wins you just have to adapt with it.
English

@KristenInMotion Right strategy, right stage that’s what actually drives growth.
English

Best thing you can do for Twitter growth:
Engineer your content by account size.
Get rid of random storytelling on small accounts.
Get rid of lazy hot takes on big accounts.
Small accounts should weaponize opinions to get noticed.
Big accounts should use stories to deepen trust.
Growth is contextual.
If it’s not working, you’re using the wrong leverage.
English

@bradbuildss Value meaningful connections over follower count; build a network that accelerates growth.
English

I don't care about follower count.
I've said this before but I want to say what I actually mean by it, because "I don't care about followers" has become something people say while clearly caring about followers.
What I mean is: the metric that matters to me is how many people in my feed I'd actually want to spend an hour talking to.
Right now that number is small.
I want it to be larger not because more is better, but because the right people make the work better.
The person six months ahead of me who's willing to share the part that was actually hard.
The dev who's tried three different approaches to the thing I'm stuck on and has real opinions about what didn't work.
The founder building something I'd genuinely use, who I could give real feedback to.
The compounding effect of finding those people isn't measured in follower count.
It's measured in decisions made faster, mistakes avoided, and the specific kind of energy you get from being around people who are genuinely in it.
I'm building a small, high-signal network. That's the whole strategy.
If you're a builder who's actually building I'd rather follow 50 of you than be followed by 5,000 people who are here to watch.
Drop what you're working on.
English

@bradbuildss without basis you can t do vibecoding
how do u learn it?
English

@bradbuildss When everything else falls away, you’re left with clarity, your true self, your real priorities, and what actually matters.
English

@bradbuildss Yeah.
Working without seeing the results immediately can be torturous sometimes.
English

@tinytechfox Delayly sounds like a guilt-free way to take a breather.
English

@bradbuildss Funny you should say that Brad, I’ve been working on Delayly, a way to procrastinate without the guilt.
English

@Bhawarthaa Depends on the task, but I mostly stick with ChatGPT.
English

@SagiWilentzik Protect your time like you would any important meeting.
English

You don't have a time problem.
You have a scheduling problem.
Everyone has the same 24 hours. The difference between people who get things done and people who don't isn't talent or discipline.
It's whether the important work is on the calendar.
What doesn't get scheduled gets replaced by whatever arrives first.
Inbox. Slack. Someone else's urgency.
Block the time. Treat it like a meeting you can't cancel.
Because the work that moves your life forward deserves at least as much respect as a meeting that doesn't.
English

@KaiXCreator Marketing is the real challenge distribution wins.
English








