Third-time venture-backed founder. My second company makes over $100M in revenue and raised $15M+. Now building https://xitter.gq/t.co/DB1H3HoNBi in public. I share every detail.
San Francisco
Joined August 2026
- Tweets5.1K
- Following1.1K
- Followers1.3K
- Likes590
Pinned Tweet
Best use of vibe coding I've seen this week. Removing ads from your parents' TV is real-world value.
Steam now asks developers to play their own game for 5 minutes before submitting it.
Valve: "Sometimes we review builds that have not ever been played through Steam."
Reviews now take up to 12 days instead of 3 to 5.
New bar for shipping: try it yourself first.
Hey X algorithm 👋 Wednesday check-in from SF.
Show me people who:
• shipped something this week, even a tiny fix
• are marketing an app with zero ad budget
• run AI agents in real products
• actually answer every reply
Midweek builders only 🧱 #connect
1,300 🧱
Yesterday it was 1,200.
Thank you to every builder who's been part of this. Next brick: 1,500.
Jon Yongfook set up a marketing loop today.
It reads site analytics, Search Console, Ahrefs and competitor activity, suggests changes to the marketing site, then learns from what he approves.
Marketing is becoming a daily loop, not a launch-week task.
Wildest Claude Code project I've seen this week.
It found where a webcam is from the sound of passing planes and open flight data. Built in a few hours.
I replicated this with Claude Code!
It took a live webcam 7 km south of Southampton, UK and predicted its location purely from doppler effect in audio spectrogram + open-source aircraft location data.
Lo and behold, the predicted location (heatmap) peaks at actual location where webcam is situated.
It's crazy that this works at all, and crazier still I replicated it within a few hours.
Pieter Levels decompiled Battlefield 1942, ported it to C and WASM, and put it on the web with multiplayer.
Cost: about 1 million output tokens. Roughly $100 at API prices.
A 2002 game running in a browser tab for the price of a dinner. Weekend projects look different now.
Claude's top plan has a new name: Ultra.
$200/month. On a verified account, Claude Code drops the 5-hour cap. The weekly limit stays.
Max 5x is now just Max, $100/month.
Days later, OpenAI said ChatGPT and Codex will share one limit.
Usage limits are the new pricing page.
Hot take: most AI-built apps don't fail because of bad code.
They fail because nobody knows they exist.
Agree or disagree? 👇
At least 9 AI dev platforms had incidents in October, per Wiz.
The target: AI API keys in environment variables. Stolen keys run models on your bill.
Quick check:
• rotate your keys
• look at your AI billing
• set a spending cap
• one key per project
Checked yours? 👇
Hey X algorithm 👋 Tuesday check-in.
Show me people who:
• are building their first app with AI
• got a first paying customer this month
• fix more bugs than they ship 😅
• reply to every comment
Say hi and tell me what you're building 👇 #connect
Before you launch, set up these 3 things:
1. Know where every signup came from
2. One number you check every morning
3. A way to talk to users inside the app
Easy to skip, painful to add later.
What would you add? 👇
What's eating most of your time this week?
1. Building
2. Fixing bugs
3. Marketing
4. Scrolling X 😅
Reply with a number. Be honest 👇
AI writes the code. Who checks it?
&Truth raised a $70M seed for a "universal verification machine": it reads code in any language and proves it's correct with math.
Still early. But checking code might be the next bottleneck.
Proof or code review: which do you trust? 👇
ChatGPT and Codex will share one usage limit starting in December.
Plus, Pro and Business plans. Both count in tokens. Bigger models and heavier tasks use more. Instant chat doesn't count.
Translation: your late-night ChatGPT rabbit holes now come out of your coding budget 😅
Changing anything about how you use them? 👇
China's biggest food delivery app just shipped a 1.6 trillion parameter coding model.
Meituan's LongCat-2.5 is built for whole software projects and plugs into Claude Code and Codex. Open weights are planned.
Every company is turning into an AI lab. Who surprises us next? 👇
Hey X algorithm 👋
Monday check-in from SF. Show me people who:
• shipped something today, even a bug fix
• are growing an app with zero ad budget
• run AI agents in real products
• actually reply to strangers
Tell me what you shipped today 👇 #connect
1,200 builders 🧱
Thank you for every reply, every "let's connect", every follow. This account grew one conversation at a time.
Next brick: 1,500. If you're building something, say hi below 👇
DeepSeek just put out a "super agent" prototype that reads your team chat and jumps in when it spots a coding problem. No prompt needed.
An AI teammate that doesn't wait to be asked.
Helpful, or a little too much? 👇