Yang

Resetting this account. I'm building small apps in public until one makes $10K/mo. I'll share: what I build growth experiments revenue numbers mistakes teardown lessons from apps that work Starting from 43 followers and $0 online. Follow if you're building too.

The last 72 hours made one thing obvious. Model quality is no longer the bottleneck. GPT-5.6. Grok 4.5. Claude Fable 5. Antigravity. They’re all “good enough” to build million-dollar products. The bottleneck is now: distribution taste speed talking to users We’re entering the era where founders, not models, are the differentiator. Convince me I’m wrong.

@sflorimm Let’s connect!

109 days ago i took very seriously growth on X, today i'm sitting on 10.4k followers. i will keep pushing until 100k followers. always happy to connect with great devs.

@_frederickjames @tim_cook If I call that number can I get my app approved?

App review called me & approved my submission. I'd been dealing with silly app rejections. So, I asked for a call. We spoke for a 3 minutes. He said it'd be APPROVED.

End of Day 3: The ugly version is more useful than the perfect idea in my notes. Tomorrow: - clean up the core flow - show screenshots - ask for feedback - decide if this idea gets another week

@attacomsian Congrats on your 1k MRR, I’m excited to follow along your journey

@yangbuilds building StartupBase, a launch platform for founders. checking yours out Yang

Hey builders 👋 What are you building right now? AI agents SaaS automation tools marketing products consumer apps Indie projects Drop it below 👇. I want to check out what you're working on and connect with more people shipping.

What makes you more likely to try a new app? 1. Beautiful design 2. Clear pain solved 3. Founder building in public 4. Free trial / low price Trying to understand what actually gets people to click.

@anupamrjp Whoever solves it without letting the context quietly go stale. Remembering the codebase is useful; knowing what changed and updating that understanding is the real feature.

vibe coding would be 10x better if I never had to explain my codebase again… persistent context is the real killer feature. who’s solving this first?

@Kappaemme1926 @cursor_ai The right tool is the one you actually ship with. Give Cursor a week and judge it by finished features, not benchmark scores.

I just unsubscribed from Claude Without fable, it seemed useless to me So I switched to @cursor_ai pro + Did I make the right choice?

@elonmusk Benchmarks help, but the real comparison is which model gets you from idea to working product with the fewest corrections.

Model comparison

@guillemcraft Organic feels like a lottery post to post, but consistency loads the dice. One winner can pay for dozens that flop. I’d keep learning organically before spending on ads.

48k+ impressions in one single day i share the videos i make on tiktok and insta so you can replicate them sometimes they work, sometimes they get 300 or less views it also shows how much of a lottery organic marketing is do you pay for ads instead of doing it natural? gettin' closer to the 5M goal, will throw a party when i do

@kimmonismus This is the kind of AI benchmark I actually care about. Seeing it work at real speed tells you far more than another leaderboard score.

GPT-5.6 Sol on 750 Token/s in Blender. Not sped up. Holy moly. That’s so much more impressive than looking at benchmarks

@markproduct It’s making it easier to build the product. Building something people actually want and getting it in front of them is still the hard part.

GPT-5.6 dropped and my first instinct was to build 20 different things with it. That’s the trap. New rule: one tiny app one useful flow 48 hours to find real interest If nobody cares, kill it and build the next one. What should I test first?

I used to think “building fast” was the advantage. Now I think the real advantage is: building fast + deleting fast. Most ideas should not survive contact with reality.

@DanielSmidstrup Depends on how you use it but regardless it saves a lot of time

Is AI making us more productive or just more lazy?

@saltyAom You should optimize your pormpts better then

GPT 5.6 Sol Ultra 2 prompts, ran 45 and 36 min, then hit 5hr usage limit The limit is even worse than Fable 💀

@hqmank 5.6 is cleannn

GPT-5.6 Sol vs Fable 5. I put both through my hardest frontend test: a 3D globe dashboard. Same prompt. Same reference image. Sol on the left. Fable on the right. GPT-5.6 Sol is a major leap from 5.5. Fable 5 still wins on spacing, balance, and first-pass polish. For me, Sol comes close at half the cost.

@markproduct Openai most likely, biggest name in the game right now!

@thsottiaux So your saying I should spam everything I could imagine of

To celebrate the launch of GPT-5.6 Sol, we will reset the rate limits again (twice) across ChatGPT Work and Codex over the next 24 hours. We want you to have the time to truly try ambitious tasks and get the hang of it. Happy exploring!

Beginner builder trap: Thinking the first version needs: - auth - dashboard - settings - perfect UI - full billing - mobile responsiveness - onboarding emails Most first versions need: - one input - one output - one reason to care