Pinned Tweet
How to Find "Lost" Tweets ? 👇
Built an X post history tool in under 2 hours.
Not a weekend. Not a sprint.
Two hours.
Used vibe coding + small, focused adjustments instead of starting from scratch.
The idea is simple:
if you pause on a post, it means you actually read it → so it gets saved locally as your reading history.
No likes. No bookmarks. No algorithm games.
Just the posts that mattered to you.
What surprised me most:
the hardest part wasn’t the code - it was realizing how much time we waste rebuilding things that already exist in some form.
With the right tools + clarity, progress feels unfairly fast.
Honestly, building this from zero would’ve taken me way longer than 2 hours.
This is what building in 2026 feels like.
you deserves unpaid sick leave
European hiring rules are genuinely insane.
For example - if you own a company in the Netherlands and one of your employees gets sick, you are required to pay at least 70% of their wages for up to 2 years.
If they can't go back to work after 2 years and you haven't done enough for their reintegration, you pay their wages for 1 more year. And you can't fire them until after that extra year!
solo devs, share 1 product that you shipped and got you at least $1 in the AI era
I'm under the impression that pretty much all the markets segments are full right now of vibe coded tools and no one makes any $
Alex Mano retweeted
Turns out that GPT-6 Luna gets almost the same job done for a fraction of the Opus 5.5 cost 🤯
GPT-6 Luna is about 97.5% cheaper input/output than Opus 5.5
Or put another way, Opus 5.5 costs 40× more than Luna for both input/output
Turns out that GPT-6 Luna gets almost the same job done for a fraction of the Opus 5.5 cost 🤯
GPT-6 Luna is about 97.5% cheaper input/output than Opus 5.5
Or put another way, Opus 5.5 costs 40× more than Luna for both input/output
Opus 5.5 is about 60% cheaper than GPT-6 Astra for both input/output tokens.
It looks like this is the trajectory of AI cost models, better performance at lower prices.
AI tooling in desktop apps may not be another server layer
Agents discovering and using tools directly from the app repository
Less infrastructure. Low latency.
everyone is vibe coding right now
building stuff that was previously built in weeks or months
guess what's the missing piece of the puzzle 🧩
the developer career path is getting weird
first we wrote the code then we prompted AI to write it
now we're reviewing pull requests from a coworker who doesn't exist
somehow.. we're still on call 🧩
if developers stop writing code for months and become dependent on ai
what happens when ai is too expensive or inaccessible or unavailable ?
what happens to the skill we stopped practicing ?
All but nobody. @levelsio
sometimes you ship a tiny product and get zero sales
but while building it you stumble upon something much more useful
suddenly the failed product becomes the starting point for a new one
sometimes shipping it's about what you discover along the way 🧩
For MVPs and tiny products, you often don't need Postgres, Redis, queues, replicas…
Just ship with SQLite.
Less infrastructure. Less complexity. More time building.
Would you use SQLite in production?





