Fixed: a blank day on the timesheet means a missed entry or no work was done.
Now that we’ve freed our minds from coding, we can focus on the real value -> Product
If you don't know what you're doing, @claudeai just changed the currency of bad design.
Where you used to pay in hours, now it's tokens.
The real cost comes down the road 💰
Agents didn't kill coding. They hid the early cost.
Code used to be slow, so bad architecture hurt as you typed. Now it's cheap, so bad design only hurts down the road.
Just discovered `bunx ccusage monthly`. Breaks down what your Claude Code plan is actually worth 📊
If you're on the Max plan, this will either make you feel great or make you realize you need to vibe code harder.
(Works with codex too, but... who uses that 😅)
"Vague issue in, shitty PR out." I disagree.
Agents are an amplifier. Good design in, leverage out. Vague design in, a mess out, just faster.
The output tracks what you bring.
Agent loops are running half of my dev work.
But I still review everything, close PRs, leave comments, challenge architectural decisions. I'm the bottleneck 🐌
Wondering where I should draw the line between trusting the loop and checking the work... 🤔💬
My small team of agents just opened 55 relevant issues.
Impressive, but purely reactive 🥱
Sentry errors, CVE patches, dev deps...
The real challenge is now moving agents from maintenance to product work 💡 but how?
Agent loops are changing dev work forever.
The core skill is no longer writing the code itself. It’s writing self-healing instructions, launching them in autonomous loops, and keeping the whole thing under control.
I'm about to fire myself as a developer.
My agents consume live context, they open their own issues, groom them, create PRs, review and even approve each other.
A year ago I wrote every line.
Now I mostly watch.
Where do you stop trusting it?
How do you find podcasts? 🎧
Is there anything like IMDb or Goodreads for podcasts?
I'd like to follow people, or topics, not necessarily specific podcasts.
Company culture isn’t about beer taps at the office or everyone wearing company branded swag. It’s giving your team a framework for making decisions without your involvement. 💡
Finally getting started with python for production code! But still struggling to understand some of the tools.
What should I use to install different python versions per project, pyenv? Then how to manage dependencies, poetry? What value does venv really add? 🧐
GPT has an opinion, and I agree.
It says if you're using pyenv together with poetry, then there's no need for venv.
But is that correct? 🤔
chat.openai.com/share/ece733…
Flavio Antelo retweeted
🌈 Ever wondered how to integrate Cloudinary with #rubyonrails seamlessly?
@fwuensche breaks it down in this article! Learn how to leverage their strengths in your projects, staying within Cloudinary's free tier.
Check it out: antelo.live/create-a-photo-g…
I've just learned about useEffectEvents in React and I think I finally found the answer to all my dependency array discussions. 🤯 react.dev/learn/separating-e…
Flavio Antelo retweeted
Mon toubib test l'utilisation de l'IA avec Doctolib : il a le compte rendu de notre entretien en quelques minutes, assez précis ! C'est étonnant.
Pour lui, gain de temps, pour moi pas besoin de répéter...
Là c'était un gros problème d'allergie : les arbres sont en bourgeon un peu plus tôt que prévu. C'est fatal pour moi !
Por acaso Bhudda era empreendedor? 🤔
"Há apenas dois erros que se pode cometer no caminho para a verdade: não ir até o fim e não começar."
Acabei de receber minha entrada pra #tropicalrb no Brasil, só que eu tô na França. Ai geinte, e agora? 💅





