No need to follow the Sylius template scheme in your app! Sylius must offer high granularity for easy template customization. Create hooks in the end-apps only where sensible, as overusing them might cause worse DX.

The price of Claude Code is discouraging. Compared to Junie, it's 4.5 times more expensive (in the cheapest Claude Code plan and priciest JetBrains AI plan). I doubt it's 4.5 times better than Junie or other AI tools.

Mono-repositories streamline open-source release orchestration and ensure components work together. Mono-repositories add complexity, requiring a robust strategy to ensure components function independently if it's one of the goals.

I love sections in README files that present other packages as alternatives with their own pros and cons, not as competitors. Kudos to the `thephpleague/object-mapper`!

Do you use a separate directory with a separate `composer.json` file for tools like PHPUnit and PHPStan? It seemed like an overcomplication to me, unless some tool blocked me from upgrading a non-dev dependency version. 😅

Great news: #FrankenPHP worker mode will be automatically supported in #Symfony in 7.4. No more additional package required! 💪 github.com/symfony/symfon… #php Kévin Dunglas Symfony

Creating a robust Docker setup (images, Dockerfile, configuration) can be challenging. In theory, containerization simplifies and unifies environment setup. However, I know many cases where developers spent excessive time setting up projects with containerization.

Learning how to use AI to learn new things faster > talking only about how AI is gonna replace us

During our sync call, we discussed AI agent tools. Someone asked which tool to use. Great question, but the choice doesn't matter initially. Pick any—Junie, Claude, Codex, etc. The key is understanding how to use these tools and their shared limitations and fundamentals.

Updated list of recommended AI agents for PHP developers 🐘🤖 - Junie AI – my top choice, but slow - Claude Code – widely tested internally - Gemini CLI – not ideal for PHP, still testing; may need `GEMINI.md` tuning

I’m using All-in Pack, so inspired by Tomas I’m sharing my discount code too 😊

I wouldn't trust Chinese-based AI for commercial projects, but at $2-3, using it for open-source projects is incredible. I built a working Playwright ↔ Behat driver prototype. It's not perfect—tests run in CLI and perform basic tasks—but polishing it will be faster than writing

Super wise CTO, period.

Interesting that GPT-5 didn't come along with a *real* Claude Code/Gemini CLI/Any other CLI-based tool competitor, including both UX and affordability (using the subscription instead of an API key).

How many more model releases do we need for folks to realize we are not getting to magical superintelligence with what we got? How many times do you have to see a model benchmaxxing to realize Humanity's Last Exam is a freaking idiotic name and that answering questions on it

Ridiculous that OpenAI claimed 74.9% on SWE-Bench just to prove they were above Opus 4.1’s 74.5%… By running it on 477 problems instead of the full 500. Their system card only says 74% too.

Fun fact: JetBrains' Junie switched to GPT-5 as its default model. For the prompt “Check changes in file X and update the unit test,” it generated twice as many steps and took longer than Sonnet.

Why can't we have a model selector and `@` file selecting in Junie as we have in the AI assistant chat 🥲.

This is a sad day for the entire #symfony community. My thoughts and love are with Leann and Beckett.