Karel

The man who killed the $10,000 GPU myth. He did it alone, from Bulgaria, with one C file. 🤯 >Meet Georgi Gerganov. >Bulgarian developer. Nobody had heard of him. >In March 2023, Meta’s LLaMA model leaked online >Within days he wrote a single C file >Called it llama.cpp >It ran a full AI model on a MacBook. No GPU. No cloud. >The entire AI industry said you needed $10,000 GPUs to run LLMs 🔥 >He proved you didn’t. On a laptop. Alone. >Also built whisper.cpp ~ same thing for voice AI > His code is the foundation of Ollama, LM Studio, and GPT4All >107,000+ GitHub stars. Fastest open-source AI project to hit 100K ever. 🚀 >In 2026 Hugging Face hired his entire team >Still ships code. Still open source. Still free. Every time you run AI locally, you’re running his work. Absolute Legend 🐐

@0x15f @enunomaduro @taylorotwell This is without knowing anything about your test suite or app, but a few thoughts off the top of my noggin. Realistically, I don't think there will be a quick solve here. My guess is item one will be the place to dig but also requires the most work.

@adamwathan What's the business model?

Quick ui.sh demo — generating multiple design ideas to choose from, no matter what tech stack you use:

Proud of our team for making this happen 🥲 Many years of work, lots of discussions, and many prototypes painfully taken back to the drawing board. *of course it isn't just the Chrome team working on this--the web is a huge collaborative effort--but you can't deny that Chrome is leading the charge here and this post represents a ton of our engineering effort, feature championing, and vision. blog.gitbutler.com/the-great-css-…

🚨 CRITICAL: Active supply chain attack on axios -- one of npm's most depended-on packages. The latest axios@1.14.1 now pulls in plain-crypto-js@4.2.1, a package that did not exist before today. This is a live compromise. This is textbook supply chain installer malware. axios has 100M+ weekly downloads. Every npm install pulling the latest version is potentially compromised right now. Socket AI analysis confirms this is malware. plain-crypto-js is an obfuscated dropper/loader that: • Deobfuscates embedded payloads and operational strings at runtime • Dynamically loads fs, os, and execSync to evade static analysis • Executes decoded shell commands • Stages and copies payload files into OS temp and Windows ProgramData directories • Deletes and renames artifacts post-execution to destroy forensic evidence If you use axios, pin your version immediately and audit your lockfiles. Do not upgrade.

@HiroKws @taylorotwell But with boost 1, you could still disable the tinker mcp tool to save a few tokens.

@taylorotwell However, the MCP for Boost needs some cleaning up. Tinker and last error aren't necessary for the MCP server because the model can directly execute Artisan commands and the CLI. That approach saves more tokens.

Laravel was so far ahead of the game with Laravel Boost tbh. Like almost a year ahead. Still so much work to do. The definition of who can be a Laravel developer is about to get very broad.

@HiroKws @taylorotwell For me, Opus often fails to escape tinker commands properly and ends up creating .php files... MCP make result way more deterministic.

@marcelpociot Did you consider adding FrankenPHP?

The next Laravel Herd update brings RustFS as a well-maintained MinIO alternative 🫡 herd.laravel.com

More than 100 participants; two prize pools for multi- and single threaded solutions; performance already pushed to sub-3 seconds. The 100M-row challenge is doing great! FYI: if you're with a company that wants to pitch in to biggen the prize pool, feel free to let me know :)

A pretty cool new feature in Tempest 3.1: generate TypeScript types from PHP code! tempestphp.com/blog/generatin…

A pretty big milestone today! We've released Tempest 3.0! tempestphp.com/blog/tempest-3

@epelboin @grok ça parle de quoi ce tweet? Véracité ?

Cutoff : en 24h le pays est à l'arrêt et on capitule. Ce serait la blitzkrieg la plus efficace jamais conçu. Ca fait plus d'un an que les militaires en parlent ouvertement dans des espaces pourtant très policés comme Linkedin et qu'on en parle dans toutes les conférence sur la souveraineté numérique. Les dégats pour les US ? Minimes, pas pire que le bordel en bourse qui a suivit l'instauration de tarifs douaniers. A un moment il va falloir faire avec la réalité.

@adamwathan Playwright

What's the most useful MCP server you have installed?

@jeffrey_way @christophrumpel You can also use Spec kit or Amazon Kiro

@christophrumpel Using the Ralph technique everyone is screaming about lately. bin/ralph.sh That reads a tasks file I have, it picks an item to work on, implements it, commits, and then repeats. It's incredibly cool.

I now create a bulleted list of tasks and issues to fix. An agent then works through them one at a time, while I take my kids to school. These are *wild* times.

What are your biggest complaints about using the web right now?

Say hello to a brand new kind of prop in Inertia: Once Props! 💨 Once Props are evaluated only once on the server and then carried over to the next page that has the same prop. This is ideal for shared data that rarely changes or data that is huge or expensive to compute.

Astuce #UX ! Ces racines de numéros de téléphone fictives, réservées pour le cinéma, peuvent être utilisées comme exemples de saisie dans vos formulaires : 01 99 00 02 61 91 03 53 01 04 65 71 05 36 49 06 39 98 cf. legifrance.gouv.fr/jorf/article_j… (article 2.5.12) #24JoursA11y #a11y

@FannyLAllegra Pour autant, je trouvais les crises plus impactantes dans le livre !

@FannyLAllegra Ils s'en sont très bien sorti pour rendre l'ensemble plus digeste !

Je vous avais dit dans le podcast que j'allais me lancer. C'est parti pour #Foundation !

@FannyLAllegra Je reconnais toutes ces qualités au livre aussi !

@FannyLAllegra Wow, j'en ai fini 2 et demi, mais le côté trop décousu entre chaque crise a eu raison de moi 😅