Yannick
@osteel.bsky.social
Backend consultant at Yellow Raincoat Ltd. Currently building pixelwatcher.io. Organiser of @phpsussex.uk. I post in French once in a while
I understand some of the feedback around Opus 5 now. It definitely is more conservative, in an annoying way. Random directives it picks from elsewhere now seem to have as much weight as my own instructions.
I’m trying to use Sonnet as my daily driver to save on token consumption, but it’s just too dumb. I end up wasting tokens (and my time) instead
I finally have some time to dedicate to PixelWatcher (I’m busy with a new contract, which is good!). I’m aiming for a public launch in the next few weeks (meaning there’s still time to join the private beta 👀)
Website Change Monitoring Made Simple | PixelWatcher
Monitor webpages and get notified about changes you care about. PixelWatcher makes website monitoring simple and affordable.
pixelwatcher.io
Always interesting to see how these companies use AI. I'm still not completely sure how to approach the review/verification side of things, but I totally agree on how it allows companies to scale more slowly
Replit — The Self-Driving Company
In the past six months, engineers at Replit have nearly tripled code output. Review times held steady. Reversions and product incidents have stayed flat. Quality metrics improved, and releases have accelerated. All the typical trade-offs you might expect have not occurred. While the code is the visible part, what's happening under the surface is much more interesting. Agents now investigate production incidents, review pull requests, answer questions, analyze business data, triage support tickets, research sales accounts, and improve the systems that power Replit Agent itself.
replit.com
I thought switching back to VSCode from Cursor would also mean fewer updates... I was wrong 🙄
I really wanted to like The Man in the High Castle, but I’m halfway through season 2, and I’m bored to death
If you're looking for a task for Claude to do semi-autonomously (and that you can monitor from your phone, for instance), try `/doctor`. It does a lot more than it used to, and will probably flag and offer to fix things currently flying under your radar
Claude Code tip: To maximise the number of 5-hour windows I get per day, I set up a dummy routine that runs at 5:30 am. It starts my first session before I wake up, so when I begin work I have a fresh session with a couple of hours left, and can fit three sessions into the day.
I'm currently trying to test more LLMs for PixelWatcher. Turns out it's very hard to find reliable providers: Scaleway times out a lot, OpenRouter is even worse. I can't even create an API key on evroc. Looking into Hugging Face now, but their interface is hurting my brain
On nous dit parfois qu'il y a beaucoup de talks Symfony au Forum PHP : c'est vrai, on adore ! Mais nous sommes aussi ouverts aux talks orientés Laravel, TYPO3, Drupal, WordPress... Parlez-nous de la diversité de PHP, répondez à l'appel à conférences du Forum PHP 2026. cfp.afup.org
I don't know if it's because I used Fable 5 for a couple of days last week, but Opus 4.8 feels incredibly dumb today.
A great summary of the current geopolitical and macroeconomic environment, and how to navigate it from an investment perspective
June 2026 Newsletter: The Wild West
June 4, 2026 Most modern investing frameworks were built in a period of expanding globalization, broad institutional trust, disinflationary growth, and U.S.-led geopolitical order. However, for years now we have been moving into a more fragmented environment: fiscal dominance, trade conflict, AI-driven information disorder, and weaker international coordination. This newsletter issue discusses the “Wild West” […]
lynalden.com
I took part in the State of PHP survey by the PHP Foundation and PHPStorm – join now too and help us map the herd! 🐘 jb.gg/stateofphp26 #StateOfPHP
The State of PHP 2026
jb.gg
One reason I’m transitioning to the role of product engineer. Product engineers are accountable for outcomes, however they get there. I don’t want my value to be tied to specific inputs while AI is reshaping those inputs.
What it feels like to work with Mythos
Claude Fable represents another big jump in AI
oneusefulthing.org
Anyone else struggling to deploy in Laravel Cloud at the moment?
I’m admittedly getting a little confused around the difference between dynamic workflows, loops and goals 🤔
Only a handful of registrations again. I'm currently resorting to DMing people directly 😅 I really want to keep this going, but my motivation also only stretches so far 😮
PHP Sussex@phpsussex.uk · 2mo ago📢 Our next meetup is this Wednesday! Christopher Miller will explore some practical use cases of Event Sourcing and show us different ways to introduce the pattern into existing codebases. 📅 Details and registration 👇 www.meetup.com/php-s... #php
Fine-tuning my slides ahead of Laravel Live UK later this month 👀 🎟️ Grab your tickets: app.tickettailor.com...
Early feedback on Opus 4.8: it seems to be way less assuming overall, which is definitely a good thing
I'm finally using a task tracker for PixelWatcher (I was using a notebook until now 😅). I settled on GitHub Issues, since I'm already using GitHub anyway, and I don't feel the need for something like Linear at the moment. Claude can talk to GitHub via `gh`, which is good enough.
Anyone following the AI space has come across the term "evals", usually in the context of training models. But what do evals look like on the application side? This article covers why I ended up needing them for PixelWatcher, and how I implemented them:
What are Evals Anyway?
AI workflows can't always be tested deterministically. As AI finds its way into our applications’ logic, a different testing strategy is needed: Evals.
yellowraincoat.co.uk
Anyone following the AI space has come across the term "evals", usually in the context of training models. But what do evals look like on the application side? This article covers why I ended up needing them for PixelWatcher, and how I implemented them:
What are Evals Anyway?
AI workflows can't always be tested deterministically. As AI finds its way into our applications’ logic, a different testing strategy is needed: Evals.
yellowraincoat.co.uk
Pretty neat and straight-to-the point explainer of the difference between jobs, actions and services, with good examples. Only thing I would change is make the action invokable medium.com/@radenadr...
📢 Our next event is up! On Wednesday 10 June, Christopher Miller Christopher Miller will show us how to time travel through Event Sourcing 🔮 📅 Details and registration 👇 www.meetup.com/php-s... #php
June | Event Sourcing Your Way out of Trouble, Wed, Jun 10, 2026, 6:30 PM | Meetup
Spring is here, the Spiegelten– er, Gardens are up, and our last PHP Sussex before the summer break is in! On Wednesday 10 June, returning speaker [Christopher Miller](htt
meetup.com