So far, our neural network can make predictions, but how does it know if those predictions are any good? That’s like having a student take a test but never getting it graded. Without feedback, there’s no way to improve. This is where loss functions come in – they’re essentially the grading system for your neural network. by Christopher Miller
When my eldest child was born, I didn’t expect it to change how I thought about software accessibility completely. But within a few weeks, I found myself doing nearly everything with, at best, one hand and at worst, my teeth or feet. by Scott Keck-Warren
Hello friends! Welcome to another Yelling At Clouds column, where I continue on with Grumpy Testing Patterns. As I write this, PHP Tek 2026 is about a month away, and even though I am unable to attend this year, I highly recommend checking out any PHP-related conferences that you can make it out to. Conferences were what allowed me to network and make connections that helped me build what has been an almost 30-year career. by Chris Hartjes
In Part V, we made session continuity boring. A user can survive node loss, drains, and routine deploys without getting kicked back to the login page. That is a meaningful step forward, but it still leaves the most common availability bottleneck untouched: the database. by Wendell Adriel
I shipped a Laravel application last year that nearly cost me my AWS account. by Eric Mann
There’s a frustrating experience most API consumers have had at some point. by Steve McDougall
A heavy database server may be overkill for local applications and medium-traffic applications. Why configure, manage, and monitor a service if you don’t have to? In this article, I’ll show you how to use SQLite in your PHP application to get a full-featured database engine without all the hassle and overhead. by Oscar Merida
Every PHP developer has used a query builder. Eloquent, Doctrine DBAL, Laravel’s DB facade. They are convenient, safe, and ubiquitous. So why would anyone build one from scratch? by Marian Pop
Check out Ben Ramsey’s fun recollection of this year’s Tek. by Ben Ramsey
Responses and Pingbacks
July 10th, 2026 at 7:55 am
Yeah, the loss function analogy really nailed it for me – never thought of grading a neural network like that before. Definitely going to keep that mental model when I’m debugging my next model.
July 20th, 2026 at 10:32 pm
Ah yeah, the loss function bit really clicked for me – totally makes sense that you need that feedback loop or you’re just throwing darts in the dark. And the accessibility point about doing everything one-handed hits different when you actually think about it.