Loren Stewart | Cheddy

@cheddybop.bsky.social

https://www.lorenstew.art/ https://cheddy.chat/

Chinese company Z.ai released their new open weight model GLM 5.2 While it’s not quite as good as the latest Claude or GPT, it’s very close, as good as those were just a few months ago And it’s open-weight so anyone can download and use it

Bar graphs of model performance, GLM is very very close to the other bars although a hair below

wow — gang, we gotta get those sveltekit numbers up npmx.dev/package/@sve.... click the 🧡 in the top right! don't let @danielroe.dev embarrass us

@sveltejs/kit - npmx

SvelteKit is the fastest way to build Svelte apps

npmx.dev

Philippe Serhal@philippeserhal.com · 2mo ago

New on @npmx.dev: 🥇🥈🥉 npmx.dev/leaderboard/... Why? Honestly, because we can just do fun things on the web. 🍿 Will the crown get to @rich-harris.dev's head? Will @danielroe.dev point out that ackshually SvelteKit is a fairer comparison and look how low the poor thing is? Will Astronauts organize?

There are two kinds of "framework-agnostic": 1️⃣ Agnostic at runtime → ship your own reactivity, adapt it into every framework 2️⃣ Agnostic at build time → compile the framework's native primitive into the core Almost everyone does 1️⃣. It's why you get a subscribe layer and a bundle tax.

Should we drop a dedicated open-source project comparing ALL major schema libs? 🤔 We could test: ⏱️ Initialization speed ✅ Validation speed ⚡ Parsing speed 📦 Download time (bundle size) If this gets traction… we launch it 🚀📊

Fabian Hiller@fabianhiller.com · 7mo ago

Just updated the @valibot.dev comparison guide to also including the new Zod Mini. Bundle sizes for a typical login schema: 🤖 Valibot: 1.37 kB 💎 Zod: 17.7 kB 💎 Zod Mini: 6.88 kB

Researchers have found two new vulnerabilities in React Server Components while attempting to exploit the patches last week. These are new issues, separate from the critical CVE last week. The patch for React2Shell remains effective for the Remote Code Execution exploit.

The death of browsing is part of the reason art is the way it is now. Our opinions are largely fed to us by algorithms. Spending a spare 15 minutes wandering around a bookstore or comic shop or video rental place was how you found stuff you wouldn't ordinarily pick up and thereby expanded your taste

violet allen@blipstress.bsky.social · 8mo ago

Bookselling is like the most "people go to the store and buy what looks cool to them without a particular agenda" type business left, and your purchases have a huge influence on what is ordered, what is displayed, and what is recommended.

Andy is a legend, and the course he has produced alongside some brilliant people including his team at @piccalil.li are top notch. If you have a project in mind for 2026, hit up @set.studio And/Or maybe make upskilling your Holiday treat to yourself this December!

Andy Bell@bell.bz · 8mo ago

A bit of a vulnerable post by me but we need your help to get us through what has been a more difficult period than 2020. It would be really appreciated if you could read and share in your networks too 💛 bell.bz/its-been-a-v...

One of my hopes for 2026 is that browser vendors take more responsibility for today's pervasively bad experiences. Nobody wants a slow site, but they also don't want to dismiss a mailing-list pop-up that blocks scrolling. Browsers need to step in: infrequently.org/2025/11/perf...

The Performance Inequality Gap, 2026 - Infrequently Noted

Embedded in this year's network and device estimates is hopeful news about the trajectory of devices and networks. It has never been easier to deliver pages quickly, but we are not collectively…

infrequently.org

Watch @atila.io demonstrate SolidStart's features like "single-flight mutations" and unveils the SolidStart 2.0 roadmap, which involves a deeper integration with Vite's native APIs and a shift towards a more modular, primitive-based architecture Full video: youtu.be/L0w0bnwzjGg