Scott Spence

@scottspence.dev

🀝 Svelte Society London co-founder 🧑 Svelte ambassador πŸ“” I blog what I learn over on scottspence.com πŸ“ˆ Product Engineer engineering products πŸ› οΈ Built svortie.com | devhub.party | skykit.blue | sveltest.dev πŸ“ UK

So excited for this β€” the Svelte community is so fun and idiosyncratic and welcoming and if this is anything like the last two conferences it will be an absolute blast. And Ljubljana is lovely. Submit a talk idea! Join the list! Would love to see you there

Svelte Society@sveltesociety.dev Β· last wk.

Svelte Summit tickets are coming in the next few days 🎟️ Join the launch list and we’ll email you a discount code when they go live. See you in Ljubljana on November 19–20. β†’ sveltesummit.com/ticket-updates

Just learned last night one of my old high school cross country teammates lost both his kids (and ex-wife) to a devastating act of violence. Our community is raising money to cover funeral services and help John with as much support as we can. www.gofundme.com/f/honoring-a...

Donate to Honoring Abigale and Dekota’s Memory, organized by Jenny Bayes

2 days ago, in a tragic act of violence, John lost both of his children at the hands … Jenny Bayes needs your support for Honoring Abigale and Dekota’s Memory

gofundme.com

Is it just me that finds gpt-5.5 better on low thinking than high/xhigh? I find high/xhigh will over engineer and over correct all the time whereas low thinking just does the job I've started using low reasoning sessions as coordinators for teams now as they don't faff around and stay on task

Latest version of sharp (for Node.js image processing): πŸ“¦ Faster and safer installation, no more install scripts πŸ™Œ Dual ESM/CJS πŸ–ΌοΈ Improved AVIF output quality and bit depth β˜€οΈ Experimental support for HDR gain maps πŸ‘· Transferable ArrayBuffer output for worker threads www.npmjs.com/package/sharp

npmjs.com

I’m getting spam emails from an email address I only use for Paddle So, they have sold the email list or there’s been a data breach? This is the number one reason I use masked emails for everything I sign up for

Belw through my OpenAI allowance for the week and for the next three hours πŸ˜… I’ll be honest results on high reasoning have been on par with low reasoning but slower πŸ˜‚ Anyway I’m trying out OpenRouter fusion now