Did Svelte finally solve state management?

Happy Friday! This is 7.dev. We’ve freed an Nvidia A100 chip from the grasps of ChatGPT and we’re using it to generate non-stop, continuous full-stack development news for you, every minute… forever…

…just kidding. This one is hand-crafted: no LLMs present. Just good ol’ fashioned writing from me (hello!)

In this issue, we’ll cover some recent news in state management (everyone’s favorite front-end development topic), tips for working with ChatGPT, and much more.

Real quick—if you missed it, we just released a free course on GitHub Copilot. It’s doing pretty well on YouTube right now, and if you haven’t had a chance to try Copilot, or just want to see how it works, you can watch it here.

(Interesting story… I A/B tested five different titles for the video before the final title, “Mastering GitHub Copilot for Beginners”, ended up performing the best. Who would’ve thought A/B testing YouTube video titles was a thing?)

One more thing: I love podcasts. If I check my stats in Overcast, I’ve saved 343 hours by using their Smart Speed feature. That’s a number that I find a little disturbing, and try not to think about too much.

If you also like podcasts, you should check out the first two episodes of the 7.dev Podcast. Every Wednesday, I’ll be covering hacks, tips, and tricks in the full-stack development world. You can also search “7.dev podcast” on anywhere you find your podcasts!

Did Svelte finally solve state management?

In a blog post this week, the Svelte team announced that Svelte 5 will include runes, a nifty feature that promises to simplify how you can update variables and handle reactivity in your applications:

> Runes are symbols that influence the Svelte compiler. Whereas Svelte today uses let, =, the export keyword and the $: label to mean specific things, runes use function syntax to achieve the same things and more.

You can read more details in the blog post, and while the team hasn’t announced a release date for Svelte 5, it looks to be a great improvement to an already quite successful project.

Using custom instructions in ChatGPT

I saw a tweet this week from Kent C. Dodds covering custom instructions in ChatGPT that I found super interesting. Custom instructions is a new feature in ChatGPT that allows you to set sane defaults for all of your interactions with ChatGPT.

For instance, Kent is using the custom instructions to indicate that whenever he interacts with ChatGPT, it should return code in a certain format and preferring certain languages. It’s a neat trick!

Odds and ends 🤔

• Wiz Research announced that Microsoft AI leaked 38TB of private data — Teams Messages, passwords, and more — as part of a large set of training data available for training AI models. Are training datasets the new cool way to leak private information?

• Vercel announced Next.js Conf 2023, which will take place on October 26th. The conference will take place in San Francisco and also be available for streaming online. You can get your online ticket using the link above. What do you think they’ll announce?

• Allow me one nerd gamer moment in this newsletter: if you grew up playing the N64, you may enjoy Ship of Harkinian — a fan project that has totally decompiled and rewritten the Zelda game Ocarina of Time (one of my top five games ever) from scratch (!) with support for 60+fps, tweaks, and more. It’s pretty remarkable, and amazing what people can do with some spare time and open-source collaboration.

Check us out on the socials 💅

Every week day, we publish a video on Instagram, TikTok, and YouTube. If you want more 7.dev than just this weekly email, follow us!

Here’s some of the stuff we published this week:

• 5 programming books you should read

• Two Git tips that will save you from EVER screwing up your projects

• Hack the Full-Stack Dev Roadmap: Junior Backend (podcast episode)