A Tasty Treats NEWSLETTER for Web Developers

Issue #28:

Do we care that LLMs are actually worse than humans at summarizing text?
A better file manager for macOS and a bunch of other tools to make your life easier
The Syntax Crew is coming to SF for GitHub Universe and we're hosting a meetup

Effortlessly build surveys with markdown

scott's find

blocks.md is a tool that takes your Markdown files and turns them into forms and web pages that are beautiful, customizable, accessible, and fully localizable.

Full Stack Cloudflare RAG

One from ben

Ever wanted to build or deploy your own ChatGPT-lite clone? Get started with this open source app, built entirely on off-the-shelf Cloudflare services.

CSS Triggers

HAND-SELECTED BY Scott

A reference for the performance implications of CSS properties.

Labs.google

Kaitlin's link

"Experiment with the future of AI" -> new tools, features in old tools, games, experiments and more.

All Hail The Algorithm

Some good stuff™ we posted this week

The Feed

New episodes 3×/week + YouTube!

8000 ESLint Errors × HTML Time Tag × 7 Meg React Bundle × CSS Modules

Fun & Profitable Side Projects for Developers

CJ × Hosting Meetups - Lunch and Learn

You Need These 30 Apps - PART 1

SHAMELESS PLUG

[Live Workshop]

Faster APIs, Better Experiences: Debugging Next.js

Join Sentry on September 15th for a hands-on workshop with Dan Mindru, co-host of the Morning Maker Show, where he'll share how he shaved 22.3 seconds off an API call using Sentry. You'll pick up practical tips on how to debug errors and performance issues using Sentry’s Tracing and Session Replay.

RSVP

Jacob

@fat

This summer we went into a hole and rewrote the bulk of Pierre off of RSC and onto a local first sync-engine called Replicache.

It's… really fast.

11:44 AM • Sep 10, 2024

Retweets

495

Likes

Read 20 replies

RSC or Replicache?

The team at Pierre ditched React Server Components and rebuilt their UI using Replicache – see the results

Gergely Orosz

@GergelyOrosz

I’ve come to abandon the use case for LLMs to summarize text because the summaries usually miss vital parts and are disappointing when I look closer. Australian government now has proof this is the case.

Another proof that LLMs have more limitations we need to understand.