autocole

Thank you for the write up! Really interesting. I may consider some of these techniques for a personal browser extension

Cool starter kit! What are your thoughts on Next.js's dev experience vs potential vendor lock-in? Is it worth considering how tied Next.js is to Vercel?

I would love to see this kind of investment. I’m still hopeful for Svelte 5 becoming a language that LLMs understand better and this would complement that future

Thank you! I have been wondering about this _so_ much. I’m glad to have some clarification. Have you written about your experience setting up the tests and writing the evals?

What is your overall impression of Cline? Why does it stick out as being on the SOTA track compared to the others trying to replace VS Code entirely?

It seems a little silly to not support some of the small additions TypeScript contributes when it could just be another flag to enable enum, namespace, constructor to property init support

Really cool write up. I love seeing a nice introduction for ECS where it’s so clear where the benefits are

Thanks for sharing! I think I started working on large teams around the same time using Angular 2 (shudders).

It’s so nice to see the learning from a multi year UI development project and the evolution of how to communicate good patterns across teams and people. It usually turns into 50% technical thinking & 50% political respect to ensure others help champion your ideas with you

Inspiring! Thanks for sharing

Nice move! With so many tools existing in this space, I want to read more about the rationale and long term vision for the project

This personalized agent is kinda how we think about things for https://forethink.ai

neat project! I love the example of how you can improve the API with the types provided by Rust. The part I didn't understand ws how this is a "port" is it literally a reimplementation of diligent Graphics Engine or is it bindings? Or is Diligent Graphics Engine a set of c++ bindings into something else and you're peeing the bindings?

I'd like to see a long term of showing both the workflow design next to the example codebase with links from the diagram into the files in the codebase. I'm hoping we see more like this become available to help everyone learn the design for AI pipelines

Interesting, but it doesn't seem to go far enough to be something I'd consider in place of my existing synchronous workflow with Cursor or async workflow with a Devon-like tool. I'd love to have seen a more in-depth example of building a tool like type checking for linting or a tool like "feature guideline + impl workflow".

I wasn't sure what this was going to be from title alone, but I appreciated the tour of practical tools and libraries available in Gleam. They could totally take some notes from how Bun does their marketing by empowering & compelling fans to start new small projects in Gleam by leveraging these libraries

So this is like Capacitor/Cordova/Phonegap?

Can you describe how you used LLMs for deployment? I'm actually doing this exact thing but I'm feeling annoyed by DevOps and codebase setup work. I wonder if I'm just being too particular about which tools I'm using rather than just going with the flow

I love this. Congrats on the launch. You all are always building something interesting

I like the terms multiplayer and collaborative

My magic wand wish is that they publish a separate rollup plugin for the rune syntax only, then allow me to write my ui in another jsx based framework

JSON_TABLE function as a way to do SELECT on a json value is pretty cool! I am curious about common patterns where people plan to use that

I’m not sure I understood the author’s point. This write up sounds like a collection of thoughts and observations about trends and the now saturated landscape of “build in public”/“indie hackers”

I’d like to have heard some perspective from people actively participating in this and how their experiences have been

I love the simplicity, but I would be hesitant to try it since I don’t know what 10MB of logs looks like. I would be interested in trying this if it could be set to auto-evicting past 10MB so I could start to understand that boundary better.

Unseemly lang is another project that goes the other direction, but I don’t remember if it has balanced brace requirements https://github.com/paulstansifer/unseemly

Very enjoyable read, and this single supervisor is similar to how I set up an application at a previous startup, where we unwrapped everything. This reminds me of one of my favorite posts https://medium.com/@mattklein123/crash-early-and-crash-often...

Can it be addressed by chunking a response into parts that can individually be checked?

I'm using the Forethink web extension for that. It keeps a local index of every page I visit for searching

https://forethink.ai

If you haven't seen a team using Feishu/LarkSuite, then I'd make an effort to see how that works. It's essentially a super office app that does the whole "everything is integrated" significantly better than Microsoft and Google have with their office products. For example, you can share a document during a call and interact with others inside the live document.