adenta

I was only getting like 200 tk/s with groq on K2, was expecting it to be faster tbh.

I think the bottleneck is file read/write tooling right now

It's too expensive!

The problem I have with shad is that it’s react based, not HTML based.

So if I want to use shad, I can’t mix and match static HTML partials in with app code. To maintain a consistent theme, I would need to make literally everything react.

Hmm, weird, maybe a bug in HN? weird this is a message instead of in the body.

Body I built Chuck, a Rust-based CLI tool that makes it easy to extract and contribute template-worthy commits from your project back to the upstream template.

I feel like we are heading towards a world with more boilerplate, and I built this to help me manage my boilerplate. See https://contexteng.ai/p/context-engineering-is-only-half for more thoughts on scaffolds and boilerplate!

This looks cool! What would be an alternative product that this could replace?

What problem was created?

Is there an expected date we can start doing industrial scale desalination with excess electricity?

I wonder if a nice middle ground would be: - recording the playwright behind the scenes and storing - trying that as a “happy path” first attempt to see if it passes - if it doesn’t pass, rebuilding it with the AI and vision models

Best of both worlds. The playwright is more of a cache than a test

nope, should I? Or should I try your thing?

To me "parametric" features seem like a nice to have vs a hard requirement. I just want to be able to talk at my computer and get back an STL to spec that I can print out

I've had some success with openSCAD and this MCP, feel free to ping if that's helpful context

https://github.com/jhacksman/OpenSCAD-MCP-Server

What's new here? I thought this already worked?

Brad, what’s your take on Rails 8 w/ SQLite? Are you gravitating towards it these days over Postgres?

I wonder if I could use this locally on my macbook. The finder applications built-in search is kinda meh.

What CI provider?

I think there are diff oil patterns on the lanes that change where you want to aim.

I tried cursor a couple months ago, and got the same “do you want a refund” email as others, that got a “sure” reply from me.

Idk. It’s just growing pains. Companies that grow quickly have problems. Imma keep using https://cline.bot and Claude 3.7.

You’ve promised a ton of people refunds that never got them. Others in this thread, and myself included

Edit: he did refund 22 mins after seeing this

> But on a modern cloud platform, there’s really no reason to permanently grant Sam Altman Google Mail access, even if you want his robots to sort your inbox. You can decouple access to your mail spool from persistent access to your account by tokenizing your OAuth tokens, so the LLM gets a placeholder token that a hardware-isolated, robot-free Fly Machine can substitute on the fly for a real one.

This kind of isolation is nice. Next logical step would be some way to prevent Fly from having access to the tokens.

This is exciting!

Glancing through the article, I can't tell, is this Cursor specific? Some of us are raw dogging VS Code with https://cline.bot, which supports MCP servers: https://cline.bot/mcp-marketplace.

Would love to try this out in Cline!

Yeah I don’t use my hands on the steering wheel when the comma is engaged. Haven’t spent much time in Tesla’s to know how it compares, someone else might know better.

It almost took an exit I didn’t want to take in Louisiana, other than that no major issues. Works well in rain.

You are still driving the car, it’s just less intense.

I swear by my comma 3. Orlando to Austin, 45 minutes hands on the wheel.

praying for Red Alert 2

I couldn't get https://docs.libretro.com/library/remote_retropad/ to work! its designed for game pads to control the emulator. I was hoping I could repurpose it to just send arbitrary key commands over a network interface, but nothing I tried worked. When I asked in the retroarch discord, they couldn't figure it out either lol.

The vision models still struggle in my experience. I got around that by reading the RAM and describing all the objects positions on screen

I think the “wa” prefix means “bad” in Japan

Email me when it launches! (In profile)

Yeah I think _a_ future of television might've been more apt. They should've made a season 5 to Snowpiercer.