skeptrune

I read this as napkin math[1] for Kube and thoroughly enjoyed. You can only find the important numbers relative to performance and scaling by trying to accomplish some kind of goal. Benchmarks are mostly bikeshedding.

[1]: https://sirupsen.com/napkin

Probably not for dedicated SaaS, but it makes sense for these new AI app builder products I think.

Claude code anonymous meetups are coming to your local area very soon lol.

This is cool! I'm unfortunately not skilled enough to be a user, but enjoyed poking around the repo and seeing how things worked.

I loved this! Good writeup and very mature response to lots of criticism they took online prior.

Wait until they turn ads on.

What a delightful piece of realistic fiction. I was very entertained and amused.

What was the motivation for making this work in tmux?

Yooo! Nice, I need to set this up. Working to get the pings will be super motivating for my new project.

I needed this. I have some level of intuitive understanding, but this reference will be great whenever I'm not quite sure if something is possible or not.

Well that's what I implemented. There are markdown docs for every HTML file and the proxy decides to serve either markdown or HTML based on the Accept header.

I learned that the golang CLI[1] is the best through my work simplifying Firecrawl[2]. However, in this case I used one available through npmjs such that it would work with `npx` for the CF worker builds.

[1]: https://github.com/JohannesKaufmann/html-to-markdown

[2]: https://github.com/devflowinc/firecrawl-simple

It is crazy how badly Tailwind bloats HTML. Tradeoffs!

Oh hey, I work at Mintlify! We shipped this as a default feature for all of our customers.

It's always better for the agent to have fewer tools and this approach means you get to avoid adding a "convert HTML to markdown" one which improves efficiency.

Also, I doubt most large-scale scrapers are running in agent loops with tool calls, so this is probably necessary for those at a minimum.

Agreed! I love that such a tried and true web standard is making a comeback because of AI.

Exactly! It can actually be a positive thing, might as well make it easy for LLMs to read.

Was this updated recently or something?

This isn't true. ChatGPT and Gemini link to sites in a similar way to how search engines have always done it. You can see the traffic show up in ahrefs or semrush.

There was a lot of conversation about this on X over the last couple days and the `Accept` request header including "text/markdown, text/plain" has emerged as kind of a new standard for AI agents requesting content such that they don't burn unnecessary inference compute processing HTML attributes and CSS.

- https://x.com/bunjavascript/status/1971934734940098971

- https://x.com/thdxr/status/1972421466953273392

- https://x.com/mintlify/status/1972315377599447390

There is an untapped market for phone-oriented IDEs. It's a super convenient way to code when you're out of the office or on the go.

I'm looking forward to my company retreat, but we are still less than 30 people. I would certainly think multiple smaller retreats are better than a big one for a company of this size.

Amazing! I have been desperately wanting this. Livecaptions doesn't seem to be maintained super well.

Huge win! Excited for better perf

yeah, lots of bitcoiners

If there's a no-code way to edit content and AI can work with the rest then I think it's fine.

It does, there's an extension you can install which is even a bit better.

Keystatic[1] is awesome. It's local-first by default and made for static site builds, so you can keep costs low by not using a server.

I've used it alongside Astro for both my personal blog[2] and two company sites[3][4]. It's worked amazingly well each time.

[1] https://keystatic.com/

[2] https://github.com/skeptrunedev/personal-site

[3] https://github.com/devflowinc/trieve/tree/main/website

[4] https://github.com/patroninc/patron/tree/main/website

It's interesting that VSCode has its own Claude code extension which works well and does most of this.

Zed has a lot of these micro-battles ahead where it has to spend money building solutions that VSCode's community shipped without their core team putting in any effort at all.

I felt seen after reading this post. NextJS was really good somewhere around v11 when it was first becoming stable and create-react-app was unusable. It's been rapidly enshittified since then.