After considering elaborate ways to make TinyMCP support different return formats, finally arrived at this "genius" 3 line solution, which brings you version 0.2.0. Keep it simple folks. https://github.com/maxim/tiny_mcp/commit/5a1cb8c9f6c04ca84d45532dc2a61e41dc0f7121#diff-cbd58482f8350668e4356d79e78e0220b1e3d1cce59968abd923f394472b6fbfL118-L119

For example, if I'm in a repo, use the top search bar, and find nothing after a couple of attempts, I cannot go back to the repo I was in without pressing back multiple times.

UX issue on Github: if you search, there is no way back to where you were before. You must either back-back-back in the browser, or enter the original url again. It's interesting, because it feels super annoying in practice, but actually not obvious from webdev standpoint.

Is it now _outdated_ to care about code quality minutia, and not losing my ability to focus insanely well on little details? I guess all of you folks in the "syntax doesn't matter" camp should be happily embracing LLMs. Working code is all that matters, right?

Even 1-shot-correct code by Claude 4 is subtly bad. Gotta read it with fine tooth comb when project matters. What's worse, "instructing LLM" brain mode feels mutually exclusive with "writing code" brain mode. It hurts to switch (and I'm an easy context switcher).

Published my local tiny_mcp server that I use to create quick simple MCP tools in Ruby. https://github.com/maxim/tiny_mcp

@henrik for me, this is where the habit of reviewing the changes comes useful. I re-read all changes in Fork.app before committing (in projects that matter). Allows me a fresh look. However been also playing with writing MCP tools for myself, maybe your idea can be made into mcp. I wrote a super simple way to make custom mcp servers in Ruby and run them all with single command: https://gist.github.com/maxim/e6d024892031ed526eb270c7baebe6c1 - quick and dirty for now, but with usage comments. Plan to publish somehow maybe.

@henrik Zed is nice, also tried it a bunch, but I can’t bring myself to switch editors for some reason, too attached. Kept my projects opened in Zed for months, used their agent stuff a bunch, but always returned to ST for focused peaceful coding. Similar story with VSCode. Now with Claude Code maybe I don’t have to switch. 🤞

@henrik did you try Claude Code? I just tried it last night and it seems like a huge upgrade compared to simple ad hoc stuff. I was trying aider before, but Claude code is night and day, so much better. I love that I can keep my editor (sublime) and have it in terminal side by side.

@henrik I wouldn't call this vibe coding. Turns out the term was originally intended to have a more narrow and reasonable meaning: https://x.com/karpathy/status/1886192184808149383.

Useful things to do with agents that aren't vibe coding:
- Converting parts of code to something.
- Asking opinions about approach.
- Asking for reviews of specific parts.
- Making non-trivial corrections in parts of code.

All totally helpful and useful in my opinion, and not what I'd put under vibe coding umbrella.

Impressive how easy it is to write an assistant that reads and answers questions about a >1200 page PDF with RubyLLM. Less than 20 mins of work. https://max.engineer/giant-pdf-llm

Is it just me, or reading text in a different style/format than it's written causes the "fresh eyes" effect? That's why I like to review code and blog posts in a different viewer than my editor. It's as if I'm reading it a day later, without actually waiting a day.

Every "services" gem: do you dislike this code? [Shows a plain ruby procedure in a controller action]. Here's how you could write it instead: [Shows a lambda-heavy mix of actors, executors, policies, contexts, and almost the same procedure in the controller action].

Have you ever had to write tests for error handling in unlimited-retry active jobs? Trying to find a way: https://discuss.rubyonrails.org/t/testing-activejobs-with-retry-on-attempts-unlimited/87724

What a great talk to watch at 1.5x. Clear and concise practical walkthrough of async features of Ruby and Rails. Well done! https://youtu.be/CfC2blkrkys

When you get lucky that the only available size of onesie is the correct one.

git absorb is actually incredible. Saves so much time and adds freedom to how I work. But don't miss the not-entirely-documented -F (--one-fixup-per-commit) flag. You probably gonna want it. https://github.com/tummychow/git-absorb

@zenspider @codefolio @rbates my god, I’ve been carefully doing this by hand for years. Does it make mistakes at all in your experience, like absorbing not the way you intended?

@henrik @n3bulous let me know if you find something good. I switched my blog to paid Commento a while ago, but while it works, it’s bare bones, and feels mostly abandoned.

@henrik that was my favorite style in the linked SO post too. But I like the namespace for grouping and the ability to work with keys. This way I could do Item::State::NEW