Partly in the Right - intertwingly.net/blog/2026/0…

You are right but it also feels short-sighted? You have the most powerful tool you ever had, you have become a 1000x maker, and you can't think of how to make your stack 10x better? Yes, Rust is fast, the type system catches bugs, but there is so much beyond it. * Why not make Ruby 10x faster? * Why not make Ruby (and therefore Rails) really great at concurrency? * Why not add a banging type system to Ruby that can provide guarantees beyond Rust? * Why not embed SQLite3 into Rails with a replication protocol so you can really compress the stack beyond a single server? * How to preemptively fix most bugs from the most critical parts of Rails, such as database pools? Yes, who cares about ActiveSearch, but there are so many ways we can fundamentally improve our stacks that go beyond asking the agent to just crap out code. Phoenix is still the only framework that ships with presence, because it turns out it is deceptively a hard problem to solve correctly, but somehow all issues are solved? I understand some folks have no interest in doing any of the above, and that's fine, but then at least pass the baton to the people who want to do exactly that. PS: I say all of this with love ❤️

Working on a few of these with @matz: Rails compiled to a native binary via Spinel, whole-app type inference with no annotations, and one binary that holds the WebSockets — no Redis. Campfire's own test suite passes on it: rubys.github.io/roundhouse/a…

Programming languages that bring meaningful design choices to the forefront and hide the noise are just as good for agents as they are for humans. Semantic compression matters. Using the right tool for the problem at hand has never been more important.

I honestly don't think that would change the conclusion. rubys.github.io/roundhouse/p… shows the meta-programming free version of the blog application in Ruby, C (clean or otherwise) would be more. intertwingly.net/blog/2026/0… goes into more depth.

What I have is a compiler, built by Claude under my guidance. The compiler is an executable, deterministic, uses no tokens. The compiler can be compiled to wasm and run in your browser: rubys.github.io/roundhouse/p…

Not Claude loops, but take a look at this page: rubys.github.io/rubyconf-202… ; where you see a programming language you like, click on that name and you will download a tgz with a README. Try it.

In the future, frameworks will become languages. These days every language has a socket library, primitive HTTP support etc. A compiler that understands models, views, and controllers can do more. I picked Rails, but I hope others follow.

> その意味でRailsこそが仕様だと言っていると読み取った Exactly. Go to github.com/rubys/roundhouse and look at the "about". You will see it in the image below:

Try different apps and different targets on rubys.github.io/roundhouse/p… ; the expansion is significant. Whether human or LLM, it is easier to reason about and make correct code that is an order of magnitude smaller. Context windows matter.

I go the other way. SQL is the best example. Sure you don't need a database, you can generate the code to replace it, but why? Databases are battle tested and optimized.

Some benchmarks: rubys.github.io/roundhouse/b…

Sam Ruby's Roundhouse transpile Rails app to plain Ruby (or other languages) without metaprogrammig. And Spinel can compile transpiled Ruby code. Not all Rails apps yet, but at least we can compile Campfire and Lobster. See github.com/rubys/roundhouse

Matz response (to Rust) is Spinel github.com/matz/spinel

Previous Deccan Ruby or next one? What if you could keep the language and replace the runtime? Explore Blog and Campfire here: rubys.github.io/roundhouse/ ; try Elixir: rubys.github.io/roundhouse/b…

Both LLMs and humans have context windows. Additionally when you write in a lower language you are specifying "how" not "what". My thesis is that Rails is like SQL, and what it needs is to be treated as a language.

I might be wrong, but I would think that enterprises would be interested in throughput and memory usage. Roundhouse (and perhaps Spinel) might not be ready for production just yet, but I would think would be worth watching.

I'm hoping that the topic of Spinel comes up. :-)

A PR against Campfire, merged by @rosapolis, and Campfire on Spinel now passes 300 out of 300 model and controller tests: intertwingly.net/blog/2026/0…