Roman Samoilov | Rage

Why is Ruby losing ground? It's not because the language is "dead" - it's because our tools haven't evolved to meet modern performance and stability needs. I built Rage to give Ruby teams the missing pieces they've been looking for in other ecosystems. A thread on the mission behind Rage:

Exploring something completely new for Rage. It’s still months away, but I’m already super excited!

@SergioRocks Hey ChatGPT, can you share the recipe of an apple pie? The constraints have moved, but not in the way you describe - the hard parts you outlined have always been the hard parts. It’s never been about the code.

Ticket-driven software development is slowly dying. For years, the workflow looked like this: - PM writes the ticket - Engineer picks it up - Engineer ships the code Repeat. That model made sense when writing code was the bottleneck. But it isn’t anymore. AI tools can generate large chunks of implementation. Scaffolding, tests, refactors, even entire features. The constraint moved. The hard part is now: - Choosing the right problem to tackle - Structuring the system correctly - Deciding what not to build That is why engineers are increasingly expected to own outcomes, not tickets. Instead of: - “Implement this API endpoint.” The work becomes: - What problem are we actually solving? - What’s the smallest thing we can ship that users will use? - How do we know it worked once it’s live? The engineers who adapt to this shift will thrive.

How much are you paying to handle 5K concurrent SSE streams + 5K HTTP DB requests in your Ruby app? I just did it on a €12/mo server using Rage. You don’t have to leave the Rails ecosystem to get Go-like performance.

I’d argue it happens daily in large monoliths. The more complex the codebase, the faster the idea of figuring out what’s inside the variable by looking at its name crumbles. Retrofitting types is a massive, error-prone headache. Which is why most (non-Ruby) teams treat types like tests - embedded from day one.

@codewithrage @brian_scanlan This kind of thing happens once in 5 years

A successful Rails codebase that ships reliably for >10 years will eventually attract multiple engineers who want to add typing/Sorbet.

Source code: github.com/rage-rb/sse-be…

Rage just benchmarked: ✅ 5K active SSE streams ✅ 5K HTTP reqs/sec hitting the DB Simultaneously. In 2 processes. On the cheapest €12/mo Hetzner server. This is how Ruby can and should perform.

The best Rage features are driven by the community. Thank you @megatux for pushing this through!

Native Server-Sent Events have landed in Rage! Streaming responses are just enumerators. Works with remote streams, queues, or any event source. Real-time APIs the Ruby way.

Gemini 3.1 Pro completely hallucinated 5 research links yesterday 🤦‍♂️ I’m officially back to 2.5 Pro for high-level designs. The funny part is that 2.5 Pro is still so much better than Opus 4.6, especially when it comes to defending its POV. Google, you’re going the wrong way!

Server-Sent Events are coming to Rage. Streaming endpoints with a clean, Ruby-native API - no extra infrastructure needed. The last piece to polish is observability.

Rage hit the #1 spot in Ruby Weekly last week! But the hardest part hasn’t been the async I/O. It’s been trying to explain to my wife (who designed the logo) that this doesn’t actually make her famous… Huge thanks to Peter Cooper and the community for the incredible support!

Background jobs shouldn’t require setting up a separate worker process and a message broker just to get started. With Rage::Deferred, you get a zero-setup background processing system built straight into your web process. • A write-ahead log ensures your jobs survive server restarts. • Fibers allow jobs to run efficiently alongside your web requests. We’re collapsing the stack so you can focus on writing Ruby, not managing infrastructure.

@5katkov I’ve been test-driving Codex for around a month, and I’m going back to Opus - Codex likes to overengineer quite a bit.

I was never fully satisfied with Opus, and been running multiple models in parallel. But Codex became my primary coding companion. I canceled subscriptions for all other models.

@getplainai I’m not convinced it’s a tradeoff. Both humans and agents struggle in the same environments. Good design is readable by anyone - human or machine.

@codewithrage Underrated insight. The real infra tradeoff for agent frameworks: do you optimize for developer ergonomics or agent comprehension? Most frameworks chose humans first. But agents need structured context, not clever DSLs. First frameworks to nail both win the next wave.

We’ve spent years teaching humans how to use frameworks. Now we need to teach agents too. Today Rage ships `rage skills install`. One command - and your coding agent understands Rage. If AI is part of your workflow, your framework should speak its language. Rage now does.

Is this how you currently work, or just the future you envision? Because in my bubble, the SDLC hasn’t changed much. Design is still there - I’m the one suggesting ideas to an agent and iterating on its proposals. Review and testing aren’t going anywhere either - highly autonomous AI code means I’m reviewing tests more rigorously. Ultimately, management relies on people to be accountable, not AI. How does this compare to your day-to-day at Cloudflare?

the software development lifecyle is dead requirements → design → code → test → review → deploy → monitor this loop is finished I wrote down my thoughts boristane.com/blog/the-softw…

Rage is officially part of Google Summer of Code 2026 as part of the Ruby org! If you want to work on async runtimes, observability, and real backend infrastructure - explore project ideas and start shaping your proposal. Let’s push Ruby forward. #GSoC github.com/rage-rb/rage/d…

When you build open source, nobody tells you that writing the code is only half the job. The real work is hitting a very specific balance: • Familiar enough to adopt • Different enough to move Ruby forward Get that wrong, and even the best architecture won’t matter.

@bradgessler @dorianmariecom Makes sense if the failure mode is deploy -> disconnect -> no app-level replay. Would love to see how you handle it.

@codewithrage @dorianmariecom The more practical concern for Rails devs re: thinking WebSockets are “enough reliability” is what happens during deployments. During that time, WebSockets connections will reset and messages can get dropped. For some apps it’s not a big deal, for others it’s a huge deal.

“Rails is great for MVPs” always felt like a backhanded compliment. Why should productivity imply a ceiling? An MVP should be the beginning of something durable - not something you expect to rebuild. Developer speed and long-term viability don’t have to be opposites.