iiinova — where your digital self keeps evolving

Radan @radan.dev · 27/09/2026

Your comment was not an observation. It was a judgement that revealed the core of the issue.

Radan @radan.dev · 23/09/2026

There are many many smart people working on or around Rails. The project is bigger than one person, no need to throw the baby out with the bath water. :)

010

Reposted by Radan

Bozhidar Batsov (a.k.a. Bug) @batsov.net · 11/09/2026

Here's a more detailed write-up about the massive #RuboCop 1.91 release metaredux.com/posts/2026/0... I've focused in the article on the new "preview" functionality (inspired by ruff and friends), but I've covered briefly the other notable improvements as well. #Ruby

metaredux.com

RuboCop 1.91: The Journey to 2.0 Begins

RuboCop 1.91 is out! It has the usual mix of new cops, new options and a small mountain of bug fixes, but the feature I want to talk about today is a single flag - --preview. It’s the first real step ...

064

Radan @radan.dev · 09/09/2026

Even at zero temperature, it would still give out readable text. The temperature, i.e. how randomly it will select the next token is what gives it “creativity“.

000

Radan @radan.dev · 20/08/2026

I'd prefer it broken down into multiple parts and released over multiple weeks. I'm not going to listen to it in one sitting so I'm going to break it into parts anyway. It would be better if you do it intentionally rather than me randomly.

120

Radan @radan.dev · 16/08/2026

I was going to do that as well, I already marked the exact book. But then it got swept away.

010

Radan @radan.dev · 31/07/2026

Great, see you there! :)

010

Radan @radan.dev · 30/07/2026

Got myself a ticket for Euruko. Who else is going?

130

Radan @radan.dev · 27/07/2026

I’m trying to find user facing products built and maintained with Gas City (gascity.com), successor to Gas Town, but I’m not finding anything. Is this just performance art or is it being used for anything with users?

gascity.com

Gas City - Software Factory Platform for Coding Agents

Gas City is the open-source software factory platform software engineers use to orchestrate CLI coding agents with Beads, formulas, packs, and the Registry.

000

Radan @radan.dev · 23/07/2026

Heck, maybe sometimes you just want to write it.

010

Radan @radan.dev · 23/07/2026

For any given piece of agent generated code you can decide to read it or not. Either can be good. The most important part is that you are deliberate about it and can explain it. Btw, "This part is not important" on one end and "I just want to read it" on the either can be enough.

130

Radan @radan.dev · 22/07/2026

I think you’re right and you’re your own proof since that advice mirrors your own early success. But I doubt it’s scalable career proofing advice. Even if Chinese labs welcome an offline unknown outside collaborator, they‘re unlikely to welcome 1000s of them. 😅

010

Radan @radan.dev · 22/07/2026

Ah yes, the gray area of the framework. You could argue that some of those deeper changes effectively produce a new framework. And the fact that you can do that easily is proof of a good design of the sub components of the original framework. Btw, I agree with half of your example. ;)

020

Radan @radan.dev · 22/07/2026

Since every framework is essentially a set of premade decisions, another way to put it is to say that a good framework has picked all the decisions which you anyway shouldn’t be making yourself while getting out of your way in the ones that are actually important for your project.

120

Radan @radan.dev · 22/07/2026

Ahh, you mean contribute to the open model’s documentation? That makes sense. Sorry, I was thrown off by the wording, I thought there‘s a way to contribute to the model itself. AFAIK, only the weights are open on the big ones. The training data and code are closed.

110

Radan @radan.dev · 21/07/2026

Wait, how do you contribute to an open weight model?

110

Radan @radan.dev · 21/07/2026

For all the people that like to say, "Coding Agents are like compilers" or "Why do you still read code, would you read the output of your compiler?" I have an amazing resource to recommend to you: craftinginterpreters.com

craftinginterpreters.com

Crafting Interpreters

Ever wanted to make your own programming language or wondered how they are designed and built?

130

Radan @radan.dev · 16/07/2026

I care about code quality, which still maters (agents also work better with good code), but it's no longer as economical for me to craft it by hand. So I end up thinking more about how to precisely explain it rather than rely on intuition. Which is both harder and more rewarding. So that's nice.

010

Radan @radan.dev · 15/07/2026

Broadcasting Turbo refresh adds an extra round trip and a debounce delay. Broadcasting the content directly fixes that, but introduces a race condition. Here's how to keep fast, immediate updates without race conditions: radan.dev/articles/tur...

radan.dev

When broadcasting a Turbo refresh is not enough: faster UX with versioned immediate updates

The simplest way to make a page collaborative with Hotwire and Rails is to subscribe to changes and broadcast a refresh when relevant models update.

031

Radan @radan.dev · 14/07/2026

Social media rewards being sure: sure that this is the absolute best way to do the thing, that not doing this or that "will leave you behind", as if it was a fact. Can that ever be true? I'm not so sure.

010

Radan @radan.dev · 14/07/2026

Also, the fact that you think you're the one who took him out for a walk just shows he's a total pro! 😁

110

Radan @radan.dev · 14/07/2026

It's extremely important for any business owner to hire a sanity officer and it's really hard to find someone more qualified than a dog.

110

Reposted by Radan

Janko Marohnić @janko.io · 13/07/2026

Shrine 3.9.0 released, mostly polishing some rough edges and increasing compatibility between plugins. shrinerb.com/docs/release... And we've reached zero issues again after a long time 🤘🏻

shrinerb.com

Shrine 3.9.0 | Shrine

New features

031

Radan @radan.dev · 10/07/2026

Totally, by default it writes it like a government notice about roadworks: Too much unnecessary detail, intent is completely unclear and it's hard to figure out if it's relevant for me. 😄

120

Radan @radan.dev · 10/07/2026

The video was useful, thanks! Is there anywhere copy-pasteable where I can see your Zed settings file?

100

Reposted by Radan

Tess Snider @malkyne.com · 08/07/2026

I am exasperated with people accusing programmers of "gatekeeping" programming. We were practically dragging people through the gate. Do you know why a mindless LLM can tell you so much great programming information? It's because we've been sharing that information for free online for decades.

923060496

Radan @radan.dev · 06/07/2026

Nice, looking forward to watching the recording when it comes out! I remember you telling me you plan to submit this, it sounded very practical. :)

110

Radan @radan.dev · 06/07/2026

Oh nice, saved it for watching. My plan is to watch it next time I sit down to work on a side project and and I'll try and do it with Zed. Thanks!

000

Radan @radan.dev · 05/07/2026

Thanks for the hint, maybe Joel will remember. 🤞 🙂

100

Radan @radan.dev · 05/07/2026

Ah, hm, I used sublime for many years before I switched to VSCode. I'm not familiar with Zed's relation to Sublime? Is there something there? Maybe I still remember something from Sublime days. :)

200

Radan @radan.dev · 05/07/2026

Thanks, but that comes second. I was thinking more for some resources on getting the most of out Zed in general. I need to transfer muscle memory but I also need to get acquainted with the editor power features in general before I can be productive in it.

210

Radan @radan.dev · 05/07/2026

Any resources you’d recommend for getting quickly up to speed on zed, bonus points of it’s specially for people coming from VSCode? I keep wanting to try Zed but can’t find the time to figure it out. A shortcut would be great, even if it was a paid course.

110

Radan @radan.dev · 01/07/2026

There's no silver bullet but Altcha is pretty good: altcha.org (TL:DR; one of the key mechanisms is making bots do proof-of-work style calculation on first visit to make the user prove they are running a full modern browser on a capable machine).

altcha.org

ALTCHA - Next-Gen Captcha and Spam Protection

ALTCHA is a SaaS platform providing Captcha alternatives and AntiSpam protection for websites with privacy in mind.

110

Radan @radan.dev · 11/06/2026

2 weeks ago, I attended RubyConf Austria in Vienna. A first day AI panel left me wondering: if programming languages are becoming irrelevant, will Ruby conferences still matter? Then, the next day, I noticed what to me looked like clues pointing to an answer. radan.dev/news/ruby-co...

radan.dev

RubyConf Austria and the future of Ruby conferences

To get from Zagreb, where I live, to Vienna, where RubyConf Austria was happening: drive 173km to Graz, take a right and go for another 198km. Sure, I’ve glossed over two country borders and some…

130

Radan @radan.dev · 10/06/2026

"Professional coders" is a very wide term. A motivated, curious professional programmer runs circles around a motivated, curious 12 year old programmer. I'm speaking from experience: I've been both of those things: I started programming when I was 10 and wrote first production code at 16. :)

000

Radan @radan.dev · 10/06/2026

Oh, I agree. It's why the mediocre LLM produced code is so mind blowingly useful: it's better than half of the code out there. But, and I think this is crucial: for the non boilerplate parts of software, most of the code out there AND the AI code is not good enough. We just put up with it.

000

Radan @radan.dev · 09/06/2026

Hm, yes, the question of are we real engineers? I think it depends. There's levels to it. I've made sand castles and I've made bridges from sticks in back gardens. Don't think that was engineering but real castle and bridge construction clearly is. The difference is just less obvious with software.

110

Radan @radan.dev · 09/06/2026

Ok, that's totally true. My comment was primed by the initial comment's example of surgery.

010

Radan @radan.dev · 09/06/2026

The key difference between coding and most other skills is that you can just start coding, make any kind of mistake with 0 consequences, learn, try again, and repeat until you figure it out. With surgery not so much ...

110

Radan @radan.dev · 08/06/2026

I walked out of the forest at dawn, holding a map, to stop the first car and ask them if they can point on the map where we are.

000

Radan @radan.dev · 07/06/2026

If you emigrate, it’s in 29days. 😁

030

Radan @radan.dev · 02/06/2026

This was innocent but it reminded me that our intuitions for where an intelligence will succeed or where it will fail are not relevant with LLMs. It's dangerous to check its work the same way you'd check another human.

010

Radan @radan.dev · 02/06/2026

I got an LLM to search for 20 articles on a topic, and, given requirements, select 5 for me to read. It picked good articles, gave me concise summarise and presented them in a numbered list: 1., 2., 3., 5. Yes, that's right, there's no number 4. 😁

100

Radan @radan.dev · 30/05/2026

Some photos from first day of RubyConf Austria:

0112

Radan @radan.dev · 28/05/2026

Definitely eat the cost. Craft an e-mail notifying the user that it won't be free going forward and frame the last 10 months as an unintentional gift. Eating the cost will build goodwill while chasing back payments is likely to drive the user away and maybe even not recoup the payments. :)

170

Reposted by Radan

Baldur Bjarnason @baldurbjarnason.com · 27/05/2026

“Big Tech’s Anti-Labor Playbook Has Come for Wikipedia” medium.com/@jakeorlowitz/wikipedia-… > Fire the engineers who know how the system works, fire the ones organizing labor, hope nothing catastrophic breaks before you can ship something splashy

medium.com

Big Tech’s Anti-Labor Playbook Has Come for Wikipedia

TLDR: In ten days last month, the Wikimedia Foundation fired the longtime lead developer of MediaWiki and disbanded the team whose entire…

33019

Radan @radan.dev · 26/05/2026

Today, the agent was not succeeding in debugging so I used the debuggers to figure it out the ancient way. It's entirely possible that I could have used the AI differently to solve it slightly faster. But man, did it feel good to be deeply focused on a single thing. So good.

040