Working on adding Replicate support to RubyLLM
github.com/crmne/ruby_l...
04.10.2025 23:15
👍 3
🔁 0
💬 0
📌 0
Working on adding Replicate support to RubyLLM
github.com/crmne/ruby_l...
Screenshot of the GH issue linked in the thread. A list of ideas we could work on.
If you’re going to RailsConf next week and are interested in pairing on HotwireCombobox during day 2’s Hack Spaces—I’ve written up some ideas for what issues we could work on.
We can also work on whatever else, or just chat. Hit me up at the teal station from 10:00AM ~ 12:40PM.
Yeah I don’t mind this syntax 👌
Yeah I guess that also works! Too verbose for my taste though 😄. You’re not alone—someone else brought it up on _the other site_.
re: `protected` methods—there’s usually so few of them you can see the access modifier immediately anyway. But I’d consider them an edge case.
Indenting private methods tells you what the intended access policy is without having to orient yourself inside the class.
Specially useful when reviewing diffs and you’re not even shown the full context by default.
This is why it’s my preference in Ruby.
Oh no! Does it happen upon submitting the form, or when attempting to confirm the subscription?
Thanks for reading! Phoenix contexts are interesting. While they make it obvious where to put things, my opinion is in the end they’re unnecessary overhead in an object oriented codebase. They’re well suited for FP though. Might write more about that in the future. Good topic!
Mil gracias Julia! 😄
Thanks Stephan! Yeah I’ll probably add that soon. This is the reason I’ve avoided it:
bsky.app/profile/jose...
Wild! I actually did not know about this. And here I am reinventing the wheel. Thanks for sharing!
I’ve just added a mailing list form if you’re interested! Check out part 2 (just posted), it’s at the bottom of the post. Thanks for reading!
Right! I’d actually suggest putting it in models (even if more than one) over using service objects. But we’ll get to that further in the series.
I just published what that might look like in part 2. Thanks for reading!
RSS is a great idea, thanks Rosa!
I just wired up buttondown for a mailing list. Check out part 2 (just posted) if you’d like to sign up 😄
The reason I’ve avoided RSS/email before is I don’t want to make sure things look good there before posting 😅. But I might just link to my site instead.
Of course! That’s why I tried the service initially. No worries about what happened way back when. It was years ago. Just wired things up again and it seems to be working beautifully. Appreciate you offering!
Here’s post 2 in my Object-Oriented Rails series.
Post 1 argued business logic doesn’t belong in active jobs. But it lacked some code examples. Now we put pen to paper explaining how to move the logic into the model and out of your job.
Also, email newsletter is live!
jose.omg.lol/posts/just-w...
Nice!
I did try buttondown once but got turned off from it when my single test email landed in spam. I’ll give it another chance though. I probably just caught a temporary bug. Nbd.
Really just looking for something that’ll let me turn off open rate tracking and won’t sell my subscriber’s data.
Thanks a bunch Zack!
I really need to get on that 😅. Really I’m just avoiding having to choose a privacy-conscious mailing list provider. I don’t know of any offhand and researching seems like super dull work.
Right! Choosing between those `_at` columns and a model would be a great topic for one of these. Same goes for the many-to-many relationships. I’ll keep those in mind!
I’m kicking off a series of blog posts called "object-oriented Rails". First off — why business logic doesn’t belong in active jobs. Let me know what you think!
jose.omg.lol/posts/its-no...
Thanks Chris! I’m really excited about this 💪
After 2.5 years at 37signals I’m moving on to a new role as a Senior Software Engineer at Harvest. I start next week! 😄
Merry Christmas Chris!
It did! 😄
Two commit messages. First one reads: "Unnecessary include". CI failed. Second one reverts the first one. CI passes.
Guess it was necessary lol
Yes! I think a lambda would’ve been fine, too.
I think that important condition is fine to be implied as long as the context is at-hand (rarely is).
That said, I don’t feel particularly strong about the name of that method. I wouldn’t insist on it. Mostly trying to illustrate the main point that code can be context-aware and thus simpler.
