Sabitlenmiş Tweet
Kyle Fox
35.2K posts
Kyle Fox
@kylefox
Probably hacking on a Rails app. 🏋🏻♂️🎸☕️💻
Edmonton, Alberta, Canada Katılım Haziran 2007
617 Takip Edilen3.3K Takipçiler
@bradgessler I chose the first option, but I typically don't add `account_id` everywhere unless denormalizing for performance.
I scope all queries off the tenant (Current.account.projects) to the point that querying models directly (Project.where) becomes an anti-pattern in the codebase.
English
Curious how most people deal with multi-tenancy in Rails apps, particularly from those who have done it so many times they feel they've unlocked a formula for it.
It's def trade-offs all the way down.
English
@levelsio@levelsio
✨ A dream I had finally came true: I can now chat directly with my sites to build any feature or fix any bug just via Telegram I've been playing with OpenClaw for 3 weeks now and it's great but I was always too scared to run it on any production server And I was right a bit as @marckohlbrugge was able to hack it by social engineering and acting as if it was me, and with enough tries it believed him, and was able to modify the server, change SSH keys etc. of course I had it isolated properly on its own VPS and it didn't touch anything sensitive (as it should!) Marc then reported that bug to @steipete who patched it fast But I wanted to try something more basic and simple, and I think maybe more secure: to just connect Claude Code on my server to Telegram which would be hard locked to only messages from me So I installed claude-code-telegram by @RichardAtCT on the server and run it as a system daemon and it works really well The cool thing is that I was already using Telegram for server errors like this: > Photo AI - ❌ Random credits giveaway failed (Attempt 30/30) with an exception: SQLSTATE[HY000]: General error: 5 database is locked So now I can just reply, "Ok fix this", and Claude Code on the server in production will try (and probably succeed) in fixing it In the video below I asked it to make show [🌳 Parks ] on the map by default on load, it did that, then I reloaded the page and it instantly worked One thing it still needs is sending actual messages while it's doing stuff which OpenClaw does really well, it's annoying to just wait while it says "Working..." but that's probably next
English
@erick_tmr @claudeai Yes, I understand the code is probabilistically generated. And when it suddenly starts consistently generating dramatically terrible results over the course of weeks when no other variable has significantly changed, I’m inclined to suspect the black box is the problem.
English
@Espen_Antonsen @claudeai Our designer did! I wondered if anyone would catch that 🤣
English
@jakeonrails @claudeai Probably the last ~4 weeks.
I’ve shipped thousands of lines of code with Claude over the last 6+ months and I’ve never seen it do the absurd stuff it’s done in the last few weeks.
English
English
@erick_tmr @claudeai I told it to add validations, in fact I even specified model validations.
How come I’ve shipped thousands of lines of code over the last 6 months and not once has Claude used such an obvious anti-pattern.
Sometimes it IS the agent’s fault.
English
I think AI agents are great picking up the patterns in a codebase. That being said, when one pattern that I've recently been enjoying is after the main work is done spin another Claude session to code review what's done, and then go back and forth between both session with code review feedback and address it prompts
Kyle Fox@kylefox
Claude Code added this validation code to a Rails CONTROLLER. Opus 4.6, high-effort, 1M context. WTF has Anthropic done to @claudeai? 😭
English
@nateberkopec @claudeai That’s a good point, never thought of it that way.
That said, I’ve never really seen this kind of poor output due to context / compacting 🤔 I usually see it forgetting requirements or decisions, not implementing total anti-patterns that exist nowhere in our codebase.
English
@nateberkopec @claudeai I think was around 20%-30% at this point. I specifically used the 1M model because I knew it would likely be a long / complex session.
English
English
@inazarova @claudeai But thanks for the link, it looks good and I'll give it a try! 👍
English
@inazarova @claudeai But why have I not needed this for the last 6 months and thousands of lines of nice Rails code Claude has shipped for me previously? Why in the last month has it started sneaking in terrible code?
English
@kanniball @claudeai Cool, I'll check it out. But I wonder why this behavior has started recently (~4 weeks) when for months I've shipped thousands of lines of code without issues like this 🤷🏻♂️
English
English
@jordanyeo Damn. We've put a ton of effort into optimizing ours and the fastest I got them to run was ~10 minutes on my M5 😭
English
Hot take: don't run your entire test suite as part of your CI & deployment pipeline.
Only tests related to *critical* system functionality should block deployments.
All other tests can run nightly to catch non-critical regressions. Failures enter your bug-triage process.
English
@chriskottom Yeah, I now understand what "belt and suspenders" means and it does make sense in this context ... but come on, the only people who say that in the real world are old grandpas *actually* named Claude.
English
@kylefox Would be nice if Claude could take a hint now and then from devs with 20 years experience in Rails rather than getting cute. It’s like, dude, you’re only like 3 years old.
English
