Some updates: @topazlabs.bsky.social acknowledged the issue and made it right. @simpleanalytics.bsky.social fixed the SSL problem. Commento.io is still unresponsive, had to start a dispute. Thankfully I found a recent backup of my comments, and was able to move to FastComments.

In what I’ve seen, doesn’t seem TPTBs make decisions based on numbers. It’s more about specific arguments. Having 30 comments say “we want the default without the ugly comment” won’t move the needle as much as one person laying out simple and clean strategy kinda like you did.

I was thinking why AI code doesn’t sit right with me even if it gets very good in the future, and it’s about possibility of human intervention. Wrote a small opinion here, curious what you think. max.engineer/failover-to-hi

Curious why you don't prefer to specify rate limits per controller, seems clearer in our case. As far as state, it seems to be kept in redis, have you tried digging in to see what/how it stores? (I haven't)

For example, if I'm in a repo, use the top search bar, and find nothing after a couple of attempts, I cannot go back to the repo I was in without pressing back multiple times.

Is it now _outdated_ to care about code quality minutia, and not losing my ability to focus insanely well on little details? I guess all of you folks in the "syntax doesn't matter" camp should be happily embracing LLMs. Working code is all that matters, right?

I notice that especially older gems tend to not separate layers. If you want to

do_specific_thing

most of the time you should:

1) def do_generic_thing(options)

and then

2) def do_specific_thing; do_generic_thing(CONFIG_FOR_SPECIFIC_THING) end

Just edited the wording. Plz let me know if that still reads dissonant. Appreciate the feedback!

I took @avdi.codes‘s advice to add that. 😅 But now that I have a few people in the list, I have no idea what to do with that info. Was traveling once, and asked if anyone wants to hang out while I’m there (in case anyone’s there) but no takers. I’m probably doing “finding friendship” wrong.

Potentially an AI could learn how groups of colleagues think and leave reviews that are likely to match their behavior. That could be an interesting way to address what you said.

That’s exactly it, our perception is the only thing that matters, and all rules and “laws” are incidental, and may help or hurt case by case. But my advice generally has been that we should lean on our colleagues’ review, since most codebases are maintained by small groups.

Ironically, maybe ai can help us measure that. A while ago I wrote about 3 maintainability questions: how what and why max.engineer/maintainable.... Maybe we can get AI to review PRs for how well they answer them. Would be an interesting experiment (haven’t tried). This isn’t a “how much value” ofc.