I know lemmy has a big antipathy towards AI but reading the details I’m not as concerned as I was when I read the posted article.
https://www.rsyslog.com/clarifying-ai-first-what-it-really-means-for-rsyslog/
I think that approach sounds pretty reasonable. To bad they went with this marketing bullshit headline.
Using AI to improve documentation (which is neglected often time anyway) or as a specifically trained LLM for support is exactly how you should use it.
The rest sounds rational too, so I wouldn’t write rsyslog down just because of that.
If it was something like “our policy is to use AI where it’s useful, so we’ve used it to reformat all the documentation” that wouldn’t worry me. Going “AI first” for “faster innovation” to “unlock the next stage of rsyslog’s evolution” is just not what I’m looking for in a logging daemon.
At least for my home network I’m going to just go with plain old GNU syslogd.
Anyone else suddenly wondering what other syslog implementations are available? Apparently the big one at the moment is syslog-ng.
We believe that AI is no longer a hype cycle, which is why we’re writing blog posts hyping us up for using AI instead of what we actually provide
also, if you want details you should
please the the more in-depth description of AI First.
TLDR “We’re enshittifying” I guess.
Yikes, honestly if I were using rsyslog, I’d figure out how to switch to another utility that would provide it without any AI implemented in the human teams dev processes.
Well… fuck.
