Recently there has been some discussion about why so many features got reverted from PostgreSQL 19. I've read blogposts, talked with people, spent some time thinking. And it got me curious.

I noticed lately that there is more mails to pgsql-bugs mailing list – basically reports of bugs, or conversations about them.

I also noticed that there are fewer emails to pgsql-general.

So figured I'll try to gather some stats, and maybe even put them in form of charts.

First, to give some perspective pgsql-hackers – very busy list:

As you can see there is data since beginning of 2024, so almost 3 years. And I do see some uptick recently in number of new threads (yellow line), but total number of emails per week is somewhat stable.

Bug reports though look different:

Reports of new bugs went up sometime between May and September. We can see it better zooming on 2026 only:

And this shows non trivial increase since mid-July.

And finally pgsql-general:

and zoom in to 2026:

Personally, I “blame" LLMs. They made more people research (and find) bugs. Which caused spike in number of mails to -bugs. But they also are used by a lot of people to answer questions. So now, there are WAY fewer mails to -general.

Incidentally this also matches that I see Slack and Discord – discussions there are almost gone. Why ask there, take a chance that someone might be unpleasant, or maybe even noone will answer, when you can get good (enough) answer in 5 minutes from some random LLM.

Interestingly, while traffic on IRC, did drop too (from average 5900 messages per month in 2024, to ~ 3200 per month in 2025, and recently to ~ 700 per month), I think that it is more active than discord and slack, which note couple of messages per week. At most.

Life of DBAs has changed. Apparently we now prefer to ask robots instead of fellow people. Whether it's good or bad, or if it even should be qualified in this way – it's up to you to decide.

On my final note I can only say that traffic to explain.depesz.com dropped significantly, and while I think that some of this drop was people moving to other visualizers, perhaps explain.dalibo.com, I kinda think/know/assume, that most of the drop is simply people trusting machines more and more with “make this query run faster", without using their own eyes/brain to read explain.

It's definitely interesting to see such big change in such a short time. And it makes me curious (and a bit worried) about future.