I, too, have noticed this and am curious to learn more about what it implies about the status of the project. At the risk of dating myself uncharitably, I rely on RSS feeds almost obsessively as I attempt to stay abreast of new developments regarding the various and sundry topics that interest me. I don’t think it unnatural to be sensitive to clues that a project is beginning to feel less obligated to communicate well with its users and supporters, even when its level of activity elsewhere remains robust; that may be more ominous than a mature project simply having less to communicate.
With all due respect to @milesrotaru1, a changelog is a soulless catalog of changesets, all presented with equal weight be they the introduction of a long-awaited feature or the correction of a misspelled code comment. I regard it as a resource primarily for identifying when particular changes were made in the code base and little else. In contrast, what made the release announcements on the blog of much higher value (to me, at least) was the curation, framing and added context provided by the author.
Lastly, in anticipation of what has become a far too ubiquitous impulse in recent years, I will implore anyone tempted to “just let A.I. generate a blog post about future releases” to ignore that temptation. No amount of mellifluous prose erupting from the LLM du jour can wallpaper over the implication that no one involved cared enough to write the post themselves; to do so would only pour gasoline on the glowing embers of my concern, not extinguish them with water. If there’s one thing the A.I. explosion has done better than anything, it’s been to hone our bullshit detectors like a razor’s edge.
For those that may have missed it, here’s a gift link to Bret Stephens’ op-ed in yesterday’s New York Times, artfully reminding us of many ineffable qualities inherent to human authorship that even the A.I. of the next century won’t be able to replicate.
