A post to the IPFS subreddit (linked in the context) seems to imply that much of IPFS-related development is winding down. To be clear, the exact details of what this means haven't yet been spelled out. I suspect that things like bootstrap nodes will continue to operate (since those are cheaper and more essential than everything else they mentioned) so Cacophony shouldn't be meaningfully impacted.

I was sort of wondering about this, since they stopped being as diligent about announcing updates, recently. Also, they have always cozied up to crypto-adjacent projects, and most of those have collapsed. They actually had useful technology and a half-decent implementation, but I don't think it is a project which can be maintained with all the extra bells-and-whistles they were supporting.

Part of the problem is that they made it too easy to not properly use the technology. While this was initially provided as a sort of crutch to help bootstrap projects, it eventually became something many projects depended upon and it fostered a culture of thinking about decentralized systems as "central-lite", which I would argue is completely backward and guaranteed to fail (although this is just my own opinion).

I suspect that they need to scale down the support overhead to be something more along the lines of what BitTorrent requires. I suspect that this is what will happen but I am not sure what their new long-term plan is.

Again, this makes me wonder if it might be worth running that experiment I have been thinking about, for years: Can I build a lighter-weight decentralized P2P key-value store to more directly solve the narrow use-case of things like Cacophony? Now, to begin with, I would need to look into things like the hash distribution of various crypto algorithms and make an experimental "loaded network test" to see how things would practically work on that design, in a real network.

Maybe, but it is hard to justify when nobody else seems to even know what I am talking about,
...Nights