Hacker News Reader

Comment by hirako2000 | Hacker News Reader

hirako2000

Exactly my thought. I'm getting sick entirely of this proliferation of free services that obviously can't scale without increasingly high funding.

There are free software for each of those services which can be stood up locally, takes more time but offers more guarantees in the long term for privacy and resilience.

View on HN

These are mostly online sharing and collaboration type services, or have sharing features. That means running it on your server and sharing with others doesn’t solve the trust issue, your just requiring that other people using these services you host have to trust you.

If you’re just using the single user services then sure, if you have the technical skills and resources doing it yourself is more private.

Perhaps you could have a federated open docs (etc) platform? Maybe it could be built on Matrix? Probably tough to implement though.

NextCloud actually started implementing ActivityPub federation a while back. I don't use NextCloud so i don't know if it works, but for sure all selfhosted programs that have some notion of cooperation should be federated otherwise you end up creating new accounts for every service on every community and that's a UX nightmare.

See also a blogpost of mine on this topic, about federated forging: https://staticadventures.netlib.re/blog/decentralized-forge/

I would be interested to read more about the Nextcloud ActivityPub features.

I know about ActivityPub related discussions back in 2018[0] which links to a blog post[1] from the author of the activity app[2] that is natively included in every Nextcloud installation. However this app does not currently implement ActivityPub.

Did I miss something?

[0] https://nextcloud.com/blog/activitypub-the-new-standard-for-... [1] https://www.schilljs.com/2018/01/10/nextcloud-2018.html [2] https://github.com/nextcloud/activity/

That's a great article. Note that there's renewed activity by the FedeProxy [0] project providing funding for federating Gitea [1] using ForgeFed [2] and ActivityPub protocols.

[0] https://fedeproxy.eu

[1] https://github.com/go-gitea/gitea/issues/16518

[2] https://forgefed.peers.community

Great writing. Radicle is definitely promising, and ipfs has been growing. I think the UX makes all the difference for adoption, and once an incentive model comes up for many to offer their local resources as available serving peers we may see a great alternative to centralised services. I would offer my disk space and fast network if I could launch a process easily enough and get credit, to exchange for cash or simply equivalent hosting on other peers.

the companies behind it (like nextcloud) make money with enterprise services and supported and maintaiend deployments for paying organizations.

since most of the cpu time is used in your browser, not in the cloud, the free offer is a relatively very cheap entry level teaser for supported, maintained and potentially on-premise deployments.

Not me. Right now everybody's worried about scaling and it's keeping them from trying any new models. It's like the world forgot how to experiment with the Internet.

We experiment, and eventually advertising or other worse forms invasions kick in. It's only fair to expect a plan. They say very little about storage limits and how they would fund exponential growth if/when it occurs.

The investment in adopting a service is consequent, with daily use the effort to migrate away keeps growing. So, no, we better not experiment with offers hosting our digital life.

This is how YouTube started and of course the promise is that ultimately it’s going to make tons of money. The problem just is that 99% of these sites fail to get to that stage.

And youtube wasn't promising a free platform, so it's okay. Here we have a company setting up libre services, vaguely explaining they fund the infrastructure, I just don't see what the future will hold since they aren't at least committing to a plan.

Charging for infrastructure is acceptable. I think libre initiative should rather charge from the get go, it reduces the risk that some day a VP joins the organisation and decide to monetise the sale of our data rather than offer a reasonable pricing model where the owner of the data is the consume, not personal data exploiters.

Reply on news.ycombinator.com