We had hoped this day would never come, but Session has now entered its final 90 days of operation. If we are unable to reach our funding goal within this period, the Session Technology Foundation (STF) will be forced to shut down.
To date, the STF has received approximately $65,000 in donations. This is enough to maintain critical Session infrastructure for the next 90 days. We are extremely grateful for the support Session has received from the community, but unfortunately this is not sufficient to retain full-time developers. As a result, all paid staff and developers will have their final working day on April 9, 2026. After this date, some team members will continue on a primarily volunteer basis to help maintain Session until July 8, 2026.
Note: I can not find any separate blog or mastodon post with this same text. This is a link directly to project's donate page. There is no new snapshot on archive.org yet.
Ok I still don't know what this program does that's interesting, but it sounds like another thing we don't need.
You recognize one word in there and because you associate it with cryptocurrency you dismiss the entire thing.
To be fair, it's a consistent red flag.
Blockchain is theoretically interesting. Very interesting, in certain niches. But 9 times out of 10, "blockchain-based" is code for "enshittified" or "a pyramid scheme scam from the start." And in the cases where its implementation is altruistic, it's still questionably sustainable or creates considerable overhead.
I mean, yeah.
But also, goddamn, are we not allowed to discuss projects with it with interest because a bunch of other assholes talked about it too much a few years back? (btw I fully expect to be having this exact conversation about genAI in about a decade)
Without looking at the protocol at all, I generally think that blockchain stuff is a solution in search of a problem, but distributed storage might be used to make the system resistant to traffic analysis, the way Hyphanet does.
looks at GitHub repo
If it's doing onion routing, then it probably is intended to be resistant to traffic analysis.
The idea is decent in theory, but not in execution. The idea is that token staking is done by node operators which makes it much harder to pull of 51% attacks as it requires hundreds of euros in money to be put aside. It also protects against poisoned nodes, which is theoretically possible on something like Tor because of how easy it is to spin those up for cheap. Besides that the token also funnels a tiny amount back towards the developers in an anonymous way that would help them during development.
In practice though they should have just went without the blockchain. I have been very interested in Session but their blockchain model was always one of the biggest things that might kill the whole project.
It costs money to run a node? That's even worse. The people most willing to pay will be the ones up to no good.
It used to be around €1500 for a full node that could be shared by up to 4 stakers. Staking is different from mining coins though. You put tokens into some sort of holding and keep ownership of them. You then "mine tokens" by having the node do work while it is holding your stake.
Wait you mean the chat users have to pay to send traffic through the mix pool? This sounds worse and worse. Is BitMessage still around?
I would say once you're observed sending data into Tor or anything resembling it, you're already compromised even if your correspondent hasn't been uniquely identified. I can't see getting excited about the app.
No, it's free. They have a whitepaper on their website: https://getsession.org/whitepaper
All in all there's a pool of tokens that gets paid out to the stakers. The full network of nodes determines what nodes are eligible by testing each other. The pool gets a constant flow of tokens over time, while transaction fees and specific purchases (like a custom username instead of one of those long IDs) feed the pool as well.
Keep in mind I'm not here to sell it. I really wish it was more like free Tor nodes, in which case I would be hosting one already.
Hmm ok, but it still sounds kind of sus. One of the insights of the Mixmaster era is that what really matters is the amount of message reordering you can do, and that's why remailers typically had 24 hours or more of latency. So I've never believed in Tor (near real time). Even with a text chat network, more than a few seconds of latency will have a significant usability hit. And also, as mentioned, using the service at all probably makes you into one of the usual suspects.
The Guardian (newspaper) handles this in an interesting way, for 1-way communication from users to the Guardian itself. They have a news reader app used by millions of subscribers to access news articles and stuff. And if you want to send them a confidential news tip, the app has a feature where you can enter a text message for their editors. The news reading protocol includes some space for this type of message in every transaction, under a layer of encryption so that an eavesdropper can't see if a message is present. Allowing user to user communication through such a scheme could easily lead to mayhem, but for sending stuff to an identified recipient (the Guardian) that has some establishment cred, it's clever.