I'm sorry, can everyone not read the actual link? They specifically say they are not training generative AIs, i.e. LLMs.
They are using the data to train non-generative AIs for stuff like emoji and channel suggestions. I.e. "you use this emoji a lot, so it's displayed first" or "people that are in these channels of yours also join these other channels you're not in yet".
This is class A misinformation being spread here, good job. It's unbelievable that I'm the first one to actually verify the truth of this post, because I wanted to share it further and that's what you do before.
It also explicitly states in the posted screengrab that the opting-out user's workspace won't contribute to the underlying models. How would that be separate from using info on their workspace as training data for any kind of model? My interpretation of that is the data would be used to inference on the models, not train them.
Ok but seriously, that is a very good reason to ban it. Who knows what would happen if the AI just fully ripped someone else’s code off that’s supposed to be like GPL licensed or something. If humans can plagiarize, than AIs can plagiarize.
But also, how are they still using CVS? CVS is so slow and so bad. Even Subversion would be an upgrade.
No, but it's the only one that big corporations know/care about... primarily because of MS's aggressive "look, it's basically free and TOTALLY the same thing, we promise" marketing strategy.
Slack [...] will never identify any of our customers or individuals as the source of any of these improvements to any third party, other than to Slack’s affiliates or sub-processors.
So you sign up to confirm that your IP is yours, while simultaneously agreeing to sell it off , but the source will be anonymous other than to who it’s sold to or anyone else Slack decides they want to know. These tech contracts and TOS should just say “we will (try) not do bad, but you agree to let us do bad, and if bad happens its your fault”.
Open-source does not equal free. There is probably a way to circumvent their paywall, but on their website, self-hosted free version has a limit of 1-on-1 for calls.
If you want a group call, you need to pay, even if you self-host.
Wait, which calls do you mean? We're hosting or own free mattermost and I've never had that problem.
We're pretty happy with mattermost overall, but it's also a small tech-savvy team.
I use Slack at work everyday. I suppose this does feel off in some way but I'm not sure I'm the right amount of upset about this? I don't really mind if they use my data if it improves my user experience, as long as the platform doesn't reveal anything sensitive or personal in a way that can be traced back to me.
Slack already does allow your admin to view all of your conversations, which is more alarming to me
The problem is where you said "as long as" because we already know companies AND the AI itself can't be trusted to not expose sensitive info inadvertently. At absolute best, it's another vector to be breached.
Instead of working on their platform to get discord users to jump ship they decide to go in the same direction.
Also pretty sure training LLMs after someone opts out is illegal?
Also pretty sure training LLMs after someone opts out is illegal?
Why? There have been a couple of lawsuits launched in various jurisdictions claiming LLM training is copyright violation but IMO they're pretty weak and none of them have reached a conclusion. The "opting" status of the writer doesn't seem relevant if copyright doesn't apply in the first place.
Nor is it up to you. But fact remains, it's not illegal until there are actually laws against it. The court cases that might determine whether current laws are against it are still ongoing.
If copyrights apply, only you and stack own the data. You can opt out but 99% of users don't. No users get any money. Google or Microsoft buys stack so only they can use the data. We only get subscription based AI, open source dies.
If copyrights don't apply, everyone owns the data. The users still don't get any money but they get free open source AI built off their work instead of closed source AI built off their work.
Having the website have copyright of the content in the context of AI training would be a fucking disaster.
It's not illegal. 2. "Law" isn't a real thing in an oligarchy, except insofar as it can be used by those with capital and resources to oppress and subjugate those they consider their lessors and to further perpetuate the system for self gain
For matrix specifically, I recommend fluffy chat on mobile and cinny for web/desktop. Most notably, they both support the not-yet-official spec on custom emojis and stickers, which I think is important for any slack-like.
For the server (since you want to self host), you'd probably want to do Synapse - it supports not being federated as well as SSO. Also it wasn't mentioned by mp3, but xmpp is another protocol that's used by many large companies for internal chat systems as well.
Does IRC still exist? I remember laughing when I first saw Slack and its early competitors because people were excited about it and when I finally used it I realized it was basically just IRC with a nicer interface. I’m assuming these offer improvements like encryption?
Nah, there's tons of features that slack has over irc. To start with inline media (images, audio, video), but most importantly lots of out of the box external integrations and webhooks.
Yeah, now there is, but I don’t think a lot of those features were in when I first used it over a decade ago. It became a lot more useful over the years.
It didn't require using arcane commands just to sign up and log in. I love IRC and will always remember it fondly, but it wasn't easy for a novice to use and that's why things like slack and discord took off.
Subscription to a software is not mutually exclusive with self-hosting. Developers deserve to earn money, especially those who do not rely on collecting data, showing ads and enshittification of their cloud platform.
It’s a business with 300 users, but only about 50 of them would even use it. The others DO need accounts for the one time per month they login. But with their pricing, and SSO plan, that’s $3000/m
mastodon.sdf.org
Hot