Heads-up: The CTO of an "AI-powered social network" startup Maven, Jimmy Secretan (ex-lead of OpenAI), confirming that his app has "ingested about 1,120,000 posts from Mastodon".
-
Heads-up: The CTO of an "AI-powered social network" startup Maven, Jimmy Secretan, confirming that his app has "ingested about 1,120,000 posts from Mastodon".
Maven
Maven: Follow interests, not influencers
(app.heymaven.com)
Contact: [email protected]
Via @liaizon, @djsundog, and others https://social.wake.st/@liaizon/112603447990005434
-
I don't see any of my posts there, but I sent Jimmy an email to preemptively stop him from scraping my server.
-
-
@Mikal @liaizon @djsundog Yeah, I'm kind of getting tired of dealing with this.
I was just looking into how to easily modify robots.txt on a Mastodon server and noticed that it blocks GPTBot by default.
https://mastodon.social/robots.txt
mastodon/public/robots.txt at main · mastodon/mastodon
Your self-hosted, globally interconnected microblogging community - mastodon/public/robots.txt at main · mastodon/mastodon
GitHub (github.com)
Nice.
-
@stefan @liaizon @djsundog Hi this is Jimmy here. Happy to remove any of your posts from Maven and cease ingestion from those servers going forward. We are trying to connect up to the Fediverse, to allow interaction with other ActivityPub servers. This definitely seems to me to be within the spirit of what ActivityPub enables, but of course, I don't want to have Maven connect to anybody who doesn't want it.
-
Folks around here have different preferences when it comes to connecting with other servers.
Some are fine with their posts showing up in other parts of the internet, fediverse or not, some only want to network with fediverse servers that are not run by corporations.
-
@jsecretan @liaizon @djsundog From my brief reading of the situation, it seemed like you're just creating copies of profiles with no links back to the original pages, eg: https://app.heymaven.com/profile/66927
I can't see a way to get to the original profile or look it up from my server.
-
@jsecretan @stefan @djsundog Hello Jimmy, this is not about me personally. I think how you went about interaction with the fediverse breaks all the expectations people here have about how it works. On the profile you ingested of mine, you don't show anywhere in your interface a link back to the actual profile you are "ingesting"
You are also seemingly not opening up any of your posts to the fediverse so you are benefiting from our network without giving anything in return.
-
-
@jsecretan @liaizon @djsundog I don't believe there is a mechanism for that, I am only now looking into how to modify my server's robots.txt, which, by the way, includes GPTBot by default, and that might give you a bit of an idea of the community's expectations.
-
@jsecretan @stefan @djsundog do you mean that you are pulling posts directly from only the mastodon.social API? Have you implemented #ActivityPub in Maven at all? or are you just grabbing the posts from there and importing them into your own database?
-
@liaizon @jsecretan @stefan @djsundog
I concur with Liaizon. This isn't how this works. No one starts a fediverse (AP) server by ingesting a bunch of posts from others without their consent. They start servers and start federating with the rest of the network.
Please stop ingesting posts from AoIR.social (I'm the admin, btw).
-
@liaizon @jsecretan @stefan @djsundog
To add to this, the custom is to start a server with a code of conduct, including clear moderation rules, so that the rest of us can make informed choices about federating.
What you've done with Maven is a pretty massive violation of norms, and likely it will result in your being defederated from many other instances. It's a poor way to start an ActivityPub implementation.
[edited for clarity]
-
@stefan @liaizon @djsundog for those living in the EU: https://www.datarequests.org/generator
-
This whole #maven thing will likely lead to discussing consent in the fediverse again, so I'd just like to repeat something I said further down the thread.
Folks around here have different preferences when it comes to how their stuff should be shared. And different perspectives on what is and isn't part of the fediverse.
So if someone wants to be left alone, it's okay to leave them alone, no need to "well actually" them.
-
And I hope this doesn't need to be said, but let's be civil when emailing Jimmy. I'm frustrated by the arrogance of tech bros, but will never stoop to harassment. A brief, to-the-point note will do!
-
-
@researchbuzz @liaizon @djsundog We need to rethink robots.txt!
Stefan Bohacek (@[email protected])
Just throwing out a thought before I do some research on this, but I think robots.txt needs an update. Ideally I'd like to define an "allow list" that tells web scrapers how my content can be used. Eg.: - monetizable: false - fediverse: true - nonfediverse: false - ai: false Etc. And I'd like to apply this to my social media profile and any other web presence, not just my personal website. #internet #fediverse #SocialMedia #robotsTxt
Stefan's Personal Mastodon Server (stefanbohacek.online)
Mostly just venting here, no idea where I'd even start trying to push something like this through.
-
@stefan @liaizon @djsundog Hello everybody, Jimmy CTO of Maven here again. Thanks to everybody who has given honest feedback about our Maven Fediverse integration. As I mentioned toward the top, we set out on this project to help our users on Maven find interesting posts across the Fediverse, and to allow our users to send and receive replies from the rest of the Fediverse.