Hey HN, long time lurker first time poster. I built a social network called PIECES. After building a private blog last year after I had to get off IG and Substack, I decided to productize it. It's here now. It has a dedicated web experience + iOS/Android.
Doing it by hand now. Besides the site itself, I have spent most of my time building out the backend automation. I imagine I can manage the first 10k users myself and opening it up slow enough that a community itself develops and will be self-policing.
I got back on instagram recently and the first thing I see when I login now is content from repost accounts and it's like, do I really want to see this? Is this what happens when we optimize for engagement blindly? A hypothesis I have is that if we make rules in the beginning this will be easier to enforce at scale. See an account that's just reposting stuff -> report it -> review
I hope this isn't coming off as snarky or anything, I am honestly interested in the project...
So you're thinking something like Twitter(X)'s community notes, but instead of flagging misinformation, it will be used to 'community flag' an account as AI? Maybe it shadow bans, maybe it brings it to moderation queue, whatever...
I've seen a few of these 'NO LLM ALLOWED' sites pop up now, but any countermeasure to LLM use is easily subverted. I think the only real way to keep the LLMs out is, as you seem to agree, some sort of user-reported system (But that seems open to fraud....difficult problem to solve!)
For sure. AI is only getting better so it's not a promise I can make but it's worth stating from the beginning. But the thing about Meta/Twitter/etc.,. is that they're financially incentivized to promote these accounts (more eyes = more ad revenue) so that makes it even worse. There will never be an explicit financial incentive to run bot accounts/repost farms on PIECES (I'm sure people will try though).
I recognize that we'll have to re-think things once we hit millions of users (I am an optimist). But for the first ten thousand or hundred thousand, I'm confident we can keep it a people-centric place and then strategize for future growth once our strategies start to be less effective.
I'm going to sound a little arrogant here but part of it is because most people don't actually believe in what they're doing. I see posts of founders who write, "hey, we started a social network, but we've since pivoted to <insert AI tool here>".
I believe in PIECES to my core. Everything I'm doing now with my full-time job so I don't have to take outside investment and saving is to get not only PIECES but my dating platform Lettr Dating off the ground. PIECES will be around in 5 years, I guarantee it, because it's the social network that I always dreamed of having - even going back in the day to 2003 when I would just post on my xanga.
This doesn't seem to have been picked up by the HN algo, which is fine. But I'm going to keep at it.
If you're in NYC, I'll be promoting PIECES and Lettr in the parks at night time, and I'll likely be hiring an intern at NYU to help me spread it around campus. If you see this and are at NYU, I'd love to chat - https://piecesof.me/p/6qrjydihiq3tikg3fb85fqcelqnhph
Hey HN, long time lurker first time poster. I built a social network called PIECES. After building a private blog last year after I had to get off IG and Substack, I decided to productize it. It's here now. It has a dedicated web experience + iOS/Android.
Would love if you tried it out!
Do you have some sort of detection for 'faceless accounts' ? Or are you moderating that by hand?
I like the idea, but how do you actually enforce the 'no bots' model?
Doing it by hand now. Besides the site itself, I have spent most of my time building out the backend automation. I imagine I can manage the first 10k users myself and opening it up slow enough that a community itself develops and will be self-policing.
I got back on instagram recently and the first thing I see when I login now is content from repost accounts and it's like, do I really want to see this? Is this what happens when we optimize for engagement blindly? A hypothesis I have is that if we make rules in the beginning this will be easier to enforce at scale. See an account that's just reposting stuff -> report it -> review
I hope this isn't coming off as snarky or anything, I am honestly interested in the project...
So you're thinking something like Twitter(X)'s community notes, but instead of flagging misinformation, it will be used to 'community flag' an account as AI? Maybe it shadow bans, maybe it brings it to moderation queue, whatever...
I've seen a few of these 'NO LLM ALLOWED' sites pop up now, but any countermeasure to LLM use is easily subverted. I think the only real way to keep the LLMs out is, as you seem to agree, some sort of user-reported system (But that seems open to fraud....difficult problem to solve!)
For sure. AI is only getting better so it's not a promise I can make but it's worth stating from the beginning. But the thing about Meta/Twitter/etc.,. is that they're financially incentivized to promote these accounts (more eyes = more ad revenue) so that makes it even worse. There will never be an explicit financial incentive to run bot accounts/repost farms on PIECES (I'm sure people will try though).
I recognize that we'll have to re-think things once we hit millions of users (I am an optimist). But for the first ten thousand or hundred thousand, I'm confident we can keep it a people-centric place and then strategize for future growth once our strategies start to be less effective.
Is this site inclusive of nonhuman identities? [0]
It seems incredibly focused on "real humans" etc, which is (common...), but my real identity doesn't exist in the physical world.
(There are also multiple of them thanks to DID, but that tends to require a significant rearchitecture to accommodate)
[0]: https://en.wikipedia.org/wiki/Otherkin
We'll have to open a committee to launch an investigation to produce a report about whether accounts held by Otherkin should be allowed.
Stay tuned and thanks for your patience.
That almost sounds satirical, but if you are serious about this, thank you. <3
I’ve seen a few posts about a “better social network” tried a few even thought about making my own.
It seems hard to get them off the ground.
I wonder if the answer to the enshittification of social networks is “no social network at all”.
I'm going to sound a little arrogant here but part of it is because most people don't actually believe in what they're doing. I see posts of founders who write, "hey, we started a social network, but we've since pivoted to <insert AI tool here>".
I believe in PIECES to my core. Everything I'm doing now with my full-time job so I don't have to take outside investment and saving is to get not only PIECES but my dating platform Lettr Dating off the ground. PIECES will be around in 5 years, I guarantee it, because it's the social network that I always dreamed of having - even going back in the day to 2003 when I would just post on my xanga.
This doesn't seem to have been picked up by the HN algo, which is fine. But I'm going to keep at it.
If you're in NYC, I'll be promoting PIECES and Lettr in the parks at night time, and I'll likely be hiring an intern at NYU to help me spread it around campus. If you see this and are at NYU, I'd love to chat - https://piecesof.me/p/6qrjydihiq3tikg3fb85fqcelqnhph
[dead]
[dead]