3 ms·

>For dealing with "bad actors", the problem is anonymity I wish it was this simple :^) Consider Facebook, where people post mostly under their own name and ph

by dexen 7y ago

>For dealing with "bad actors", the problem is anonymity I wish it was this simple :^) Consider Facebook, where people post mostly under their own name and photo. Even a casual visit to Facebook quickly reveals your hypothesis is wrong; it's nearly the exact opposite of reality. Contrast that with HN, where users range from pseudonymous to fully anonymous, and where discussion is kept to a much higher standard. It bears repeating - the discourse on HN is kept to a much higher standard. Active moderation and community guiding, performed by intelligent agents, is the real answer to the woes. Anything automatic, anything with a guaranteed outcomes will be 'gamed' and put to bad ends. --edit-- There's also the separate but equally important matter of privacy. As internet spaces became both the town square and also gentlemens' clubs and also private homes to multitude of discourses, we need privacy from various actors' prying eyes.

derekp7 7y ago
The only way this works on HN is through active efforts to remove some of the anonymity -- I'm not talking about actually finding a contributor's real-world identity, but making sure that the same person doesn't create 5000 accounts and just switch to the next one when one gets banned. This requires at the minimum logging IP address that someone signed up from, and treating with suspicion IP addresses that are pooled (such as from a VPN or TOR, etc). There are also other heuristics to prevent sock puppets and voting rings, each of these removing some amount of anonymity. Whereas what i proposed, people can still maintain full anonymity, and if they really need a separate ID so that they can log in to the same forum under a different ID than they normally use (for things like whistle blowing, etc) then they have that option -- just buy another $5 (or cheaper, or more expensive) certificate. It would be up to the forum (or the auto moderation system, for assigning points), to decide how much to trust a user with a new cert, based on how much they paid for it (as an example), or how much PII they provided to the cert provider. Another thing this can do is allow someone to be anonymous, but with a chosen pseudonym, and allow their reputation from one site to benefit them on another site (by proving that a given reputation score belongs to them). Of course, this is almost sounding too much like a "social credit score", so it may not be acceptable on those grounds.
- dexen 7y ago
  >The only way Way to start with a false premise. The main way on HN is efforts, both by mods and by users, to buoy up good content, and push down bad content. >[is to] remove some of the anonymity Not even close. Preventing cheap creation of new identities isn't inherently tied to identification of users; there already deployed alternatives. One example, TeamSpeak uses proof of work - a computational process that's expensive to perform and cheap to verify - to protect identity creation. Works wonders, and allows anonymity just fine. At any rate, the thrust of my argument is different, and cuts deeper into your idea: any automated process, any guaranteed outcome, will be subverted and put to bad ends. To grow & maintain a good community you need active management by intelligent agents (ordinarily, people). And not just by selected mods - it helps when typical user is well aware what is up and able to fend for himself.
downerending 7y ago
Discussion here may be kept to a higher standard, but that also means that a lot of things I'd be interested in are censored. Not sure what the solution is, but it would be nice to have some sort of ML-based approach that would tune the content I see to my wishes, rather than the wishes of the average denizen or the moderators.
- dexen 7y ago
  >a lot of things I'd be interested in are censored Same for me. Frankly I accept HN is a space to discuss a limited range of subjects, and keep other subjects to elsewhere. It's annoying to a degree, but it the longer run it works.

by dexen 7y ago

derekp7 7y ago

The only way this works on HN is through active efforts to remove some of the anonymity -- I'm not talking about actually finding a contributor's real-world identity, but making sure that the same person doesn't create 5000 accounts and just switch to the next one when one gets banned. This requires at the minimum logging IP address that someone signed up from, and treating with suspicion IP addresses that are pooled (such as from a VPN or TOR, etc). There are also other heuristics to prevent sock puppets and voting rings, each of these removing some amount of anonymity. Whereas what i proposed, people can still maintain full anonymity, and if they really need a separate ID so that they can log in to the same forum under a different ID than they normally use (for things like whistle blowing, etc) then they have that option -- just buy another $5 (or cheaper, or more expensive) certificate. It would be up to the forum (or the auto moderation system, for assigning points), to decide how much to trust a user with a new cert, based on how much they paid for it (as an example), or how much PII they provided to the cert provider. Another thing this can do is allow someone to be anonymous, but with a chosen pseudonym, and allow their reputation from one site to benefit them on another site (by proving that a given reputation score belongs to them). Of course, this is almost sounding too much like a "social credit score", so it may not be acceptable on those grounds.

dexen 7y ago
>The only way Way to start with a false premise. The main way on HN is efforts, both by mods and by users, to buoy up good content, and push down bad content. >[is to] remove some of the anonymity Not even close. Preventing cheap creation of new identities isn't inherently tied to identification of users; there already deployed alternatives. One example, TeamSpeak uses proof of work - a computational process that's expensive to perform and cheap to verify - to protect identity creation. Works wonders, and allows anonymity just fine. At any rate, the thrust of my argument is different, and cuts deeper into your idea: any automated process, any guaranteed outcome, will be subverted and put to bad ends. To grow & maintain a good community you need active management by intelligent agents (ordinarily, people). And not just by selected mods - it helps when typical user is well aware what is up and able to fend for himself.

downerending 7y ago

Discussion here may be kept to a higher standard, but that also means that a lot of things I'd be interested in are censored. Not sure what the solution is, but it would be nice to have some sort of ML-based approach that would tune the content I see to my wishes, rather than the wishes of the average denizen or the moderators.

dexen 7y ago
>a lot of things I'd be interested in are censored Same for me. Frankly I accept HN is a space to discuss a limited range of subjects, and keep other subjects to elsewhere. It's annoying to a degree, but it the longer run it works.