This app was a really really bad idea from the start. PG and YC often tell people to look for ideas that "seem bad" but this is a whole new league. Even if used properly by 100% of people 100% of the time it would still be a bad idea and fundamentally could never have worked. I"m just surprised someone ponied up the cash to see it realised - perhaps theres an ulterior motive. I don't know.
> Even if used properly by 100% of people 100% of the time it would still be a bad idea and fundamentally could never have worked.
Then how do you explain the success of Yelp, which is basically the same thing but for businesses? Or the school system, which is already designed around rating people? Or safeorscam, which was the same thing but for drug dealers? Or LinkedIn testimonials?
Certainly there are a ton of challenges to coming up with a good implementation — business, social, epistemelogical, etc. But I think the possibility space is too large to rule out the possibility of a feasible implementation.
People != Corporations. Although I can see some similarities between Yelp and Peeple, they are for rating/reviewing two fundamentally different entities.
> Then how do you explain the success of Yelp, which is basically the same thing but for businesses?
Businesses aren't people. You can't bully a business into suicide, for example.
> Or the school system, which is already designed around rating people?
On academic merits. And sure, there's social rating, but it's not a public Internet page saying "untrustworthy liar" that anyone can find.
> Or safeorscam, which was the same thing but for drug dealers?
See businesses.
> Or LinkedIn testimonials?
Perhaps a fair comparison, but that's not evaluating objective worth or something, it's evaluating employability and is as cordial as business tends to be.
There is a significant difference between rating someone in a market context and rating someone in a social context. People hold different norms between the two, and what Peeple failed to realize is the reaction people have when those norms cross over and are violated.
Imagine rating your friend's dinner party the same way you'd rate a restaurant on Yelp, and you can get a sense of the difference.
> There is a significant difference between rating someone in a market context and rating someone in a social context.
And that's why I said that coming up with an implementation that accounts for these sorts of issues would obviously be challenging. These specific founders may or may not be smart enough to do it, but that doesn't say anything about the inherent possibility of doing so.
I'm not saying I have a good solution, because I don't, but I'm also not willing to bet that there is no solution out of hand. And while I fully expect that this startup won't be successful, if only because of the difficulty of the problem, I also don't think it's fair to trash talk the founders without at least taking a look at the actual mechanics of their system.
People are extremely sensitive about being judged.
If someone had posted something ruinous for someone else's reputation, which also happened to be false, there's always libel lawsuits. Beyond that, to quote Mrs. Feynman's question to her famous husband, "what do you care what other people think?"
It's a whole lot more than 'sensitivity'. It's an anonymous abuse engine which holds no merit to begin with. What, exactly, would be the benefit of such an app?
"Hey, I just met this great girl and she seems nice - oh, here on Peeple I see that 27 people in high school said she smells funny and one person says she murdered a baby."
With no verification or checks, what exactly is the point of this app, other than to be used to abuse people to suicide, ruin marriages, get people fired or other similar nefarious tactics? Are people really going to be all that interested in using an app to say 'hey this dude seems swell'.
What's going to prevent people from accusing random men of pedophilia, or random women of cheating on their husbands, or saying some person uses cocaine all of the time at their employer, without proof of any of these things or even a real connection to someone?
>What, exactly, would be the benefit of such an app?
Extortion, blackmail, peer pressure and public humiliation come to mind. I don't know what the benefit of honest reviews would be for anyone, though.
This certainly seems like an idea better suited for the black market, however, since all the ways I can think of to make money at it are bound to be illegal somewhere, if not everywhere.
> Then how do you explain the success of Yelp, which is basically the same thing but for businesses?
Yelp is successful as a business, but I consider its impact on businesses to be negative, because Yelp has a perverse incentive to let businesses influence Yelp reviews. Naturally, businesses complain about Yelp, even though consumers like it. For the same reason, people would complain about a rating system for people.
> Or the school system, which is already designed around rating people?
This is perhaps the most toxic and least successful part of the modern education system.
There's no way this idea was completely new. I think anyone who mulled this over would have realized all of the legal implications as well as how pissed off most people would be. However you could probably make a similar app where its all good things only. "Compliments.com check and see if anyone has complimented you!"
From what I read, they are building this app with the intention of it being like LinkedIn where the hope is for positive referrals. But of course we know there will be negative reviews. You can't leave anonymous reviews, which could be some incentive for most people to be decent. Of course there's a million things that could go wrong. And, if nobody pays attention to the site then it won't work anyway.
Why do people seem to overwhelmingly consider this such a bad/despicable idea? Doesn't it seem like it would prevent some motivation for people not to be assholes? Are people worried about inevitable misuse (bullying?)?
We all do this rating in real life. Sometimes people talk about their ratings of other people with each other ("yeah, I know john, he's really annoying"). You know which of your friends are good and which are not... Why not make this open knowledge?
Anonymous people can't leave bad reviews was their plan to curtail that.
Also, anonymity means people are less likely to self-censor. So you get more honesty. Sometimes honesty isn't nice. If you want statements that are only ever nice things to say, you'll have to accept that many of them are lies.
From what I understand, you can't rate people anonymously. Of course, this probably doesn't prevent someone from going to the effort of creating a fake profile, but I don't think anonymity is the primary factor here.
Suppose Alice tells me that Bob is flaky and can't be relied on. I need a lot of information in order to be able to evaluate that claim.
* Maybe Alice and Bob were in a relationship a few months ago, and there was a bad breakup. Alice's experience of Bob would probably be different from my experience of Bob.
* Maybe Alice also told me bad things about Carol and Dan, so I think that Alice just has a dim view of people in general, and isn't willing to give them the benefit of the doubt.
* Maybe Alice heard everything from Eve, so I need to consider whether I trust Eve.
* Maybe Bob got chosen over Alice for a promotion at work, even though Alice has seniority. So Alice might have an axe to grind.
* Maybe Alice hangs out with Frank and Grace, who bully Bob. So Bob naturally wouldn't want to be friends with people who bully him, even though Alice isn't doing anything wrong.
When I get this information from Alice in person, I can ask questions or just listen to Alice's tone of voice to figure out the kind of things I want to know about Bob. Putting the information in text form in an application makes it much harder to evaluate Alice's claims, because it only make's Alice's opinion open knowledge.
Suppose the app was smart enough to do some clever analysis.
For example, if all the people who leave bad reviews of Bob are all friends with each other, it should count as a single bad review (so as to not double-count the evidence), and if Alice left bad reviews of several of your friends, it should know that you will probably find Alice untrustworthy, because she dislikes the people you like.
Not that we can trust Peeple or any other app to get that right, but it sounds like it should be possible.
> Suppose the app was smart enough to do some clever analysis.
When Alice gives a bad review of Bob, I think to myself, "Why is Alice giving a bad review of Bob?" I'm putting myself in Alice's shoes. There is absolutely no hope that a computer program would be sophisticated enough to do that, not yet, not for at least a decade (translation: possibly not in my lifetime).
Computers can't even solve the "simple" problem of matching people in romantic relationships. When we throw computers at the more general problem of human interactions and trust, we should naturally expect them to completely botch it.
This is something I was thinking about as a solution for traffic violations: have drivers rate each other; increase the "trustworthiness factor" of anyone who thumbs down few people and those people are also negatively rated by others and/or have accidents; decrease the multiplier for those who just negatively rate everybody and so on.
Unfortunately, I think you can hit some mines related to https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theore... but who knows? Maybe it won't happen given that you don't want a transitive relation, you want a numeric score for each driver. (Such a score could be used to influence insurance premiums; as an anarchist, I'm generally thinking about solutions for problems like "suppose we don't have patrol cops".)
I see this as a bad idea because it is personal and it is in print. When people have a bad experience with each other, time tends to heal the situation and many learn from the experience and most move on. Perhaps many situations could be chalked up to personality chemistry or bad timing or health issues.
When the experiences are put in print and published, they drudge up the past and make a dynamic thing a static thing. The printed version of a bad experience won't heal and forgive as people do and the rating won't reinvent itself as people do.
Even credit history is not kept for more than seven years if I'm not mistaken. People do change.
(1) People ratings are far too subjective to be considered "knowledge". We have none of the context needed to evaluate the validity of these ratings—the kind of thing you can suss out when gossiping in person.
(2) People can and do change, but these ratings would persist. It makes fresh starts and self-improvement that much more difficult.
(3) People can and do forgive, but these ratings would serve as permanent reminders of bad deeds.
(4) What problem do these ratings solve? I see many problems created, but with no real benefit.
Criminal records have (almost) the same problems; what makes them superior?
Incidentally, you could solve (2) by "dimming" the relevance of a review as time goes on, eventually clearing it altogether. You could also have a "faster dimming speed" for younger people than for older ones, on the grounds that the young are stupid and the old are probably harder to change.
Criminal records are presumably based upon written laws, public evaluation of collected evidence, trial, and conviction. That process doesn't always arrive at the right outcome, but at least it strives to be correct. And later, you can find those records and make up your own mind about the person with the rap sheet, if needed.
That can't be the case with personal ratings, because there's no public evaluation or context or validation of each rating.
You talk about "dimming" as if it solves a problem, but what overall problem do personal ratings solve, really? "Dimming" is a solution patched on top of a non-problem. I don't get it.
To quote John Oliver: "Have you ever been on the Internet?" Because people are vile and and love whitchhuts. They love destroing social reputation and status, witch is you litteraly need to survive.
Some friends in Miami built this same concept 8 years ago. It was called TheGorb.com (Good or Bad). They hired PHDs to come up with algorithms for ranking, and all sorts of complex variables to factor. In the end no one cared or wanted to rate others enough to make it stick.
John Oliver talked about this site last night on his show. What caught my attention is that it seemed like the founders were clueless, but I only know it from the show.
45 comments
[ 3.4 ms ] story [ 96.8 ms ] threadThen how do you explain the success of Yelp, which is basically the same thing but for businesses? Or the school system, which is already designed around rating people? Or safeorscam, which was the same thing but for drug dealers? Or LinkedIn testimonials?
Certainly there are a ton of challenges to coming up with a good implementation — business, social, epistemelogical, etc. But I think the possibility space is too large to rule out the possibility of a feasible implementation.
Businesses aren't people. You can't bully a business into suicide, for example.
> Or the school system, which is already designed around rating people?
On academic merits. And sure, there's social rating, but it's not a public Internet page saying "untrustworthy liar" that anyone can find.
> Or safeorscam, which was the same thing but for drug dealers?
See businesses.
> Or LinkedIn testimonials?
Perhaps a fair comparison, but that's not evaluating objective worth or something, it's evaluating employability and is as cordial as business tends to be.
Imagine rating your friend's dinner party the same way you'd rate a restaurant on Yelp, and you can get a sense of the difference.
And that's why I said that coming up with an implementation that accounts for these sorts of issues would obviously be challenging. These specific founders may or may not be smart enough to do it, but that doesn't say anything about the inherent possibility of doing so.
I'm not saying I have a good solution, because I don't, but I'm also not willing to bet that there is no solution out of hand. And while I fully expect that this startup won't be successful, if only because of the difficulty of the problem, I also don't think it's fair to trash talk the founders without at least taking a look at the actual mechanics of their system.
Extorting businesses to suppress bad ratings, probably.
> the school system
Do you really think the quality of education is "successful" or something to be proud about?
> LinkedIn testimonials
That's a source for good reviews, not mixed reviews.
If someone had posted something ruinous for someone else's reputation, which also happened to be false, there's always libel lawsuits. Beyond that, to quote Mrs. Feynman's question to her famous husband, "what do you care what other people think?"
"Hey, I just met this great girl and she seems nice - oh, here on Peeple I see that 27 people in high school said she smells funny and one person says she murdered a baby."
With no verification or checks, what exactly is the point of this app, other than to be used to abuse people to suicide, ruin marriages, get people fired or other similar nefarious tactics? Are people really going to be all that interested in using an app to say 'hey this dude seems swell'.
What's going to prevent people from accusing random men of pedophilia, or random women of cheating on their husbands, or saying some person uses cocaine all of the time at their employer, without proof of any of these things or even a real connection to someone?
Extortion, blackmail, peer pressure and public humiliation come to mind. I don't know what the benefit of honest reviews would be for anyone, though.
This certainly seems like an idea better suited for the black market, however, since all the ways I can think of to make money at it are bound to be illegal somewhere, if not everywhere.
and do you really assume school rating system is a correct method?
Yelp is successful as a business, but I consider its impact on businesses to be negative, because Yelp has a perverse incentive to let businesses influence Yelp reviews. Naturally, businesses complain about Yelp, even though consumers like it. For the same reason, people would complain about a rating system for people.
> Or the school system, which is already designed around rating people?
This is perhaps the most toxic and least successful part of the modern education system.
Kind of a combination of involuntary facebook account and unselective LinkedIn testimonials.
This isn't even a new idea. There's apps to rate ex-boyfriends and such that are obviously focused on being negative. Just for example http://www.ratedexes.com/add-person/john-carreiro-251332-vic...
We all do this rating in real life. Sometimes people talk about their ratings of other people with each other ("yeah, I know john, he's really annoying"). You know which of your friends are good and which are not... Why not make this open knowledge?
Also, anonymity means people are less likely to self-censor. So you get more honesty. Sometimes honesty isn't nice. If you want statements that are only ever nice things to say, you'll have to accept that many of them are lies.
Suppose Alice tells me that Bob is flaky and can't be relied on. I need a lot of information in order to be able to evaluate that claim.
* Maybe Alice and Bob were in a relationship a few months ago, and there was a bad breakup. Alice's experience of Bob would probably be different from my experience of Bob.
* Maybe Alice also told me bad things about Carol and Dan, so I think that Alice just has a dim view of people in general, and isn't willing to give them the benefit of the doubt.
* Maybe Alice heard everything from Eve, so I need to consider whether I trust Eve.
* Maybe Bob got chosen over Alice for a promotion at work, even though Alice has seniority. So Alice might have an axe to grind.
* Maybe Alice hangs out with Frank and Grace, who bully Bob. So Bob naturally wouldn't want to be friends with people who bully him, even though Alice isn't doing anything wrong.
When I get this information from Alice in person, I can ask questions or just listen to Alice's tone of voice to figure out the kind of things I want to know about Bob. Putting the information in text form in an application makes it much harder to evaluate Alice's claims, because it only make's Alice's opinion open knowledge.
For example, if all the people who leave bad reviews of Bob are all friends with each other, it should count as a single bad review (so as to not double-count the evidence), and if Alice left bad reviews of several of your friends, it should know that you will probably find Alice untrustworthy, because she dislikes the people you like.
Not that we can trust Peeple or any other app to get that right, but it sounds like it should be possible.
When Alice gives a bad review of Bob, I think to myself, "Why is Alice giving a bad review of Bob?" I'm putting myself in Alice's shoes. There is absolutely no hope that a computer program would be sophisticated enough to do that, not yet, not for at least a decade (translation: possibly not in my lifetime).
Computers can't even solve the "simple" problem of matching people in romantic relationships. When we throw computers at the more general problem of human interactions and trust, we should naturally expect them to completely botch it.
Unfortunately, I think you can hit some mines related to https://en.wikipedia.org/wiki/Arrow%27s_impossibility_theore... but who knows? Maybe it won't happen given that you don't want a transitive relation, you want a numeric score for each driver. (Such a score could be used to influence insurance premiums; as an anarchist, I'm generally thinking about solutions for problems like "suppose we don't have patrol cops".)
When the experiences are put in print and published, they drudge up the past and make a dynamic thing a static thing. The printed version of a bad experience won't heal and forgive as people do and the rating won't reinvent itself as people do.
Even credit history is not kept for more than seven years if I'm not mistaken. People do change.
(1) People ratings are far too subjective to be considered "knowledge". We have none of the context needed to evaluate the validity of these ratings—the kind of thing you can suss out when gossiping in person.
(2) People can and do change, but these ratings would persist. It makes fresh starts and self-improvement that much more difficult.
(3) People can and do forgive, but these ratings would serve as permanent reminders of bad deeds.
(4) What problem do these ratings solve? I see many problems created, but with no real benefit.
Incidentally, you could solve (2) by "dimming" the relevance of a review as time goes on, eventually clearing it altogether. You could also have a "faster dimming speed" for younger people than for older ones, on the grounds that the young are stupid and the old are probably harder to change.
That can't be the case with personal ratings, because there's no public evaluation or context or validation of each rating.
You talk about "dimming" as if it solves a problem, but what overall problem do personal ratings solve, really? "Dimming" is a solution patched on top of a non-problem. I don't get it.
To quote John Oliver: "Have you ever been on the Internet?" Because people are vile and and love whitchhuts. They love destroing social reputation and status, witch is you litteraly need to survive.
https://www.youtube.com/watch?v=TnZa18F8QW4
submitted here:https://news.ycombinator.com/item?id=10331184
Edit: That is about the whole peeple outrage.
Community s05e08, "App Development and Condiments" https://en.wikipedia.org/wiki/App_Development_and_Condiments
On video[2]: "I wouldn't want this app to just be positive (...) we don't live in a fairtale land (...) it would be pointless if it was all positive."
[1]: https://linkedin.com/pulse/julia-cordray-ceo-peeple-creating...
[2]: https://vid.me/ukIW