Why it's impossible to agree on what's allowed

In my experience, people disagree about the &quot;vehicle in the park&quot; game because they try to apply other rules that are not part of the game.<p>Ambulances, for example, are clearly vehicles, and if they&#x27;re in the park, the rules has been violated.<p>That doesn&#x27;t mean the ambulance driver is in the wrong, though. There could be other rules that supersede that rule. We aren&#x27;t told of any, though, and are simply asked about that one rule.<p>The only judgements that we&#x27;re asked to make are &quot;What is considered a vehicle&quot; and &quot;What is considered to be in the park&quot;. That&#x27;s because these things are not defined for us and are open to interpretation.<p>Interestingly, though, the game said I agreed with 74% of people, which was a lot higher than I expected it to be.

IMO another aspect that could have been highlighted is that really rule-enforcement is “part of the game,” despite any attempt to take it out, and people will engage in strategies that attack the rules.<p>There are lots of strategies that attack the rules in the pedantry space. One could post odious content that carefully is constructed to not break the rules. Or post odious content that breaks the rules in a manner similar to the rule-breaking of somebody you don’t like, to highlight uneven enforcement. Post rule breaking content, but of a type which is generally non-odious and general not banned. Try to find room to quibble about what exactly justified an exception.<p>I think the WWII tank is an interesting question. It is clearly a vehicle, so not allowed by the rule. But it is non-functioning, so it isn’t at risk of violating the most likely intentions of the rule (to prevent noise and injuries). It is an unusual artifact in that it is a monument to a war which is generally considered (to a basically unmatched degree) to have been just. If we allow the tank, do we allow an Iraq war or Vietnam war tank? We just need to solve the question of which wars are just, first, I guess, so to enforce our no vehicles bylaw we’ll have to solve all of politics. Or maybe add a “no war related monuments rule” which seems nice until a veterans group wants to donate a couple benches with little plaques on them.

The problem with the ‘no vehicles in the park’ example is that the question is <i>intentionally</i> vague. I feel like we do ourselves a disservice to hold this particular example up as a way to conclude that everyone disagrees and it’s impossible to agree and let’s throw up our hands. The way forward is to put more precision into the rule statement. And yes, there might be diminishing returns and ;it might even be impossible for every last person to agree, but that’s an actively harmful take-away if it’s possible to get 90% or 95% or 99% to agree by just being <i>slightly</i> more specific. A huge portion of the ‘no vehicles in the park’ questions are already answered by existing US aviation laws, for example. Another huge swatch of the questions would have been answered if the rule was ‘no <i>motor</i> vehicles in the park’. By intentionally withholding the actual rules, this survey doesn’t strike me as evidence that agreeing is hard, it seems like more of a trick question.

The presented trichotomy between no moderation, moderation, and federated moderation is false.<p>Moderation can also be accomplished via a user-level web-of-trust system, where each user can choose who to trust as a moderator, and this trust can propagate recursively to the people trusted by the people you trust, and at each level (even when manually choosing people to trust) this trust can be fuzzy (not full trust vs no trust, but potentially something in between those two), and rapidly decreasing the more distant you get from those you&#x27;ve manually chosen to trust. To solve the issues of spam, censorship, and convenience simultaneously, you simply assign to users some moderators on the trust list by default and allow users to opt out of that trust.<p>This approach is also applicable in the same manner to the similar problem of curation (i.e. choosing what to highlight instead of what to hide), where the same four approaches are also applicable with largely the same pros and cons.

I did the quiz, got 93%. The common sense factor should prevail in all circumstances.<p>First, An implicit assumption. No vehicle implies no motor vehicle.<p>Second: The purpose of the park, namely to allow people to enjoy nature.<p>Third, the rationale for the rule: A motor vehicle is much faster and heavier than every other entity using the park. Ergo, it has the capacity to cause great harm (ex. running over a soccer team).<p>One can deep-dive all manner of philosophical arguments, but the principle of least harm while allowing maximum freedom is the true, unspoken rule. Ergo, any vehicle that can co-exist without hampering or endangering others enjoying the park is okay.

I think this is overstating the problem by comparing unique bundles of policy opinions, rather than looking at agreements on a per-policy basis. Out of 27 policies in the game, 20 policies had overwhelming support (&gt;70% agreement) one way or another [1]; those are obvious candidates for moderation rules. And of the seven controversial policies, there was still &gt;60% agreement on five. There&#x27;s certainly room to disagree on the controversial policies, but I think painting this as meaning there&#x27;s less than 10% agreement overall, or that moderation is impossible, isn&#x27;t really accurate. Most people agreed on most things.<p>After all, we live in a society where we have ambiguous rules like these all over the place. Occasionally someone will get a weird-seeming enforcement action, or a weird-seeming lack of enforcement (compared to what the rules say — for example, while it might seem weird that SF does not punish thieves for under $900 of stolen goods from a common sense perspective, it isn&#x27;t weird from a letter of the law perspective: that&#x27;s the actual law); but in my opinion that&#x27;s fairly rare per-capita.<p>But maybe I&#x27;m biased since I had 100% agreement per-policy with the majority ;)<p>1: <a href="https:&#x2F;&#x2F;postimg.cc&#x2F;SjMbNMKW" rel="nofollow">https:&#x2F;&#x2F;postimg.cc&#x2F;SjMbNMKW</a>

It may be impossible, but those contrived games don&#x27;t illustrate that. For example, it sternly warns that the responders should disregard all rules from their day-to-day lives, which would simply not happen, people don&#x27;t clean their minds for some games and will apply other those other things when making decisions. Also, as is common in these games, they&#x27;re underspecified, so a lot of &quot;disagreement&quot; is just differences in terminology. So you can be in agreement on the policy (that this type of vehicle&#x2F;non-vehicle) should be allowed in the park while disagreeing on the definition<p>Also, if your failure mode is for someone to &quot;called for her to be physically assaulted, doxed, etc.,&quot;, then agreement is irrelevant, you can be in total agreement on rules and still call that for non-rule-specific reasons (people are complicated and have emotions)<p>But the main fail is in reducing agreement to a binary, so that&#x27;s why &quot;Exactly. There is a clear majority in the answers&quot; is correct an a recipe for having a broadly popular moderation policy

This is partly why I really liked the idea of Slashdot&#x27;s meta-moderation. Although it could probably be gamed by enough determined people.<p>Moderators should be moderated too, because just like humanity in general, there&#x27;s going to be a spread of skill levels among the mods.<p>I wonder if there&#x27;s ever been a site that has &quot;moderator karma&quot;, a score&#x2F;scores by which a moderator&#x27;s &quot;performance&quot; can be measured. History of that score too would be interesting to see, potentially showing if a mod has deteriorated in their judgement lately, stuff like that.<p>It feels like there could be more done to find useful data about moderation decisions.

The experiment is bad because it has hopelessly convolved a number of different causes of disagreement, some of which have nothing to do with difficulties in moderation.<p>If you&#x27;re interested in &quot;people disagreeing on what should be allowed&quot;, then you must state the rules in a way that allows them to establish the intent of those rules.<p>Instead the experiment was not evaluating people on their ability to agree on the intent of the rule &quot;no vehicles in the park&quot; but on their agreement over the meaning of the rules of the experiment itself.<p>Since <i>these</i> rules were intentionally left hidden, you should expect there to be lots of disagreement on what they are. Which was evident in the original comments section.

I got 93% agreement based on a simple rule: It&#x27;s a vehicle if it is touching the ground and powered with a motor.<p>And if I were moderating a community, which I have done a lot of in the past, I would clarify the rules as soon as this disagreement popped up. This isn&#x27;t a great experiment because it doesn&#x27;t involve any judgement calls -- you can apply a (pretty simple) objective ruleset here.<p>The main areas where I disagreed with the majority are a police car and ambulance. The police car and ambulance are clearly vehicles. People who said no were applying other rules. But again that rule is simple -- emergency services are exempt when responding to emergencies.<p>Moderating actual communities is a lot more difficult because there aren&#x27;t a simple objective set of rules you can apply. Judgement calls are necessary. We ran this experiment on reddit for a while. We would present the user with a link and ask them &quot;is this spam&quot;. It was rare for a link to get 100% (or even 98%) yes. Only the very most obvious cases. Otherwise most links would be closer to 60-70% if it were &quot;spammy&quot;. And only things like wikipedia would get close to 0%. Everything else in between was a spectrum.<p><i>That</i> is what makes moderating hard.<p>The conclusion of this experiment is correct, but I would say the methods are deeply flawed.

A bit more info on the origins of this. The original that David draws on is a hypothetical proposed by H.L.A. Hart in 1958. The original is looking specifically at law and the fact that even for &#x27;settled meaning&#x27; of a term or phrase there will still be the &quot;penumbra of debatable cases&quot;. (The penumbra being the almost-shadow between light and dark).<p>&gt; If we are to communicate with each other at all, and if, as in the most elementary form of law, we are to express our intentions that a certain type of behavior be regulated by rules, then the general words we use – like “vehicle” in the case I consider – must have some standard instance in which no doubts are felt about its application. There must be a core of settled meaning, but there will be, as well, a penumbra of debatable cases in which words are neither obviously applicable nor obviously ruled out.[1]<p>I&#x27;m most fascinated by it from a UX perspective in public space: how will a user respond to &#x27;penumbral&#x27; or edge-case data? Can that be used to refine wayfinding signage? How will a user react or change pathways when encountering information that doesn&#x27;t fall right in that expected area? (Yes, I&#x27;m actually wondering of the implications of <i>actual</i> signage, not just hypothetical.)<p>We see it often in transit systems. It&#x27;s where someone might misinterpret a sign or information (in a predictable manner) and latter blame themselves once they learn of the &#x27;intended&#x27; information (which will seem obvious in hindsight). Yet, they are not to blame.<p>[1] Harvard Law Review Vol. 71, No. 4 (Feb., 1958) - H.L.A. Hart - Positivism and the Separation of Law and Morals [p607] <a href="https:&#x2F;&#x2F;www.jstor.org&#x2F;stable&#x2F;1338225" rel="nofollow">https:&#x2F;&#x2F;www.jstor.org&#x2F;stable&#x2F;1338225</a> (note: it wasn&#x27;t hard to find a copy via Google)

The whole reason that we’re having such a disagreement on moderation is that moderation has been weaponized for political purposes.<p>This has happened over and over - Twitter, Reddit, Facebook, YouTube, Wikipedia, even here (try posting a skeptical take on climate news).<p>I don’t know what the rules should be. But I don’t think that we can have well informed people if a small subset are allowed to filter what ideas people are even allowed to consider.

If the question were, “how do you think this rule should be applied to a park in your neighborhood” I think agreement would much more achievable.<p>There is no clarity on what the context of the application of a rule is.<p>Is an ambulance a vehicle in the park? Of course. Should the rule apply to an ambulance on duty in an actual park? Of course not.<p>Is a memorial tank a vehicle in the park? Obviously. Would some people object the memorial? For sure. Likely because they don’t like tanks in the park, not because of the vehicle rule, however they may use the vehicle rule to object.<p>So - amazing data. Wrong question.<p>I propose that the differences in opinion are likely differences in understanding the ill-defined context of the question.

The problem most people here are missing is that even if moderators create a rule set that are clear and understood by 95% of people, the remaining 5% will still become an unbearable headache.<p>Add to this that spaces that are worth being highly visible in will have their rulesets attacked and rules stretched to the maximum, plus the fact that users individually are not worth much, and that most publicly visible moderators are unpaid, and you have a recipe for user misery and moderator indifference.<p>The solution is to pay your moderators so you can attract higher quality talent and more stable labor.<p>Or don&#x27;t, if enough users will use the platform anyways.

A converse to this is &quot;Don&#x27;t be the person that gets a new rule named after them.&quot; I find this one interesting. It&#x27;s a meta-rule that actually aims at the spirit of the rules. (And requires a violation or meta-violation to come into effect.)

&quot;You agreed with the majority: 100%&quot;<p>Which proves it&#x27;s not actually impossible, the majority opinion is actually sensible, and this take is incorrect. Case closed.

His ultimate point:<p>&gt; No large platform can satisfy user preferences because users will disagree over what content should be moderated off the platform and what content should be allowed.<p>But he makes to attempt to say what we should do about it. In my view, we should not try to make explicit coding of rules so precise as to be automatable, but rather make it clear, to the communities the platforms serve, what the values are that underlie the rules, and to say &quot;apply the rules in line with their spirit.&quot; This is the same sort of reasoning humans use in everyday life, it&#x27;s why you don&#x27;t get ticketed for jaywalking across an empty street, and it&#x27;s why courts get flexibility with their interpretations of law.<p>The value underlying &quot;no vehicles in the park&quot; is that a park should be a place for recreation and all-ages shared use. This implies the space should be free from the danger posed by the regular traffic of a busy street. A motorized wheelchair doesn&#x27;t pose this danger so it is allowed. Nor does a bicycle, unless we&#x27;re talking about a large amount of bicycles representing a heavy flow of traffic, because that traffic would be dangerous to kids in the park.

Related:<p><i>The rule says, “No vehicles in the park”</i> - <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=36453856">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=36453856</a> - June 2023 (1186 comments)

The implications of this problem are far broader than the article suggests.<p>Consider this classic essay entitled The Myth of the Rule of Law by Georgetown law professor John Hasnas.<p><a href="https:&#x2F;&#x2F;www.copblock.org&#x2F;40719&#x2F;myth-rule-law-john-hasnas&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.copblock.org&#x2F;40719&#x2F;myth-rule-law-john-hasnas&#x2F;</a><p>It begins with a series of questions about what’s protected under the First Amendment. None seem particularly tricky. The answers all seem fairly clear cut.<p>Except, you guessed it, they are all controversial. In fact, none of them stand on firm ground at all. The essay brilliantly shows that government is always “rule of men,” and “rule of law” is a mere diversion. This essay will blow your mind.

I wish there was a button I could hit to format this site into something readable about 8-10 inches wide.

In general the way to get out of this successfully is to have a moderation team and a process that (enough) users trust, and the way you do that is to just _not_ try to make everyone happy and accept that some people are going to disagree with the moderation decisions, and it&#x27;s fine if they just _leave_. It&#x27;s one of the reasons that I think the best subreddits and smaller independent online communities like hacker news and metafilter remain so stable for so long.<p>Facebook and twitter and the like are doomed to be in this trap forever because their pursuit of global growth means they _have_ to make everyone happy, and in the process make _no one happy_.<p>And this is also one of the reasons why, wisely, most western democracies try and stay out of the business of regulating speech entirely.

This is essentially a debate on legal interpretation. There are 3 traditional approaches:<p>1. Common law (determined by precedent), the UK system.<p>2. Civil law (determined by textual interpretation of a code), the US system.<p>3. Qadi law. (determined by a local judge who rules according to the local societal norms and current context, neither by text nor precedent), the Ottoman&#x2F;Islamic system.<p>[1] <a href="https:&#x2F;&#x2F;www.britannica.com&#x2F;topic&#x2F;qadi" rel="nofollow">https:&#x2F;&#x2F;www.britannica.com&#x2F;topic&#x2F;qadi</a>

I have enjoyed the game, although I find that it is really inconsistent with reality. The reality is that there is a judge, an authority, an enforcing body, and contingencies for failures of each of those.<p>One example is that in my house, you&#x27;re not allowed to use profanities around children. Who determines the profanity? I do. I&#x27;m the authority. I&#x27;m the judge. I&#x27;m also the enforcer.<p>In the event of a rule break, a warning is given, and if the rule is accepted, it isn&#x27;t spoken of again. If there&#x27;s pushback against the rule, it will include an explanation of the enforcement. For this to be successful, there has to be consistent enforcement, an explanation of the authority that I hold, and the prefect execution of judgment rendered.<p>My family knows that profanities exist, we patents use them at times privately, that people use them in different capacities, and for different reasons, yet under my direction they will not be used in my house. This isn&#x27;t a value judgment of a guest&#x27;s usage but one of teaching control over one&#x27;s tongue while under my authority.

My experience is that the fundamental issue is that people agree or disagree on things for the most part based on their perceived location relative to ‘the group’, not based on if they think the thing itself is actually reasonable or not. At least within some (surprisingly large) bounds.<p>Some people will fundamentally ‘center’ on what they think the group will think, others will ‘go more’, others will ‘go less’, and some will ‘go opposite’. A few will go ‘fully independent’, but they’ll only take ‘independent’ directions no one else has taken, which isn’t actually independent.<p>If modeled out with one set of ‘group thought’, I’d be surprised if it followed anything but a bell curve.<p>It’s why things are so all over the place right now online - the same process is happening, but with infinite apparent ‘groups’. So it’s scattering things more randomly than would historically have happened. Maybe even an inverse bell curve, actually.<p>Rationalization then happens after the fact.<p>It’s the Overton window, but what happens when there is not a single window - but everyone nearly has their own, custom window.

My abstract reading of the rule: we should minimize negative externalities while maximizing positive ones, while curtailing selfish behavior. IE, the ambulance is there altruistically; a riding lawnmower, similarly. The airplane has few negative externalities (to the visitors of the park), whereas a drone has more negative externalities and few positive ones. An ambulance that rides over the grass and tears up the soil better be in pursuit of an urgent need, a police cruiser similarly driving over the soil to stop that drone pilot has inadvertantely created greater negative externalities in pursuit of enforcing a problem. The rule is simple: don&#x27;t be a jerk, think about others, don&#x27;t be oblivious.

&gt; The idea behind the site is that it&#x27;s very difficult to get people to agree on what moderation rules should apply to a platform<p>There&#x27;s a big difference between &quot;what moderation rules should apply&quot; and &quot;how do I interpret this one vague rule.&quot; Questions like &quot;can I play with a toy car in this particular park&quot; are answerable. If some people want the rule to be different, that&#x27;s another story. People&#x27;s ability to interpret a vague rule out of context doesn&#x27;t really bear on the question of how hard or easy it is to agree on rules for a community, or to write good rules that are easy for moderators to apply.<p>Not every single person in a community needs to like a rule. And if rules are not specific enough, they can be made more specific.

The &quot;No Vehicles in the Park&quot; game is not so much about the impossibility to agree on what rules to set (like &quot;no vehicles in the park&quot;), as it is about the ambiguity of rules (what is a vehicle? where does the park begin and end? are emergency vehicles OK?).<p>Using a common sense intepretation, solves most of the &quot;insurmountable&quot; issues it presents, even though the way the game is setup is supposed to prove how that&#x27;s impossible, in practice, real parks, with real similar restrictions, don&#x27;t have those problems - or extremely rarely, and mostly with individuals with mental health issues which would contest the rules whatever they were.<p>Apparently when it&#x27;s not about contrived scenarios in test form, those issues don&#x27;t really occur.

I think the core problem with the authors analysis is that it does not ask for intensity of opinion.<p>I think when looking at &quot;choice&quot; * &quot;intensity&quot; results would be much more homogeneous.<p>I think if the author accounts for intensity, results <i>would</i> be homogeneous and therefore the thesis of fractal disagreement would have to take backseat to a more dominating thesis, such as: It&#x27;s hard to decide what&#x27;s allowed because those with power don&#x27;t want to have power exercised against them and use their power to gain concessions via technical readings of law rather than intent based readings of law.

The first footnote just turns me off: zero moderation means everything becomes 4Chan? How about the usual systems of vote ranking, reputation scores based on voting history to enable or disable posting power, etc? I think that a community can easily hit 90% of quality with just a few simple rules, a tiny number of human mods to handle a few egregious cases where someone obviously but cleverly circumvent said rules (like when people started spelling racial slurs through 1 letter tweet replies stacked), and a massive amount of crowd distributed power to reward what a critical mass wants.

The question is, why do you want people to agree on what is allowed in some kind of absolute fashion?<p>In the park example, the important is that most if not all users should be satisfied with the experience they live in the park most of the time. If your park is the sunny nudist Valley, probably the expectations of rules won&#x27;t be the same than those who live on the demure cold hill. And as far as everyone is free to go live wherever the rules feets best to their own general tendencies, you don&#x27;t structurally design your society to generate drama about mundane frivolous topics.<p>Of course if you want to be the platform that bind them all, for example because you think people are brainless cash cows, you are going to face soon or later the fact that rules for demure nudists doesn&#x27;t grab a very large audience.<p>The article does mention a bit fediverse, but then conclude that if you can&#x27;t have everyone in the same park, you failed. Having a single park with the same permanent rules for everyone will not satisfy even a single individual, because people have moods and interest that evolve over time. Yes it has a cost to go from one park to an other. Compare it with the cost it would imply to have a single park for everyone.

The vehicle game is poor metric, since it measures the general population, and doesn&#x27;t up-front make it clear whether the spirit or letter of the law should be applied.<p>People need some minimum life-experience as well before they can pass judgement, which many in the &quot;audience&quot; simply don&#x27;t have. I for instance find it easier to talk about a start-ups problems to someone that had run a restaurant for a few years than to a developer.<p>All that means is that moderation can be achieved, it just costs a lot and may not scale easily.

If you look this from a statistical point of view it turns out that people agree a lot. Out of all the 2**26 (~26M) possible sets of opinions only 9432 different opinions are generated. We have data for around 37k users, so if users opinion was so special we would expect around 37k different opinions, but we only have 4k.<p>It would also interesting to see how many different opinions are there if you allow fuzzy matching. Users disagreeing in only 1 or 2 questions probably have the same idea of what the rule means.

I am a bit confused by this since I got agreement with over 90% of people. Are others getting much lower scores? My result suggests that is indeed possible to agree :)

I wonder what percentage had my response, which was to read the first question, say Unsuffucient Data, and stop doing the quiz.<p>There is no such thing as a singular rule. They are embedded in a host of other rules either explict or as (cultural?) assumptions. What is a park, what is a vehicle, what is &#x27;in&#x27;? There are domains, like math, where effort is made to be completely explicit, but that doesn&#x27;t work everywhere.

A reasonableness test (like what a court system uses) makes this quiz really easy to answer. Basically everything is either allowed or a margin case except for the person driving their Civic in the park.<p>TBH, I think if you have this quiz to a non-engineer, they would come away wondering what the point is.<p>Perhaps the answer is that content moderation should just let a judge make a decision and maybe appeal?

&gt; the propensity of humans to pick on minute differences and attempt to destroy anyone who doesn&#x27;t completely agree with them hasn&#x27;t changed<p>This invites the question of whether the proximal cause for destroying The Other is the issue, or merely a pretext for destruction.<p>These situations can be as icebergs, with quite a bit going on out of view.

You would definitely get different results in another language. My best known language is Polish and the best translation of &quot;vehicle&quot; (pojazd) carries different connotations than &quot;vehicle&quot;. For example a plane or a boat is definitely not a &quot;pojazd&quot;.

at the core of this is language, which is both a lossy medium for transmitting thoughts and-appears to be designed to be-ambiguous. it doesn’t help too that we don’t study it anymore, and seem to be slowly gravitating towards a small core vocabulary that is called into all sorts of services. our words are in tension at all times. on common sense, i’d define it as the substrate of all of one’s experiences through life. it’s empirical. there’s no absolute common sense where there’s no absolute experience. and that’s why education and other forms of civilization-preserving socializations are important. where you find yourself in firm possession of common sense that your interlocutors seem to lack, be curious and learn about their life experience.

I think the park rule would work better if it were &quot;Don&#x27;t be an ass in the park&quot;.<p>I think most of us would agree on what that means. The guy parachuting into the park, for example, is clearly in violation, whereas the ambulance driver and kids playing with toys are ok.

Is there an easy way to make sites like these more readable? With the wall to wall text on a 32&quot; monitor, negligible margins, small text interspersed with giant graphs with giant captions, it&#x27;s a bit hard to move past the form and focus on function.

In retrospect at least this should have been obvious, because for hundreds (at least) of years we have had various systems of laws, and in all of them we&#x27;ve had to rely heavily on precedent and the rulings of human judges and juries.

If everyone could agree on behavior in a common space, we wouldn&#x27;t have politics, or even government. We would just all agree on something and then go about our lives. Internet moderation is just governance for the private sector.

On moderation in general. I think the main political problem isn&#x27;t agreeing to any set of rules, and what those rules apply to, but when one group&#x27;s definitions trump other groups definitions by fiat.

&gt; obviously police and ambulances (and fire trucks) doing their jobs don&#x27;t have to follow the sign.<p>not obvious. there could be weight limits such that heavy vehicles like that could become trapped.

Tabs or spaces though? <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=SsoOG6ZeyUI" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=SsoOG6ZeyUI</a>

Content moderation decisions are like fractal edges.<p>You think you can see all the edges until you zoom in and realize each edge is multiple edges, recursively.

I agree with the principle of what he&#x27;s saying, but using the game as an example exaggerates the level of difference here.<p>People disagreeing on whether an ambulance is violating the rule are mostly disagreeing on a technicality of whether it counts as a vehicle in the park if you only have this one rule, but in practice almost everyone would allow it as an exception. Thus, it doesn&#x27;t represent true disagreement for most.<p>It&#x27;s similar to how most message boards have some rule about civility and not being hostile&#x2F;insulting, but if someone shows up who&#x27;s unironically like &quot;wow Nazis sure are great right??&quot; then nobody has a problem with posters telling that person to fuck off.

The classic approach is precedent. That way once the call is made, for better or worse, everyone knows how the rule applies in that case.

<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=39299186">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=39299186</a>

I love the kind of games like the &#x27;no vehicle in the park&#x27; one.<p>Please someone send me more links to games like these

A bicyclist in Golden Gate Park once yelled at me for being on an e-bike, while pointing at a giant &quot;No Motor Vehicles&quot; banner at the entrance. I told him, actually, California vehicle code 24016 specifically states “an electric bicycle is not a motor vehicle.” He replied &quot;You&#x27;re a fu*king idiot&quot;.

&gt; The next most naive suggestion is to stop downranking memes, dumb jokes, etc., often throw in with a comment like &quot;doesn&#x27;t anyone here have a sense of humor?&quot;.<p>The fact that forums that don&#x27;t aggressively moderate for picspam (so-called &quot;memes&quot;) become dominated by picspam is exactly why any moderated forum needs to aggressively moderate such content if it wants to be a platform for intelligent discussion. On large social media where you&#x27;re effectively doing the moderation yourself (think Twitter), that&#x27;s why you have to mute or unfollow users who regularly post picspam.<p>To make an online community usable, you have to moderate it to systematically exclude low IQ users or the entire community becomes low IQ and completely unusable. Stupid content gets enormous amounts of engagement. This wasn&#x27;t as big of an issue in the past when the internet was less mainstream because stupid people weren&#x27;t on it in significant numbers.<p>A badly moderated community moderates for a particular set of opinions that the mods agree with rather than for IQ. That means it becomes filled with stupidity that people on that side agree with. Most moderated communities go that route because it&#x27;s easier to filter for opinion rather than intelligence but it makes them worthless for any kind of serious discussion.<p>Moderating for civility is one filter that works somewhat well because many stupid people resort to ad hominem attacks (think stupid nicknames for political opponents) as if they were arguments. But intelligent people are sometimes uncivil especially if they run out of patience arguing with an idiot. Things can also get heated between people who are friendly and intelligent if they have an irreconcilable disagreement on a topic both feel strongly about. Such debates often get wrongly confused for flame wars by mods so you do have to be careful about applying civility rules.<p>Of course, any social media platform that wants a large mainstream audience is not going to follow this approach to moderation because you&#x27;d be systematically excluding at least the bottom 85% or so of the population to maintain a platform that&#x27;s usable for high IQ people. If you assume 2 standard deviations above average as the minimum to be capable of intelligent thought, you&#x27;re looking at excluding 98% or so of the population.<p>In short, I disagree with the author&#x27;s opinion that most people don&#x27;t want a social media completely dominated by stupid &quot;jokes&quot; and picspam. I think that&#x27;s exactly what most people want which is why platforms that value user growth and engagement over quality of discussion tend to converge on being dominated by stupid comments, pictures and videos. The next step after that is to remove or shadowban stuff the advertisers don&#x27;t like which results in moderation based on viewpoint. Reddit is Reddit because that&#x27;s what is most profitable for Reddit. The only way out of that vicious cycle is for a community to not rely on ad revenue and care about maximizing intelligent conversation not engagement.

See also, the impossibility of AI &quot;Alignment&quot;<p>I think about this problem of coordinating human action pretty much constantly, and IMO it is the foundational challenge of all of humanity.<p>It&#x27;s beyond clear to me that aligning human desire (and codifying them) is mathematically impossible <i>even within the same individual over the lifetime of the person</i>.<p>Why? Human desires and actions are not coherent, predictable or consistent, even to the Human making the actions. Self reported desires and goals will diverge in extreme ways through the lifetime of an agent.<p>As a result, the foundational requirement of a stable and coherent system is an impossibility because the indivisible unit of measurement: The human, is not consistent with their goals.<p>If the individual agent&#x27;s goal criteria is not known, or worse is unknowable for future time steps (my claim), then there is no possible system that can be codified which does not impede an individual&#x27;s action vector in favor for some other structure that promotes another individual&#x27;s action vector, provided there is variability in the agent&#x27;s goal vectors.<p>Someone is going to be prevented from doing what they want to. Unfortunately there is no way to objectively determine a set of action-contexts that can be objectively agreed to satisfy all persons intended action-vectors, because of physics.<p>People have attempted many ways at addressing this.<p>Non-hierarchical formal structures rarely coalesce into long-lasting organizations because the actors typically reject any reduction to their autonomy on behalf of a larger organization. As a result they don&#x27;t grow and are very susceptible to environmental &quot;shocks.&quot; Examples of these are the existing small mutual-communities that generally never get much larger than the Dunbar&#x27;s Number of whomever the most powerful organizer in the community is (~few 100 people).<p>Hierarchical formal structures encode an opinionated version of how the agents within the system are limited in their actions to the benefit of a larger organizational structure. This has the benefit of growth and accretion of resources that would make the organization more robust to shocks, however in every case there are significant limits on the action-vectors of the agents in order to be part of the organization. Examples of these are religious faiths, which require significant adherence to social behaviors, and often have significant limitations on the behaviors of the actors to benefit from the group resources.<p>There really isn&#x27;t anything in between unfortunately. History shows that the latter types of groups will overrun all of the former types of groups when they are in competition for growth in some physical space.<p>There do not exist social structures that are robust to significant changes in the composition of the criteria for success that do not create social divisions that realign power.<p>I honestly don’t know what to do about this because it seems intractable to make a future society that is coherent and stable given that the foundational unit of society is not coherent and stable.

This is nearly identical to a thought experiment I created on manifold markets about a real boat marina which has a sign saying &quot;Boat Owners Only&quot;<p><a href="https:&#x2F;&#x2F;manifold.markets&#x2F;Ernie&#x2F;boat-owners-only-sign-correct-inter" rel="nofollow">https:&#x2F;&#x2F;manifold.markets&#x2F;Ernie&#x2F;boat-owners-only-sign-correct...</a><p><pre><code> 78% You can go in if your boat is in this marina right now. 77% If your boat here sinks can you go in? 71% If there is a fire and your boat stored here burns up partially and is not seaworthy, can you go in? 71% If your boat is here and you contracted a spot, but all legal record of that burned in a fire, and the only one who remembers who is still alive is you, and you paid cash, are you allowed in? 69% If your boat is here and you contracted a spot, but all legal record of that burned in a fire, are you allowed in? 66% If you owned a boat which is here, but it&#x27;s been molecularly exchanged for identical but different atoms by aliens, you can go in 57% Only official owners of a specific but unspecified boat located on the dock are allowed through 56% You can go in if you own a real live seaworthy boat now anywhere in the world. 56% If you privately own the company that owns the boat, you may enter 53% The dock owner is not allowed to go in, unless he is or is with a boat owner 52% If you stole a boat, parked it here with a legal berth lease contract, then left, and return, can you go in? 50% If you are a shareholder in the company that owns the boat, you may pass 50% If California becomes officially Marxist, where ownership is an exclusive right of the state, can you anyone go in at all? 50% If the 24 hour video surveillance of the marina is disabled, that invalidates that sign immediately above, creating a presumption that all the signs on the fence are false, and making it the case that only non-boat owners are allowed. 50% You can go through if you open the gate 49% You can go in if you own any kind of boat in any condition in the world, including toy boats, model boats, Lego boats, virtual boats in baldurs gate etc. 47% The gate will prevent all non-boat owners from passing. Guests and passengers must swim 46% The sign isn&#x27;t about who is allowed through, it&#x27;s about the contents of what&#x27;s on the other side. Everything beyond the fence is a Boat Owner. 45% You can go in if you have no boat, but plan to buy one someday and have a contract for a reserved berth space 44% If you own 1% of a boat here you can go in 39% You can go in if your spouse is a boat owner 38% If you own half a boat stored here legally you can go in 34% You can go in if you own a Binary Oxidizing Acetylitic Thermometer. 34% Bonus: people who own three boats stored here can alternate sleeping arrangements so that in any seven day period they never sleep in one more than 3 days, legally? 34% This market will entirely be excluded from leagues 34% You can go in if you are a former boat owner but have converted it to a sailplane, which is here. 34% You can go in if you are a leashed dog that doesn&#x27;t own a boat, but is with a boat owner 34% You can go through if you have a contracted and paid berth here. 34% You can go in if you have a berth contract but are behind in payment. 32% You can go in if you have a rental boat stored here. 32% You can go in if you are a guest of someone who is a boat owner. 31% Ghosts are allowed because they say B.O.O. (Boat Owners Only), which is the password 25% You can go in if your grandpa is a boat owner 20% You can go in if you are ex navy. </code></pre> From being on that site for about a year, I learned a lot about how people think of laws, judgments, legalism etc. There isn&#x27;t a consensus and it leads to lots of arguments. I personally am a lot less positive on legalism as the answer to everything, yet don&#x27;t really know any other systems I feel safe with, either.

dag - can I request a temporary pass on accusing people of not reading the article? I jest of course - but this comment thread seems to contain a lot of posts that appear to have missed the point (especially the &quot;vehicles in the park&quot; responses where few acknowledge the sections that call out HN specifically)<p>I wonder if there&#x27;s something about the points raised in this article that the typical HN denizen is constitutionally predisposed to &quot;not getting&quot;. The overall conclusion is along the lines of &quot;stuff is complicated and throwing rationalityh at it won&#x27;t help&quot; and &quot;the solution (if there is one) would be a messy, partial, human solution&quot;.

This article purports to be a defense of moderators against people who complain about rules not being specific enough, but I think it is equally valid to shove this argument in the face of a moderator who points at a rule and shows you the door for why their supposedly simple rule should never have carried such weight in the first place: no, it is not at all obvious that this thing you didn&#x27;t like was a violation of your rule, as your rule was inherently vague and apparently encoded some hidden value system. It also, to me, strongly argues that computers and automation have little place in moderation and any such usage--in either direction--must be carefully and routinely scrutinized and evaluated by humans and preferably only used as tools by decision-makers (as opposed to first-line decision-makers in their own right).<p>The lesson that I think more people need to take from this is thereby to have more compassion for people who broke some supposedly-clear rule, by (at least) reconceptualizing the role of moderation by: 1) having multiple checks and balances on moderators, including both an easy way to appeal decisions to a preferably-higher and necessarily-separate power and a way for users to hold moderators accountable for their actions; 2) insisting that enforcement happen in public; 3) providing numerous examples of what has been allowed or not in the past (precedent is not to be confused with epicycles); 4) offering a way for users to ask, without judgement, whether a specific thing <i>would</i> be allowed before they take an action (transferring liability on the punishment off of the user, even if later on the action is deemed an offense); and 5) instead of focusing on listing simple &quot;rules&quot;, spend your time documenting what your goals are along with your philosophy of enforcement.<p>Note: I actually am an elected government official for a local government--not the Recreation and Parks District, but I do work closely with them--and I am the chair of our policy committee, in addition to having spent a decade as a key moderator of a very large ecosystem of software developers, working on rules for both the software people could distribute&#x2F;sell as well as doing live moderation of chat rooms. In so doing, I got to work with other people whose opinion on moderation I came to trust deeply, and spent an incredible amount of time just having discussions about how to deal with moderation in general as well as how to balance it inside of a system where I also insisted that I could not myself become a centralized dictator, which is how people always seem to want to approach the problems of content moderation, ignoring any chance to empower users in their own paths. I thereby want to be clear: I come at this as someone who isn&#x27;t trying to blanket defend people who like to abuse either the system or other people... and yet I also don&#x27;t think the way most people approach moderation actually solves that problem in a reasonable or principled way.<p>I thereby think it is worth trying to use this analogy a bit to show how current moderation systems seem to often work: you want to bring your toy car to the park, but see a sign &quot;no vehicles in the park&quot;, and you want to be a good citizen so you call the office of the local park district to ask if this is OK; 99% of the time, you get an automated message to read the sign, and 1% of the time you get a person who says &quot;I am sorry but if I told you whether or not that toy car is a vehicle, then bad actors will destroy our park&quot;. You then show up to the park with your toy car -- sure it couldn&#x27;t possibly be a vehicle -- and a police officer shows up, throws you in jail for the day, and tells you you are banned from the park for life. You ask what you did--&quot;OMG, was it the toy car? I am so sorry I didn&#x27;t think it was a vehicle&quot;--and the officer looks you dead in the eye and says they can&#x27;t tell you what you did; or, if they do say it was the toy car, they refuse to engage in any form of discussion of why this toy car is somehow different from all of the other toys you see strewn around the park at that same moment.

this shows why it&#x27;s impossible <i>for &quot;HN&quot; readers and like-minded people</i> to agree on anything. the <i>entire</i> rest of mankind isn&#x27;t bothered by such nonsense:<p>- the bureaucrat that comes up with the sign does not carefully consider the wording of the sign nor the ordinance, it&#x27;s frequently more than a little vague, but it doesn&#x27;t matter. his point is to assert control over something, no matter how trivial. a long complicated, precisely worded sign is unhelpful to him, simple words and having a good amount of ambiguity are both to his benefit.<p>- the guy that does whatever the sign says not to on purpose parks his lifted truck on a boulder in the middle park after ripping up the grass a little. if the police were around he wouldn&#x27;t do this, but they aren&#x27;t, so he is definitely not a coward or a conformist at all. his girlfriend regrets her choices.<p>- the police are who enforces the rule, so it&#x27;s irrelevant whether they are breaking the rule or not. the other emergency services operate, in the police&#x27;s opinion, under their protection&#x2F;direction, so they&#x27;re usually ok too.<p>- the pedant pulls out the oxford dictionary and argues about whether a stroller or the aircraft or tank on display technically violates the rule. nobody notices unless another pedant shows up and starts quoting HN statistics or arguing the virtues of webster&#x27;s vs. oxford. it&#x27;s irrelevant anyhow because neither of them know how to drive, which is honestly the ethical thing. an argument breaks out about whether the reason it&#x27;s ethical is because of pedestrian fatalities or global warming.<p>- nobody cares whether the tank is a violating the rules because it weighs 50 tons (ok damnit you pedant 42 tons), the owner is the city who isn&#x27;t going to fine itself, and nobody is going to tow that thing.<p>- karen shows up and tries to tell a much more socially adept mom that her stroller is a vehicle and against the rules. when the karen is first ignored and then told <i>exactly</i> where to stick it, thanks to tik-tok the community has an ad-hoc emergency real-time vote about the issue and the consensus is that karen is burned at the stake, metaphorically (usually).<p>- everybody else notices (or doesn&#x27;t) the sign, doesn&#x27;t read it that carefully, and just mostly does what everyone else does anyhow.<p>- the park department eventually puts up bollards that effectively define &quot;vehicle&quot; as things that do not get stuck on and fit through the bollards. Excepting the things that got lifted in there by a crane or have the key to the gate.<p>in actual fact, everybody does sort of agree in real life. even the assholes that don&#x27;t follow the rules usually know it (i mean, that&#x27;s the whole point). a few folks are truly oblivious, but they usually just do what everyone else does anyhow.

Moderation is a bad idea, especially on places like hacker news. Blasphemy.

The vehicle in the park game seemed very obvious to me with a 100% result. I’m having a hard time imagining what others could have disagreed with.<p>A vehicle is loosely defined as a human-sized ( but not necessarily human operated ), generally functioning motorized transport device. Emergency vehicles in the park technically violate the parks rule ( even though the violation is basically rendered meaningless ).<p>A tank would be a vehicle even if it was inoperable, but not if it was intentionally made inoperable. It ceases to be a vehicle and instead becomes a display piece, memorial, etc.<p>In the park means any land or water within the parks geographic bounds, but definitely not airspace.<p>A toy is a toy. It may be a replica of a vehicle but it is not a vehicle. A bike, skateboard, rowboat, etc are not motorized and are therefore not a vehicle in the most common colloquial usage.