What does a TRUMP presidency look like?

[DGuller]

It's much more fair than the alternative, considering the relative power of the parties.  An employer hiring a fuckup is of course bad for them, and unlike many these days, I do believe that employers have a right to not get fucked. 

However, the alternative situation could lead to an employee being frozen out of their industry by their vindictive former employer, and that former employer may be vindictive as part of a strategy to keep their non-yet-former employees more pliable.

But this rental price app you're criticizing has not been shown to be vindictive.  It is presumably impersonal.

I don't get your point.

[Admiral Yi]

I don't get your point.

You object to bad references because they have the potential to be based on vindictiveness.  This objection can not apply to the rental app.

[DGuller]

I don't get your point.

You object to bad references because they have the potential to be based on vindictiveness.  This objection can not apply to the rental app.

One can have many objections to coordinated economic actions.  The don't all have to apply to every example of them.

[Admiral Yi]

One can have many objections to coordinated economic actions.  The don't all have to apply to every example of them.

Sure, but you're basing your argument on the basis of analogy.  "We object to this, therefore we must object to this."  If there are fundamental differences between the things being analogized then it doesn't work.

[Sheilbh]

This is absolutely dystopian shit:
https://www.reddit.com/r/CringeTikToks/s/PNokJGs7RN

A bunch of armed and masked men with no convincing identification thst they are police, put a child into an unmarked civilian car.

Like, the fuck?!

Edit: maybe it's an adult not a child but it only makes marginally less terrifying.

Also the Chicago video. I've said before I think that ICE is lawless and that's not an accident or bad apples. It is the policy. I'd add again that I think this is "American" in the broadest sense - masked paramilitary agents of the state pulling people off the streets, acting on dubious legality etc. I can't help but think of this happening in the rest of the America's in the 20th century fight against Communism and the left and Aime Cesaire's line about how the violence of imperial powers returns to the metropole. How the fascist regimes turned violence and mechanisms of repression onto Europeans that had previously been used by Europeans only on non-Europeans, discretely, imperially. I think Orwell made similar points.

Someone earlier made the comparison with stormtroopers - I think people are talking about conditioning the military to fight like "warriors" in American cities. I'm not sure that'll work. My read is that the American military is committed to its constitutional norms, civil government etc and that is deep in their culture (to an extent that I think they're possibly the last standing institution with a sense of "non-partisanship" about them). But if you need a paramilitary force to go out and be the willing agents of coercion - then I think that's what ICE are absolutely being conditioned to do.

I also think that the right had a weird idea of the partisanship of the military. I think this was Hegseth's pitch - that officers might not like him but the rank and file would love having a real man/"warrior" like him as a leader and he'd basically be able to forge that connection over the heads of officers if necessary. In effect, that the military was ready to be MAGA-fied, they just didn't know it yet. I don't think that's true or that it's happened. By contrast I think ICE, in part because of the political pushback against them (which is correct) are absolutely politicised and aligned with what Trump wants.

[Zoupa]

I don't get your point.

You object to bad references because they have the potential to be based on vindictiveness.  This objection can not apply to the rental app.

It can. A vindictive landlord could start the snowball effect. "The app" is not an omniscient impartial AI.

[DGuller]

One can have many objections to coordinated economic actions.  The don't all have to apply to every example of them.

Sure, but you're basing your argument on the basis of analogy.  "We object to this, therefore we must object to this."  If there are fundamental differences between the things being analogized then it doesn't work.

Let's rewind and follow the whole sequence of events.  Let's also allow for the possibility that every time you take an analogy somewhere else, it sometimes helps to give local reasoning, which may not carry back three steps back to the origin.

Here is the sequences:

Yi:  Guy welched on me.  Why can't I boycott him?
DG:  You alone cannot boycott someone.  There has to be a coordinated effort for there to be a boycott.
Yi:  I tell someone else the guy welched on me, I guess then it is a boycott?
DG:  Yes, take whispers far enough, and it can very well be.  That's why employers don't whisper about former employees, because it can be taken as a boycott.
Yi:  You think it's fair and just for employers to not be able to whisper about their employees?
DG:  Replying specifically to the situation with employers, yes, I argued it's more fair than the alternative.  I didn't think far enough to realize that trying to find the best reasons for the employer example would come back to bite me, hence I tried to find the best reason that specifically applied to references.

Let's focus on the common thread rather than the fact that compounding the natural imperfection of analogies can eventually add up.  What's common in all the analogies is that the decision of one market actor, when communicated to other actors, can create a coordination that freezes someone out.  One landlord evicting the tenant gets communicated through background checks and prevents the tenant even from an opportunity to get another place.  One gambler refuses to deal with another and badmouths them to other gambles so that they also refuse to deal with them.  One employer decides to fire an employee and through references makes them appear to be a bad hire to other employers.

The justification for spreading the information is secondary; what matters is the resulting coordination that turns private discretion into systemic exclusion. Once that happens, the market stops being free in any meaningful sense.

[Admiral Yi]

It can. A vindictive landlord could start the snowball effect. "The app" is not an omniscient impartial AI.

Perhaps we should ask DGuller for some info on how it works.  Do landlords rate their tenants or does it record evictions?

[Razgovory]

When I had to find a new place my eviction could be found with a simple search on the Missouri court site.  Basically no one would rent to me.  My options were living in a motel, or a senior living facility.  I picked the senior living facility.  It is certainly not ideal

[DGuller]

It can. A vindictive landlord could start the snowball effect. "The app" is not an omniscient impartial AI.

Perhaps we should ask DGuller for some info on how it works.  Do landlords rate their tenants or does it record evictions?

I can't speak authoritatively specifically about how landlords work, but I have quite a lot of professional experience with information vendors in general.  Aggregating the information and selling it is big business, and it tends to be a monopoly or oligopoly because the network effect is in play (usually you have to give data to get data, although everyone can scrape the public records).  Credit bureaus were the first ones to the game, but now there are vendors for almost anything you can think of.  It makes much more sense for employers, or landlords, or anything else, to pay a small fee to someone else to do the background check, rather than do it themselves.

The next logical step in the sequence is to not just sell data, but also sell scores.  All these histories of credit card payments, mortgage payments, car payments, etc., are messy and somewhat unstructured.  A score like 690 is easy and quick to understand.  I don't know for a fact if there is such a thing as a tenant score, but I would be very surprised if there weren't.  Most vendors I deal with professionally have a score to sell in additional to the data itself.

The worry that I expressed to begin with is that having these vendors is tying together all the supposedly independent market actors.  The vendor space is monopolistic or oligopolistic, so all the supposedly independent actors can be actually acting in concert, because they're reacting to all the same scores.  The problem with scores is that even if they're directionally correct, there is always a lot of nuance missed.  In this brave new world, you don't get a chance to explain your nuanced situation, you're just a number on a report, and the same number to everyone regardless of how representative that number is.

[Admiral Yi]

Let's focus on the common thread rather than the fact that compounding the natural imperfection of analogies can eventually add up.  What's common in all the analogies is that the decision of one market actor, when communicated to other actors, can create a coordination that freezes someone out.  One landlord evicting the tenant gets communicated through background checks and prevents the tenant even from an opportunity to get another place.  One gambler refuses to deal with another and badmouths them to other gambles so that they also refuse to deal with them.  One employer decides to fire an employee and through references makes them appear to be a bad hire to other employers.

The justification for spreading the information is secondary; what matters is the resulting coordination that turns private discretion into systemic exclusion. Once that happens, the market stops being free in any meaningful sense.

Yes, the one common thread is that information is disseminated.  The bad tenant and the bad driver and the welcher might prefer that information not be disseminated, but everyone on the other side of the deal is happy to have it.  There's nothing unfair or unjust about the legal acquisition of infomation and acting on it.

The point I agreed with is the moral hazard; we feel bad when a person doesn't have a place to live even if it's a result of their own bad acts.  The point I disagreed with is imposing the costs of fixing the moral hazard only on the landlords.  They should feel exactly the same sympathy as the rest of us and therefore pay exactly the same.

[Zoupa]

A review of evictions in DC concluded that in 93% of cases, the eviction was for non-payment of rent (full/partial).

Here is the increase in rent from 2009-2019. Keep in mind that since 2020, rent has increased a further 30%.



There's no need for me to post the minimum wage increase over the past 15 years, as it hasn't moved.

[DGuller]

Yes, the one common thread is that information is disseminated.  The bad tenant and the bad driver and the welcher might prefer that information not be disseminated, but everyone on the other side of the deal is happy to have it.  There's nothing unfair or unjust about the legal acquisition of infomation and acting on it.

I'll go back to one of my earlier points, in that an individual may just be an unlucky residual of a statistical model.  What you call "information" may be lot less deterministic than you make it sound.  Models are directionally right on average if they're competently built, but they may still be systematically wrong on an individual level.  Car insurance company doesn't observe you every second you're behind the wheel, they estimate whether you're a "good driver" or a "bad driver" based on some criteria, which are valid but incomplete.

In the past the "models" were the human judgment.  Human judgment is faulty, usually much more faulty than a statistical model, but they're faulty in a diversified way.  One human would get you wrong one way, another would get you wrong in a different way, a third one would get you exactly right.  Somewhere along the way someone will lease you an apartment, or give you a job, or write you an insurance policy.

These days everyone will estimate your risk the same, but that doesn't mean they will estimate it correctly.  They'll just be right or wrong in the same way.  The ones that are unlucky to be wrongly estimated in an adverse direction are shit out of luck, with no recourse and no second chances.

[Admiral Yi]

OK, that sounds like a different argument than the one you were making previously.

Out of curiosity, what would be an example of a bad tenant that was unfairly rated?

[DGuller]

OK, that sounds like a different argument than the one you were making previously.

It's the same damn argument I made in the very first post.   For reference:

The issue where tenants with eviction history can't find an apartment is also part of a much bigger problem not tied to homelessness.  It's one of the older AI problems I was always concerned about, where many seemingly independent actors all buy a vendor solution, and in effect become a monopolist even without intending to do so.

At individual level, it makes sense to screen out tenants with a bad "tenant score", whatever the reasons happen to be.  At a collective level, though, people who fall on the wrong side of the model, either for good reasons, or for reasons of being an unlucky residual, get frozen out of the market entirely.  Previously you may find some luck with some landlord, but if all of them now use same or similar solutions to screen out potential nightmare tenants, then these potential nightmare tenants are fucked.

In car insurance, we have insurers of last resort or risk pools, because we understand that even justifiably bad risks often need a car to functional in a US society.  That's why there are various schemes to force insurers to deal with them, even if all the good models as well as common sense tells us they're a loss.  I think it's way past time the same concept was extended to housing, or access to financial services.

Out of curiosity, what would be an example of a bad tenant that was unfairly rated?

The word "fair" is actually very loaded.  Even people in good faith can view it differently, and there are a lot of people right now hostile to AI that are using it in very bad faith.  I will avoid using this word here.

Based on the context, I think you're asking for an example of where an information about a tenant or a tenant score from a model would provide a misleading picture of the tenant's actual risk.  If that is what you're asking, a simple example would be a tenant with $100k in bank account forgetting their rent autopay card has expired, and having a landlord or their management company file an eviction notice first and resolving it later.  Even if you get it resolved, the case is now public record.  It's a clerical error, but frankly an understandable one, and not one that indicates financial unreliability.