Attributing "thinking" is also a mistake. Anthropic have shown that the "thoughts" it produces to explain what it's "thinking" are (like all the rest of its output) just plausible text, unrelated to the actual node activation happening inside the model: https://transformer-circuits.pub/2025/attribution-graphs/bio...
These tools don't have the capacity for introspection, and they are not doing anything that really resembles the thinking done by a human or an animal.
in my view both human and LLMs perform most of these aspects of thinking in a similar way … I suspect that a large fraction of the human brain performs LLM like language manipulation
for sure, today’s LLMs lack the last two on the list, and there is probably a rational debate to be had whether these can just emerge in the substrate provided by the LLM-like setting or whether the brain provides some hardwired additions to loop around focus selection and perceive-model-decide-act that will need to be grafted on to LLMs to achieve AGI
Unenforceable rules are bad, but if you tweak the rule to always require some sort of authorship statement (e.g. "I wrote this by hand" or "I wrote this with Claude"), then the honor system will mostly achieve the desired goal of calibrating code review effort.
If it's any consolation, living in Blade Runner will be optional! You'll also have the option of living in full-dive VR where it's permanently 1999. No AI in sight, just print outs of MapQuest directions.
An LLM can trivially instruct someone to take medications with adverse interactions, steer a mental health crisis toward suicide, or make a compelling case that a particular ethnic group is the cause of your society's biggest problem so they should be eliminated. Words can't kill people, but words can definitely lead to deaths.
Part of the problem is due to the marketing of LLMs as more capable and trustworthy than they really are.
And the safety testing actually makes this worse, because it leads people to trust that LLMs are less likely to give dangerous advice, when they could still do so.
Spend 15 minutes talking to a person in their 20's about how they use ChatGPT to work through issues in their personal lives and you'll see how much they already trust the "advice" and other information produced by LLMs.
Netflix needs to do a Black Mirror episode where either a sentient AI pretends that it's "dumber" than it is while secretly plotting to overthrow humanity. Either that or a LLM is hacked by deep state actors that provides similar manipulated advice.
It's not just young people. My boss (originally a programmer) agreed with me that there's lots of problems using ChatGPT for our products and programs as it gives the wrong answers too often, but tgen 30 seconds later told me that it was apparently great at giving medical advice.
...later someone higher-up decided that it's actually great at programming as well, and so now we all believe it's incredibly useful and necessary for us to be able to do our daily work
Most doctors will prescribe antibiotics for viral infections just to get you out and the next guy in, they have zero interest in sitting there to troubleshoot with you.
For this reason o3 is way better than most of the doctors I've had access to, to the point where my PCP just writes whatever I brought in because she can't follow 3/4 of it.
Yes, the answers are often wrong and incomplete, and it's up to you to guide the model to sort it out, but it's just like vibe coding: if you put in the steering effort, you can get a decent output.
Would it be better if you could hire an actual professional to do it? Of course. But most of us are priced out of that level of care.
Family in my case. There are two reasons they do this. A lot of people like medicine - they think it justifies the cost of the visit, and there's a real placebo effect (which is not an oxymoron as many might think).
The second is that many viral infections can, in rare scenarios, lead to bacterial infections. For instance a random flu can leave one more susceptible to developing pneumonia. Throwing antibiotics at everything is a defensive measure to help ward of malpractice lawsuits. Even if frivolous, it's something no doctor wants to deal with, but some absurd number - something like 1 in 15 per year, will.
I can co-sign this being bi-coastal. in the US not once have I or my 12-year old kid been prescribed antibiotics. on three ocassions in europe I had to take my kid to the doctor and each time antibiotics were prescribed (never consumed)
No. They are sort of good at reading a list of symptoms and having a decent chance of coming up with whatever their training set thinks is a likely diagnosis. They seem to be rather erratic, though, and they also make stuff up that would be trivially rejected by anyone with actual understanding of medicine (and by the same LLM if you quiz it harder!). They also seem to suffer from what I believe is known as the “reverse problem”: if the training set contains “condition A causes symptoms B and C” but not “if you have symptoms B and C then condition A should be considered”, then LLMs seem very likely to miss the diagnosis. Interestingly, real doctors seem to have the same problem, and I suspect that medicine is full of supposedly rare conditions that are actually not so rare but that most doctors are incapable of diagnosing.
This is analogous to saying a computer can be used to do bad things if it is loaded with the right software. Coincidentally, people do load computers with the right software to do bad things, yet people are overwhelmingly opposed to measures that would stifle such things.
If you hook up a chat bot to a chat interface, or add tool use, it is probable that it will eventually output something that it should not and that output will cause a problem. Preventing that is an unsolved problem, just as preventing people from abusing computers is an unsolved problem.
(1) Execute yes (with or without arguments, whatever you desire).
(2) Let the program run as long as you desire.
(3) When you stop desiring the program to spit out your argument,
(4) Stop the program.
Between (3) and (4) some time must pass. During this time the program is behaving in an undesired way. Ergo, yes is not a counter example of the GP's claim.
I upvoted your reply for its clever (ab)use of ambiguity to say otherwise to a fairly open and shut case.
That said, I suspect the other person was actually agreeing with me, and tried to state that software incorporating LLMs would eventually malfunction by stating that this is true for all software. The yes program was an obvious counter example. It is almost certain that all LLMs will eventually generate some output that is undesired given that it is determining the next token to output based on probabilities. I say almost only because I do not know how to prove the conjecture. There is also some ambiguity in what is a LLM, as the first L means large and nobody has made a precise definition of what is large. If you look at literature from several years ago, you will find people saying 100 million parameters is large, while some people these days will refuse to use the term LLM to describe a model of that size.
Radical rebuttal of this idea: if you hire an assassin then you are responsible too (even more so, actually), even if you only told them stuff over the phone.
Table saws sold all over the world are inspected and certified by trusted third parties to ensure they operate safely. They are illegal to sell without the approval seal.
Moreover, table saws sold in the United States & EU (at least) have at least 3 safety features (riving knife, blade guard, antikickback device) designed to prevent personal injury while operating the machine. They are illegal to sell without these features.
Then of course there are additional devices like sawstop, but it is not mandatory yet as far as I'm aware. Should be in a few years though.
LLMs have none of those board labels or safety features, so I'm not sure what your point was exactly?
They are somewhat self regulated, as they can cause permament damage to the company that releases them, and they are meant for general consumers without any training, unlike table saws that are meant for trained people.
An example is the first Microsoft bot that started to go extreme rightwing when people realized how to make it go that direction. Grok had a similar issue recently.
Google had racial issues with its image generation (and earlier with image detection). Again something that people don't forget.
Also an OpenAI 4o release was encouraging stupid things to people when they asked stupid questions and they just had to roll it back recently.
Of course I'm not saying that that's the real reason (somehow they never say that the problem is with performance for not releasing stuff), but safety matters with consumer products.
An LLM is gonna convince you to treat your wound with quack medics instead of seeing a doctor, which will eventually result the limb being chopped to save you from gangrene.
You can perfectly use an LLM to attack someone. Your sentence is very weird as it comes off as a denial of things that have been happening for months and are ramping up. Examples abound: generate scam letters, find security flaws in a codebase, extract personal information from publicly-available-yet-not-previously-known locations, generate attack software customized for particular targets, generate untraceable hit offers and then post them on anonymized Internet services on your behalf, etc. etc.
> An LLM can trivially make a compelling case that a particular ethnic group is the cause of your society's biggest problem so they should be eliminate
This is an extraordinary claim.
I trust that the vast majority of people are good and would ignore such garbage.
Even assuming that an LLM can trivially build a compelling case to convince someone who is not already murderous to go on a killing spree to kill a large group of people, one killer has limited impact radius.
For contrast, many books and religious texts, have vastly more influence and convincing power over huge groups of people. And they have demonstrably caused widespread death or other harm. And yet we don’t censor or ban them.
This drug comes with warnings: “Taking acetaminophen and drinking alcohol in large amounts can be risky. Large amounts of either of these substances can cause liver damage. Acetaminophen can also interact with warfarin, carbamazepine (Tegretol), and cholestyramine. It can also interact with antibiotics like isoniazid and rifampin.”
The problem is “safety” prevents users from using LLMs to meet their requirements.
We typically don’t critique the requirements of users, at least not in functionality.
The marketing angle is that this measure is needed because LLMs are “so powerful it would be unethical not to!”
AI marketers are continually emphasizing how powerful their software is. “Safety”
reinforces this.
“Safety” also brings up many of the debates “mis/disinformation” brings up. Misinformation concerns consistently overestimate the power of social media.
I’d feel much better if “safety” focused on preventing unexpected behavior, rather than evaluating the motives of users.
At the end of the day an LM is just a machine that talks. It might say silly things, bad things, nonsensical things, or even crazy insane things. But end the end of the day it just talks. Words don't kill.
Was there? It seems like that was the perfect natural experiment then. So what was the outcome? Was there a sudden rash of holocausts the year that publishing started again?
2016 was also first Trump, Brexit, and roughly when the AfD (who are metaphorically wading ankle deep in the waters of legal trouble of this topic) made the transition from "joke party" to "political threat".
Major book publishers have sensitivity readers that evaluate whether or not a book can be "safely" published nowadays. And even historically there have always been at least a few things publishers would refuse to print.
GP said major publishers. There's nothing stopping you from printing out your book and spiral binding it by hand, if that's what it takes to get your ideas into the world. Companies having standards for what they publish isn't censorship.
does your CPU, your OS, your web browser come with ~~built-in censorship~~ safety filters too?
AI 'safety' is one of the most neurotic twitter-era nanny bullshit things in existence, blatantly obviously invented to regulate small competitors out of existence.
It isn’t. This is dismissive without first thinking through the difference of application.
AI safety is about proactive safety. Such an example: if an AI model could be used to screen hiring applications, making sure it doesn’t have any weighted racial biases.
The difference here is that it’s not reactive. Reading a book with a racial bias would be the inverse; where you would be reacting to that information.
That’s the basis of proper AI safety in a nutshell
As someone who has reviewed people’s résumés that they submitted with job applications in the past, I find it difficult to imagine this. The résumés that I saw had no racial information. I suppose the names might have some correlation to such information, but anyone feeding these things into a LLM for evaluation would likely censor the name to avoid bias. I do not see an opportunity for proactive safety in the LLM design here. It is not even clear that they even are evaluating whether there is bias in such a scenario when someone did not properly sanitize inputs.
Luckily, this is something that can be studied and has been. Sticking a stereotypically Black name on a resume on average substantially decreases the likelihood that the applicant will get past a resume screen, compared to the same resume with a generic or stereotypically White name:
That is a terrible study. The stereotypically black names are not just stereotypically black, they are stereotypical for the underclass of trashy people. You would also see much higher rejection rates if you slapped stereotypical white underclass names like "Bubba" or "Cleetus" on resumes. As is almost always the case, this claim of racism in America is really classism and has little to do with race.
"Names from N.C. speeding tickets were selected from the most common names where at least 90% of individuals are reported to belong to the relevant race and gender group."
If you're deploying LLM-based decision making that affects lives, you should be the one held responsible for the results. If you don't want to do due diligence on automation, you can screen manually instead.
okay. and? there are no AI 'safety' laws in the US.
without OpenAI, Anthropic and Google's fearmongering, AI 'safety' would exist only in the delusional minds of people who take sci-fi way too seriously.
for fuck's sake, how more obvious could they be? sama himself went on a world tour begging for laws and regulations, only to purge safetyists a year later. if you believe that he and the rest of his ilk are motivated by anything other than profit, smh tbh fam.
it's all deceit and delusion. China will crush them all, inshallah.
Most poor people rent. They are already forced to move when the land they live on gets more valuable and they can't afford the new lease. Taxing land redirects part of the rent they already pay to serve their interests (assuming good government).
Land value taxes are politically impossible in America, so there's no need to make the people who like LVT look bad. We already look bad!
> Taxing land will be passed through 100% to the renter
Replies like this genuinely confuse me. How do you think rent prices are set now? I’ll tell you, they’re generally set to maintain a given occupancy rate, which is to say, they’re set as high as the market allows. The market being a group of renters which make an income, of which landlords generally take a third or more.
If we implemented a LVT tomorrow, the renters don’t get additional capacity to pay rent as mana from heaven. The rents wouldn’t budge.
Any claim otherwise requires it be the case that there’s capacity to raise rents that landlords aren’t currently utilizing, i.e. that landlords are undercharging renters en masse. I have never seen evidence in support of such a claim.
> How do you think rent prices are set now? I’ll tell you, they’re generally set to maintain a given occupancy rate, which is to say, they’re set as high as the market allows.
"What the market allows" is affected by taxes.
Suppose the risk-adjusted return to investing in real estate in some area is equivalent to the overall market, and rents are currently $1000/mo. Then if new people move into the area, rents start to go up from higher demand, but that brings rents above the overall market rate of return, which makes it profitable for developers to invest money in building new buildings, which increases supply to match the increase in demand and acts to mitigate the increase in rents.
Now suppose you raise taxes on property ownership, so that to match the overall market rate of return, rents would have to be $1500/mo in order to cover all the existing costs plus the new taxes. Then nobody is building anything new until rents hit $1500/mo, because new construction isn't profitable until then. So when new people move in, there is an increase in demand but no corresponding increase in supply, and that persists until rents get above $1500/mo. By which point the renters are paying the new taxes.
Notice that this is effectively the same thing that happens when you impose restrictive zoning rules like single family zoning. They effectively act as a tax because now you have to buy a lot more land for each housing unit you want to add, so the cost of construction goes up and rents with it.
This actually gets at one of the main criticisms land value tax (LVT) proponents have of traditional property taxes, which tax both land and buildings.
Let’s unpack your example with some rough numbers to see where the logic leads.
You assume rents are currently $1,000/month and that a new tax causes them to rise to $1,500/month to maintain the same return. That implies an added tax burden of $500/month per lot. If the lot is 10,000 sq-ft (common for single-family homes where I live), that’s about $0.05/sq-ft/month, or $0.60/sq-ft/year in land tax.
Now let’s look at whether this disincentivizes development.
Say a developer replaces that single-family home with a fourplex:
* Each unit is ~1,000 sq-ft
- Construction cost is ~$270/sq-ft, so total is ~$1.08M
* Required rent for a 5% return = $1,125/month per unit
- The land tax ($500/month) is now split over 4 units = $125/month per unit
* Total required rent = $1,250/month per unit, or $1.25/sq-ft/month
Compare that to the single-family home:
Rent = $1,500/month, also $1.25/sq-ft/month
So the developer earns the same return per square foot, and houses more people on the same land. Renters gain a cheaper overall option at the same cost per square foot. New development isn't disincentivized — it's neutral or even encouraged under a pure LVT.
Where your argument usually does apply is with regular property taxes, because those are assessed on both the land and the structure:
Single-family home is taxed on $240K (structure) + land
Fourplex is taxed on $1.08M (structure) + land
Per unit, that’s $270K in taxable improvements vs. $240K
Unless the land is very expensive, higher-density development pays more tax per unit, even though it uses the land more efficiently. And that penalty grows with scale (e.g. high-rises).
That's the core issue LVT proponents focus on: property taxes tend to penalize building, while land value tax does not. In fact, LVT often makes better use of land more attractive by decoupling the tax burden from how much you invest in construction.
> You assume rents are currently $1,000/month and that a new tax causes them to rise to $1,500/month to maintain the same return. That implies an added tax burden of $500/month per lot.
That implies an added tax burden of $500/month per unit. It's only $500/month per lot if the lot is only expected to have one unit.
> New development isn't disincentivized — it's neutral or even encouraged under a pure LVT.
You're getting this result by comparing it to existing property tax, which even more heavily disincentivizes new development.
Consider what happens if you only compare LVT to itself, i.e. to see what happens if you raise the amount of LVT by $500/lot.
If you build the fourplex, the amount of LVT increases by $500/lot or $125/unit. Now incentive to build the fourplex is lost unless the tenants are going to eat another $125/mo in rent.
And wait a minute here, under the existing property tax system the government was getting $500/mo/unit in property tax. If the typical plot is going to have a fourplex on it and the tax is expected to raise the same amount of revenue as the old property tax then it needs to be $2000/lot and thereby likewise increase the required rents by $500/mo/unit.
Where the incentive changes is that now you'll want to build not just fourplexes but highrises to try to dilute the LVT over more units. But now you've got a weird incentive: If existing rents are high enough then you want to build a highrise. But as soon as they drop below the point that building a highrise is profitable -- and highrises have a higher construction cost per unit -- you certainly don't want to build anything smaller than that, because the higher LVT per unit for smaller structures outweighs the lower construction cost per unit for smaller structures. So as soon as prevailing rents can't justify any more highrises, you get no more construction at all. Even if there are still a bunch of abandoned lots.
> That implies an added tax burden of $500/month per unit. It's only $500/month per lot if the lot is only expected to have one unit.
In the starting scenario the housing supply are single family homes, so there's one unit per lot to begin with. That it happens to be $500/unit is a coincidence of math rather than a fixed statement. Governments don't levy taxes (particularly an LVT) on a per unit basis, it's only ever on a per lot basis.
> Consider what happens if you only compare LVT to itself, i.e. to see what happens if you raise the amount of LVT by $500/lot.
The point of an LVT isn't to raise it by itself, it's to trail changes in land value.
Even if you were to arbitrarily raise it by $500/lot, the price is still more affordable for the fourplex renters than the house renters. So you should still expect less negative profit from the fourplex than the single family home. That the profit is negative will disincentivize future development, but it would also includes maintenance on the single family homes. And it'd be least bad for the fourplexes, because it's more spread out.
> And wait a minute here, under the existing property tax system the government was getting $500/mo/unit in property tax.
They were getting $500/lot, that happened to coincide with per unit but is non essential. And can't be, an LVT can only ever be charged at lot level. It's a _land_ value tax, not a unit value tax.
> If the typical plot is going to have a fourplex on it and the tax is expected to raise the same amount of revenue as the old property tax then it needs to be $2000/lot and thereby likewise increase the required rents by $500/mo/unit.
Huh? The government is overseeing a certain number of lots. Total revenue given a certain tax per lot is independent of the number of units on the lot. Also the costs, for local government, are largely per lot, in the sense that the maintenance on infrastructure dominated by lot level servicing rather than incremental uses of it. The exception to this would be schools. There's no reason to think that the tax needs to increase to be equal per uni.
> Where the incentive changes is that now you'll want to build not just fourplexes but highrises to try to dilute the LVT over more units.
This implies that the tax is not set per unit, which contradicts your previous paragraph. I happen to think this is correct, incidentally.
> So as soon as prevailing rents can't justify any more highrises, you get no more construction at all. Even if there are still a bunch of abandoned lots.
Construction is generally less discrete than you're suggesting. There's generally various buildings of various degrees of age (and various levels of depreciation) in a city at a given time. Also, most cities have a spread of land values, with land being most highly valued (and thus incentivized to be most dense) in the core and less so as you spread out.
And LVT would, rather than the discrete steps you suggest, probably look like: as it increases there's an incentive to sprawl, and as long as that population is still bringing traffic and receipts into the urban core the urban core will be incentivized to redevelop to higher density. That will bring people in from the previous sprawl, until there's additional demand at which point the sprawl either increases in density or moves further out, etc.
> So when new people move in, there is an increase in demand
but the presumption that there will be a stream of people moving in is false. The people would just not move to a place with high rent (all else being quite equal).
So high rents will lower the incoming demand. Aka, the current rent is almost always the max the market will bear, regardless of people moving in (or out) - so you can assume it's at some sort of equilibrium.
A new tax on the land will cause the owner to obsorb it, and under that equilibrium assumption above, they will not be able to raise the rent to compensate. Unless, the new tax has an equal and opposite tax credit to renters! Who now have more money, and would be able to accept a higher rent as they've become slightly richer.
> but the presumption that there will be a stream of people moving in is false. The people would just not move to a place with high rent (all else being quite equal).
People move into places where there are jobs. All else is not equal, because other places don't have those jobs. Notice that higher rents correlate with population growth. Under your theory, if rents went up, shouldn't the local population decline rather than increase?
Also notice that the overall population increases over time as a result of births and immigration, and that existing structures are occasionally damaged in various ways and have to be renovated or rebuilt. The result is that the steady-state is rents going up unless you have active investment in new housing.
> So high rents will lower the incoming demand. Aka, the current rent is almost always the max the market will bear, regardless of people moving in (or out) - so you can assume it's at some sort of equilibrium.
"What the market will bear" is supply and demand. If you add costs to supplying more housing, supply won't increase until the price matches the new costs. It changes the intersection point with the same demand curve.
> Unless, the new tax has an equal and opposite tax credit to renters! Who now have more money, and would be able to accept a higher rent as they've become slightly richer.
Your assumption is that renters can't pay higher rents. It's that they won't pay higher rents if lower rents are available, which is caused by competition between landlords and increases in supply from new construction. Take that away and people have to outbid each other for the now more constrained number of units, and then people have to pay higher rents and have it come out of their other spending, or take on more debt. Because you pay for housing or you're homeless.
Let's assume what you say is true. Then the reason it's impossible in the US is because landlords would do everything in their power to stop it, because they'd be the ones eating the cost.
I don’t think so, actually. Real estate is a pretty small sector of the economy (maybe 13 or 14% of GDP according to Google. That’s about as much as manufacturing, but not politically unassailable.
The real reason these sorts of reforms will never kick in is that roughly 2/3 of Americans are in owner occupied housing, it’s the largest asset on most of their balance sheets, and an LVT will in many cases effectively zero that out.
So it’s the homeowners (particularly the older cohorts) which will vote against this policy to the detriment of the younger cohort.
Which amounts to: “We propose to fix this problem by seizing the value embodied in the homes of 2/3 of Americans, while leaving their mortgages intact.”
That being not politically tenable is a feature of democracy, not a bug.
It is kind of a bug, because otherwise how do you fix it?
Housing got set up as a pyramid scheme where the price keeps going up because supply is constrained, so younger people have to take out ever-bigger mortgages in order to have somewhere to live. Nobody wants the prices to be unaffordable until they buy a house, but once they do they're locked in to the high price they bought in at and want prices to go up even more. When you have a majority of people on the hellbound side of the line then you're locked into a death spiral that violently explodes once the prices get so high that the majority can no longer afford a home.
fixing it implies it is a problem in the first place. It isn't a problem.
There is no right to ownership of a property (despite it being called the american dream). As long as rental vacancy is sufficient to not cause mass homelessness, it isnt a problem. And if does become a problem, the issue becomes one of production of supply and density (related to allowed zoning), and not the price and taxation (of land).
And it is through the fight to afford to buy property that productivity and innovation can be forged.
> fixing it implies it is a problem in the first place. It isn't a problem.
Isn't it?
80% of people own a home, so they vote for policies that increase housing costs. Then 70% of people can afford to own a home, so they vote for policies that increase housing costs.
Eventually only 45% of people can afford a home, so the now-majority of renters vote for different policies. The entire edifice built on ever-increasing home values collapses because policies to support their continued increase are no longer favored by the majority. Without the expectation of continued price increases, speculators divest. There is a massive housing crash and all the people who did own a home end up underwater on their mortgages, but they're no longer in the majority so the now-majority only cheers because housing costs have finally come down and they're no longer paying two thirds of their income in rent. The now-underwater homeowners hand the keys back to the bank, because why pay a $500,000 mortgage when you can now get a house for $250,000?
That's a pending disaster. Maybe we should consider policies that get housing prices down in a controlled way, instead of causing a giant recession?
> There is no right to ownership of a property (despite it being called the american dream). As long as rental vacancy is sufficient to not cause mass homelessness, it isnt a problem.
"As long as there isn't mass homelessness, having landlords milk workers for everything they have is fine"? You get "can't afford medicine" and "can't afford nutritious food" long before you get homelessness. You get "can't afford to start a business" long before even those things. That's bad. That's a problem.
> And if does become a problem, the issue becomes one of production of supply and density (related to allowed zoning), and not the price and taxation (of land).
Those things are related. If taxation is higher then investment in new construction goes down.
Land value tax also does something extra weird. Suppose the annual tax on a plot of land is a million dollars. Meanwhile the amortized construction cost of putting one house on it is $30k annually, a 10 unit complex is $400k and a 25 unit complex is $1.5M annually, because taller buildings cost more per unit. Then the annual cost per unit with the LVT is $1.03M for the house, $140k for the 10 unit complex and $100k for the 25 unit complex. Even though the 25 unit complex costs twice as much per unit than the house to build, and 50% more per unit than the 10 unit complex, and even though the local population density may only require 5 housing units on each plot of land.
So the result is that you get a lot of 25+ unit condos/apartments, and then a lot of abandoned plots with nothing on them at all, because it's not cost effective to build on them unless you're going to build at least 25 units, but it's also not worth the construction cost of 25 units unless rents stay above $100k/year/unit. So you get enough towers to get rents down to the cost of the tower plus the LVT on the land under it and then construction stops because it both costs too much to build more towers (higher construction cost per unit) and costs too much to build smaller structures on the remaining land (LVT kills you with fewer units per plot).
> And it is through the fight to afford to buy property that productivity and innovation can be forged.
High rents cause poverty and crime, not productivity and innovation.
i'm up-chain from who you replied to. Thanks for typing that. You spent a lot more time contemplating this than I did; my comment was off-the-cuff. i find people tend to be blind to certain, "externalities". I am people, too, so sometimes i prod to get discussion going and learn something.
There are many other forms of government which are less subject to the influence of the people and their silly preferences.
If you don’t want to switch forms of government, formulating a holistic plan, addressing weaknesses in the plan, laying out how the transition will work, educating the people about it, and convincing them why your plan is better is what I’d recommend.
The general problem is the "democracy is two wolves and a sheep voting on what's for dinner" thing. If you can get at least 51% of people to vote for "I've got mine" then you've created a system of injustice. Which is nominally why we use a republic rather than a direct democracy.
But those checks and balances have been eroded over time, or weren't strong enough to begin with, so then the majority e.g. votes to constrain new construction because they value their own short-term financial gain over the greater harm they're causing to the minority.
And then the same dynamic plays out for some other issue, but this time you're in the minority and you take -50 damage instead of a +10 gain. So the problem is that it's in everyone's interest to stop this from happening in general, but only in the minority's interest to do it on any particular issue, and democracy in and of itself isn't structured to handle that.
Wealth that isn’t concentrating on the 1% will be diluted as part of inheritance. Those from a poor background will continue to be screwed no matter how well they are doing. Tabitha who makes clogs for shitzus as a job (standard person on uk real Estate shows) will co time to live rent free massive houses while those working and living in the cities massive rent and commuting will do tie to pay far more tax.
That’ll be true when my generation is old and when yours is old as well.
My house is well sized for our family of four. I don’t plan to move out when the kids go to college nor as long as I’m able to live independently. People who don’t like that are entitled to their opinion but not to my acting in accordance with their wishes.
Thanks, I understand now. That makes sense. It may be though that as the older generation shrinks, the newer generation enacts new zoning regulations. That could change the composition and mix in current residential areas now dominated by large single family houses.
Huge red flag, this isnt the case. It isnt a perfectly liquid market.
>If we implemented a LVT tomorrow, the renters don’t get additional capacity to pay rent as mana from heaven.
No of course not, the process would probably take 12 months, as home owners identify their losses, and instruct rental agencies to end leases and increase prices. Accepting that while the house might be on the market for longer than usual, attracting a higher income tenant and getting them to sign for term is going to long term resolve the new hole in their budget.
>that landlords are undercharging renters en masse. I have never seen evidence in support of such a claim.
I am living this discussion right now. So this is amusing me.
I rented my current property in 2020 for 320 p/w. It was probably worth 400 p/w I had a handshake agreement with the rental agency that they would leave the price where it is, and I would complete as much maintenance myself before raising a fuss.
In 2022, the owners had to sell the property to pay for medical treatment. The old owners had no mortgage, and loved having a fixed stable income. Their biggest risk was having the property untenanted for any length of time. So they left the price alone while the world moved on.
The new owner has raised the rent every 12 months since. The previous owners sold at a massive gain, and the new owner was struggling to pay the mortgage. My calculation based on sale price is that their mortgage payments were in excess of 380 p/w and, considering the sale grandfathered my no maintenance agreement, are looking at considerable maintenance costs.
Current cost to me is ~500 per week.
Issues: Market is currently roughly (we have big discussions about this) permitting 520 - 540 pw for similar properties.
My barrier is that its a substantial cost to move. So I negotiate all the price rises with the real estate in terms of property improvements.
If they saw a taxation increase imposed by an LVT, they would not hesitate to pass it on to me over 12 months. I would have to reevaluate the costs to move. If the same increase occurred roughly evenly across where I want to live, I would probably just have to make it work. I cant go anywhere to avoid the LVT increase, and it costs money to relocate. I would have to further erode my spending in other areas to account for it. And I could.
This really comes back to what the LVT is designed to do. Its designed to enforce the most efficient use of the land. The outcome is intrinsic to the goal. The land is not being used most efficiently (I rue the day that my landlord figures out that we have the exact amount of land required to battleaxe the block, turning my backyard into another property) which for humans, as opposed to spherical cows, is actually fantastic in a lot of cases. Under an LVT my landlord would be required to run the property in the most taxation efficient manner, which is counter to my interests as a renter.
But yeah, the rental market is not liquid. There's costs to move. Costs for having the property on the market untenanted for significant time. Costs to bring a property up to spec between renters etc etc. And lots of people live in that gap. To the point where our local real estate industry body actually issued guidance to real estates that they should raise the rent as frequently as possible because it wasn't getting done very often, suppressing prices.
> No of course not, the process would probably take 12 months, as home owners identify their losses, and instruct rental agencies to end leases and increase prices.
Why don't home owners just instruct rental agencies right now to end leases and increase prices? Why do they need to wait for some hypothetical tax to kick in?
Because their tenant can just go down the street and rent from another landlord whose costs did not increase by $10-20K/yr.
In a world where LVTs increased taxes on all nearby landlords, that threat is an empty one. Sure, they can threaten to go move 10 miles away or eventually to go move into a beehive apartment that will be built 3 years from now to optimally house humans while not leaving a shred of land under-used.
> Because their tenant can just go down the street and rent from another landlord whose costs did not increase by $10-20K/yr.
Well, that only works iff landlords can pass down these taxes. That's exactly what we are discussing. So this is begging the question.
> In a world where LVTs increased taxes on all nearby landlords, that threat is an empty one. [...]
It doesn't matter. Even if only left-handed landlords paid LVT, rents wouldn't change depending on the handedness of your landlord. (However, left-handed landlords would likely sell to righthanded landlords, because the property is worth more to them.)
It is not begging the question to identify that a key economic factor would change under an LVT to be different from today. This change would affect the number of landlords who could profitably offer units for rent at today’s prices, which I posit would affect their willingness to supply units at that price, shifting their supply curve.
Your own observation that rents would not change based on handedness of your landlord implies that you understand this.
(You can change the supply of land on the market, by eg giving a tax subsidy for keeping land off the market. That's what the British council tax does in many counties: you get a rebate, if your property ain't occupied.
But LVT specifically does not depend on what you do with the property. It's due one way or another, thus it has no influence on the supply of land.)
The supply of land is not fixed, its highly regulated.
In my country, the state governments have land release targets roughly matching population growth, they have never met these targets.
Then there's zoning. We have wild agricultural zoning practices that prevent agricultural land, even shitty agricultural land, from becoming residential.
The land that is 'released' also already exists. It's just in government hands.
> Then there's zoning. We have wild agricultural zoning practices that prevent agricultural land, even shitty agricultural land, from becoming residential.
Yes, there are land restrictions. But that limits land use. Not the total amount of land.
In 2025, a certain supply of units is available for rent in some relevant market. In 2026 an LVT is enacted shifting taxes onto land owners (and away from income earners).
Do you expect rents at the end of 2026 to be higher or lower than now in that market?
In your scenario, we are enacting an LVT and lowering some other tax (say, income taxes).
The LVT itself will have no impact on rents. However, lowering income taxes will increase rents, independent of LVT. (Eg if they'd just lowered income taxes, because some rich sheikh wrote the government a big check for fun.)
That's assuming all else being equal, eg number of units available for rent.
Any tax levied on landlords will impact rents. Any business where the input costs increase, has an effect on prices. If you have a heck of a lot of competition AND the businesses can eat the increase you might get away with it. But LVT is a regressive tax that doesnt care about income, so the "businesses" aka landlords in this scenario can (and probably will) have tax that reduces their income to a deficit. Theres no getting around this, you need to stop pretending that LVT is magic.
Why do want to do everything in your power to ignore it? Stop trying to use your ignorance to selectively ignore arguments.
Regressive taxes hurt people which is why progressive taxation exists.
>So you are also arguing conversely that giving landowners some free gifts will lead to a drop in rents?
No, to drop rents you need both an increase in supply and a reduced costs for suppliers.
>Prices are set by demand and supply. For a cost to be passed on, supply must shrink.
This is a bonkers statement. "For a cost to be passed on supply must shrink". Only in a perfectly elastic supply/demand market is this true, and only if I sneak "relative to demand" in there. We have been around and around on this point. But its been demonstrated false over and over again.
Whats worse, is the other outcome identified, is supply shrinking. So not only are you wrong about prices, your statement is already inclusive of the effects of LVT.
>There's nothing in the LVT that would cause supply to shrink.
Ok you have to be a religious zealot to still be spouting this.
If the cost of supply houses to the rental market becomes higher than the income expected from supplying houses to the rental market, supply to the rental market will shrink.
LVT is a regressive tax, in that it doesn't care about income. So landlords with tight margins are going to be just as affected as landords with good margins. Any landlord making 1 dollar more than rental income, faced with a new 100 dollar LVT, is going to either withdraw their property or raise rent at least as far as 99 dollars to make it worthwhile to hold the asset.
Simply repeating your religious mantra does not change this obvious fact.
Fixed term rental contracts in many jurisdictions, and lack of motivation, particularly for landlords buying property primarily for anticipated capital gains who are quite happy with a "reliable" tenant paying no much more than their costs, but won't be prepared to do the same thing when the monthly cost goes up (and most of the anticipated capital gain disappears...)
If you go from covering your costs and expecting to making a capital gain to not covering your costs and not expecting to make a capital gain, you're much more likely to put the rent up
Approximately all landlords are already charging the maximum they can. Especially commercial landlords. Otherwise, they would break the fiduciary duties to their shareholders.
>Approximately all landlords are already charging the maximum they can. Especially commercial landlords. Otherwise, they would break the fiduciary duties to their shareholders.
Assuming all landlords have shareholders is the first clue you aren't on target here.
You got lucky with the old owner. They gave you a sweetheart deal.
I hope your new landlord will lower your rent, when they refinance to a cheaper mortgage. (Or are only higher mortgage rates pass through? What's the difference?)
>You got lucky with the old owner. They gave you a sweetheart deal.
I shopped around and found a good deal. It including moving to a suburb with cheaper rents and still making sure I found something good.
But this is the market. Its not full of spherical cows, its full of human beings making human decisions.
I found humans who valued longevity over pure profit. This is not a unique scenario, people use rents as retirement income all the time. They wanted a consistent tenant, and the suburb I wanted to move to is famous for inconsistent tenants.
I valued low prices and outdoor space. We found a price we both agreed on. I know of other examples of this happening near where I live, it isnt unique.
Likewise, costs being passed directly to the renter, not unique.
Your question that I answered wasn't about LVT, so shoehorning LVT back in here seems irrelevant.
If your question is, AH BUT SOMETIMES THEY DONT INCREASE RENT ALWAYS SO MAYBE THEY WONT INCREASE IT WITH AN LVT. But we are talking about a situation where the cost to supply a rental property has increased dramatically.
Landlords will fit into 2 categories.
1. Landlords who can absorb the cost.
2. Landlords who cannot absorb the cost.
Landlords have 3 options.
1. Try and absorb the cost.
2. Pass on the cost through increased rent (possibly leaving the property untenanted)
3. Sell
Landlords who cannot absorb the cost either need to pass it on or sell. Theres no "absorb it" long term for them. Likely the vast majority will try to pass the cost on before doing so.
The question remains whether landlords who can absorb the cost, increase prices. If enough of the market bites on the rent increases they likely will.
Also, there's another leg here. If significant numbers of landlords remove their properties from market, scarcity will drive up rental prices anyway. We literally did this to ourselves years ago when we removed negative gearing. Negative gearing has a tendency to drive up rental prices, but it also drives up supply of rental properties. We removed it for a year or two and the result was a dramatic increase in rental prices due to lack of supply. Renters cannot become home owners fast enough to ensure elastic demand, that's why house ownership : rentals is often seen as arbitrage. You take a product from one market and make it available to consumers in another market.
> If significant numbers of landlords remove their properties from market, scarcity will drive up rental prices anyway.
unless you're imagining they're destroying the property, you cannot "remove" it from the market (or they decide to leave it vacant deliberately - an illogical a choice as destroying the property).
If they sell to a owner occupier, it removes that buyer's original rental demand, by the same amount as the removal of the rental property from the market.
If they live in it themselves, it's the same (just no money changing hands).
Its not that the house is destroyed, its that they aren't spherical cows that immediately exist on the sale market. Renters cant become owners instantly. Landlords cant sell properties instantly. These things take time. Housing is not a perfectly elastic market.
This was a long reply and I’d like to honor it by addressing what you bring up, but there’s a lot so forgive me for jumping around.
Let’s start by noting that we’ve now shifted the argument from “100% of this tax will be charged to renters” to, “this will get passed through to renters who were previously being undercharged.”
It’s hard to directly argue with your anecdote because I don’t know where you are or what your rental market is like, so I’ll address the argument more broadly.
Let’s examine how common that arrangement is. To the best of my search small time landlords of the variety you mention own something on the order of 35 and 40% of rental units. The rest, primarily multifamily (apartments) are owned by corporate landlords.
The corporate landlords were sued by the last admin for price fixing using realpage. I don’t know if or how that case was resolved but I think it’s safe to conclude they’re probably not undercharging.
For the small time landlords, probably some aren’t undercharging and some are. After all, how hard is checking neighborhood rents on Zillow once a year? Even if none are, we’re still admitting at least 60% of renters aren’t getting the kind of deals you mention. So this is a minority case, and probably shouldn’t be the basis of policy.
Addressing this point:
> Under an LVT my landlord would be required to run the property in the most taxation efficient manner, which is counter to my interests as a renter.
Even granting that that’s true, I’d argue it’s not persuasive. The government needs some amount of money to function. LVT is one source of that money, income taxes are another, sales taxes are another, wealth taxes are another, etc.
So your argument fundamentally resolves down to “other members of society should make up the deficit in taxes I would otherwise pay so that I (and others in my situation) can enjoy a yard or more space than I otherwise would.”
And the basic question here is, why? In what way does it benefit literally anyone else in society that you have a bigger yard? This is a blunt and perhaps impolite way to put it, but it’s true.
Going back to those alternatives, we can counterfactually raise income taxes on some waiter bussing tables or a SWE slinging code at Facebook, on an author with a copyright, or on someone who owns a business. But, if we do, we should expect less of all of those services. That serves as a reason to avoid such a tax.
In contrast, with the yard, I can’t think of a single such service provided or reason to avoid the tax. And that’s the crux of the Georgist argument more than railroads or slumlords. It’s the empty lot, or the lawn as we now call it.
>The government needs some amount of money to function.
The issue at hand isnt "Does the government need money to function" unless we are going full MMT derp we can all accept a level of taxation is necessary.
The issue at hand is whether LVT is a better distribution of the burden of taxation.
>So your argument fundamentally resolves down to “other members of society should make up the deficit in taxes I would otherwise pay so that I (and others in my situation) can enjoy a yard or more space than I otherwise would.”
My argument is largely that by shifting the burden of tax from a progressive tax regime, that is only targeting those by an amount calculated on what they can afford, to a tax regime thats largely dictated by circumstance and subjective valuation, you will hurt a lot of people who dont deserve it.
If we go too far in this direction we will have to define basic moral principles, because that seems to be where you are heading. I dont necessarily think that earning income should mark you for punishment either. But just as not taxing the homeless is fairly universal, not taxing people who organised their retirement and arent a burden is also fairly universal. Your mileage may vary.
Not under an LVT. The amount of rent a property does bring is more a property of the buildings and improvements and only minorly of the land.
The amount of rent that it could bring if maximally used is relevant to an LVT, but not what it actually does. (This is the entire point of an LVT vs a property tax.)
Yes, the Fed controls how much money is worth, ie the price level. They 'print' more or less money so that the change in the general price level (aka 'inflation') hits their target.
They are sometimes a bit clumsy, yes. But they manage to get long running inflation to 2%.
(Especially they can always create more inflation by just buying more and more stuff with freshly printed money. If that ever fails, that means they have unlocked an 'infinite wealth' cheat: just keep printing and buying until you own the whole world. Obviously, the dollar would lose value before that, hence proving that they can always create inflation.
Lowering inflation isn't always possible, yes. But it's always possible as long as they have any assets on their balance sheet: incrementally sell those assets and 'bury' the money received in the sale.
As long as the Fed ain't out of assets, they can lower inflation.)
> they can always create more inflation by just buying more and more stuff with freshly printed money.
the GFC would like to chat about that.
The feds cannot nudge more than a tiny bit in either direction. Look how much money was printed during and post GFC, to try make inflation higher, and it never did.
The bank of japan tried the same a decade earlier too, in japan. Inflation never took hold no matter what they tried.
> As long as the Fed ain't out of assets
the Feds can never be out of assets, as they just print more. What they cannot control is the hearts and minds of all economic participants - and that is where the root of inflation (or deflation) comes from.
They didn't take their own advice, that they previously gave to the Japanese central bank in the 1990s. Instead, they started paying interest on excess reserves to neuter their own buying.
> Inflation never took hold no matter what they tried.
They tried approximately nothing. As I said, if you found a way to keep printing money without raising inflation, that's actually an opportunity, not a problem: just keep printing money and gradually buy the rest of the world for a bit of paper and ink. (Well in reality, it's only database entries. Which are even cheaper.)
If they can't create inflation, they should go and take lessons from Argentina, Turkey or Zimbabwe.
> the Feds can never be out of assets, as they just print more.
The Fed prints liabilities, not assets. Domestic money is a liability for the Fed. Assets are the stuff they bought with the money, like eg bonds or gold.
> Taxing land will be passed through 100% to the renter.
Taxing land, though, forcing more economical use of said land. It's wasteful to spend your land on single units to rent, and then you're not going to be competitive when other developers are putting in apartments around you.
So, you're going to be smart and try to score a better unit/land ratio. Maybe you'll build apartments, or dingbats, or townhomes. The end result is that, per tenant, they're going to be paying significantly less rent.
The reason we build sparse and stupid housing is because it works and it makes money for the builder. But it really, really shouldn't - at least not all the time. We have the wrong incentives.
> Taxing land will be passed through 100% to the renter.
Then I suppose it's settled. All that economic theory, and all those left-wing and right-wing economists that unanimously agree with that research, is all meaningless, because of your declarative statement.
I haven't been able to figure out how there's a moat for AI products that, if they work as advertised, can build a bridge over any most with near zero user effort.
The moat for AI products will be, as is so often the case, user data. In this case, your personal history of interactions with a given AI.
The author predicts a land grab where AI companies try to scoop up as much personal data on you as they can as fast as they can, which renders them significantly more personalized to you than other AIs. That's the moat.
Analogous to Facebook managing to scoop up your entire social graph. Other social networks popped up, but there was no incentive to use them because you didn't have your social graph setup there and it was really hard to rebuild.
You're under the mistaken impression that you will own this data. What you are describing is akin to saying you will just export your friends list, posts, messages, pictures, etc from facebook to some other social network.
Maybe you are technically able to run your own AIs (and I guess they would somehow be supported by these other platforms?). But most people are not.
It's hard to export my friends list because there's no API for automated tools to use. AI can export your friends list today because it can see your friends list in the browser. I think the same thing will happen for chatbot memory. I don't need to run my own AI to do this. Every AI company will build their own importer.
It's not that python doesn't have a clean way to subtract a year, it's that "subtract a year" is imprecise. There's a clean way to subtract 365 days, and there's a clean way to set the year one year earlier. But if you're doing the second thing, is python supposed to silently change to March 1 when you change the year from a leap day? There's no way around handling edge cases.
This has been a factor in contract law for longer than computers have existed. "Subtract a year" or "Add a year" aren't really imprecise, it's just that there are two ways you can interpret it: 365 days, or 12 months. When adding, you get the same result: end of February, the 28th. Subtracting, you get one of March 1st for 365 days and February 28th for 12 months.
Most jurisdictions use the 12 month standard, fwiw. I believe that's the ISO ruling as well, but I wasn't able to confirm that.
To be a bit blunt, these questions are entirely irrelevant, since you know the answer as well as I do: the calendar year, the one Western law and international business uses uniformly. It's very clever of you to mention that other concepts of the year exist. Kind of. Would have been more clever if you'd realized that they aren't germane, any more than the Islamic, Jewish, Indian, or Chinese years are.
What is it? Does it attempt to return a datetime that's the same time as the input but 365 calendar days earlier (ignoring leap seconds? compensating for time zone changes?), or does it subtract 31536000 seconds from the current datetime? Because those aren't necessarily the same and it's ambiguous which one you mean by "subtract 365 days"
Too many of the identity and data problems are network effect problems, and ad revenue is the best business model that has been discovered for tackling those problems. The business model is incompatible with openness and user sovereignty: the dollars go to the platforms with the most users, and the other platforms die.
We can't get the Mozilla we want until they discover a network effect business model that's actually compatible with empowered users. Until each user gets to steer every advertising dollar on the internet where they want them to be spent, there's plenty of room for innovation.
Houston is the only major Texas city without zoning, and Houston still has plenty of parking requirements. As a result, both housing for people and housing for cars is really cheap. At this rate, Houston is going to figure out how to fix its planning mistakes of the past much faster than more regulated cities are going to decide to build enough housing for the people who want to live there. I'd never bet against Houston, or Texas in general.
How about we ease up on both zoning restrictions and parking requirements?
I’d bet against Texas in the idea that their working class/average citizen population is going to be quite oppressed until women have meaningful rights there.
Hard to have a functioning society when a near literal half of your population is seen at best as second class citizens.
Houston had a female mayor, who happened to also be lesbian, for the maximum amount of terms. I get your general point about Texas, but at least Houston is probably the most diverse, multicultural, and liberal of the big Texas cities.
The parent I’m replying to is partially discussing betting against Texas, especially regarding property values - if you think half the populace within a state being viewed as second class citizens doesn’t affect property values, I think you should reconsider.
How many rich tech people from SV/Cali who paid top dollar for Texas property are now quickly fleeing after the recent political happenings that stripped the rights of half the voting population? I’ve seen quite a few personal anecdotes on HN & also a general trendline across publicly accessible real estate data that point to them ditching their Texas properties quite quickly. This most definitely affects property values, & at least touches on zoning issues in the grand scheme of things.
> How many rich tech people from SV/Cali who paid top dollar for Texas property are now quickly fleeing after the recent political happenings that stripped the rights of half the voting population?
Not very many at all, in my estimation. Those "rich tech people from SV/Cali" moved to Texas in the first place for their bottom lines, not because they thought Texas was the vanguard of progressive politics. I doubt a socially conservative state becoming more socially conservative is going to trigger many reverse moves.
If I had to guess, the extremely restrictive abortion laws (prohibited after six weeks of gestation with very few exceptions) that you're obliquely referring to are not going to stand for very long, either.
> How many rich tech people from SV/Cali who paid top dollar for Texas property are now quickly fleeing after the recent political happenings that stripped the rights of half the voting population?
The rich tech people who moved from California to Texas often were doing it specfically because of support for the political party that did that stripping (sometimes centered on tax/regulatory policy rather than culture war issues, but still supporting the party knowing the culture war stance was part of the package.)
The lazy solution to rivalry getting out of control is bicameralism. Make tree-based governance where most of the action is, but design another chamber that can veto it without the same rivalries involved.
