Who knew that the one and only solution to a housing crisis (besides forcefully removing people from the area) would actually help the problem of a housing crisis.
The takeaway from this should be "Never listen to anyone who offers a fix for high housing costs if they say anything other than "Build more housing""
Around here in Oregon, there's a truly bizarre subcurrent of people who think that building more housing makes prices go up, because something mumble mumble builders blah blah luxury.
I suspect these people are looking for any rationalization for their emotional desire for Portland to turn back the clock 10-20 years to being a slightly grungy and relatively inexpensive city. Like back in 2009 when I paid <$500 a month for a room in a shared single family home in a slightly dangerous outer NE neighborhood. Probably the most visible change since that period is the explosion of multistory development in SE and NE.
Unfortunately for them, an urban doom loop will not be kind to all the things that made Portland livable to begin with. The budget deficits looming city-wide are grim. On the housing front, our only hope is that the statewide ban on single family zoning plus urban growth boundaries will continue to structurally encourage density via infill.
Same in Montreal, people say that prices are going up because of the luxury condos. Where 'luxury condo' == 300 sqf boxes in towers built near highways.
It’s a completely reasonable concern if new luxury housing is replacing older rundown construction. If the population growth rate is large enough, poorer people get evicted from the older housing that is then occupied by newer richer residents, and they can’t trade down to other remaining older housing, just out.
Because a 5 million dollar luxury apartment built for foreign investors and an affordable 2 bed for a nurse arent competing in the same market any more than a bike and a ferrari are.
It wouldnt be such a problem if the free market could create land, could create public transportation to link up housing to cheap land or could be kept from creating "apartment bitcoins" with a 100% land value tax but it refuses to and it wont.
So, while the margins on luxury housing are ~15% and affordable is 4-6% the free market will rationally maximize the utility of property construction firms...
It's the same market, because those investors are already there whether or not new luxury housing is built. The people buying the ultra-high-end stuff will, if it's not available, buy slightly lower high-end stuff, and so on down the chain ad infinitum, resulting in higher prices for those low-end two-bedrooms. Conversely, the very-high-end stuff being built then frees up some fraction of housing capacity all the way down the chain, lowering prices. https://www.nmhc.org/research-insight/research-notes/2024/wh...
> inner city land is scarce
Inner city land here in Portland has plenty of surface parking lots, one- and two-family homes, and old buildings that have gone unused, sometimes for literally years†.
† This part also gets into the bizarre rules around commercial building valuations and mortgages, which can make it a better choice for the building owner (at the expense of the city) to leave a building vacant indefinitely and eat the commercial mortgage payments than to rent it out cheap and have the mortgage provider suddenly call in the entire loan because the building valuation went down.
No, the people who buy the ultra high end stuff will not buy it if it is not a good investment.
Tax them more and they will decide that they dont need a huge and opulent second home in the middle of a city where all the jobs are.
>Inner city land here in Portland has plenty of surface parking lots, one- and two-family homes, and old buildings that have gone unused, sometimes for literally years.
Until we catch up on literally decades of underbuilding housing, real estate in most US cities will be a good investment. The only question is which cities are gooder investments than others and how much money the investor has.
It's not really relevant to bring in radical overhauls to our economic system when people are talking about what politically-possible changes might help improve the housing market.
> So, while the margins on luxury housing are ~15% and affordable is 4-6% the free market will do what it always does...
Volume isn't something that companies tend to ignore.
In any case, there's a chain of purchasing; if people have to move somewhere then they buy whatever is available at the top of their budget. If there's a shortage of housing then they buy whatever was originally the cheaper housing. You can see this in London where in some areas 2-bedroom terraced houses are well over £1 million.
American cities are some of the least dense in the world! You can't make more land, but you can build up. Taxation can even encourage this with land value taxes.
Change "Build more housing" to "Increase supply of available housing". Those are distinct. There is a LOT of unused housing that sits empty. If you changed the incentives a bit, that housing would not sit empty, and thus the supply of available housing increases.
Citation needed. I've never seen an-one back that up. Note that people who move need to occupy two units and scheduling the next renter to arrive on time is hard so some units must be empty while other people have two units.
Austin vacancy rate sitting at 6%. Getting below 3% is generally seen as impossible so 6 isnt that bad. Vast majority of vacancies are rehabs and people moving.
I feel like housing density is one of those subjects where I will get called a capitalist bootlicker or an Orwellian communist for suggesting that we should let the private market build more dense housing if there’s market demand for it. I don’t know how the “deregulate density” position can be simultaneously a far right and far left position, but here we are in 2025.
For many people yea, but I’ve also had many conversations where very left friends and acquaintances will argue that we shouldn’t density because:
- more dense housing makes housing more expensive (and trying to explain why this isn’t true to them is usually met with some form of “nuh uh, capitalist simp”
- if housing is built by developers and not the government then new housing will be such bad quality that you might as well die instead of have a home
- anything short of the proletariat revolution and the end of history isn’t good enough, so we should oppose density so that the market fails even more and capitalism can finally collapse.
(This is all in Seattle, I’m sure the cocktail party conversations vary around the country)
Dense housing is great until you're the first one to build in your block and then someone slaps a few 50 story buildings near you and your city hall doesn't give a damn about water usage and sewage.
Source: living in a developing country where regulations suck and am facing this awful problem.
A generalization of this solution: If you're faced with a seemingly intractable socioeconomic problem, nine point eight times out of ten the problem will become less severe with Supply Growth. What's more: the easier-to-predict second-order negative impacts of that Supply Growth also become solved because of the harder-to-predict third-order impacts.
One good example of this might end up being the American energy policy under Trump (with emphasis on the word "might"). Energy costs being high can be well-solved by increasing the supply of energy (drilling, etc). The easy-to-predict second-order negative impact of this is likely to be a harsher global warming curve. But, a harder to predict third-order impact might be increased innovation due to lowered costs of energy, enabling humanity to out-innovate global warming. Again, this is speculation, but that's how this generalization works, nine point eight times out of ten.
With housing: You increase the supply. An easy-to-predict second order impact is: housing values go down because of increased supply. But: A hard-to-predict third order impact is: Because there's more people, more businesses open shop, and that access to services actually raises your property value.
> a harder to predict third-order impact might be increased innovation due to lowered costs of energy, enabling humanity to out-innovate global warming.
This is already happening.
Texas is building insane amounts of solar capacity because there's very little environmental regulation there, which makes solar cheap to build.
Meanwhile, California is stuck with non-renewables because it's impossible to build anything new in California.
> One good example of this might end up being the American energy policy under Trump (with emphasis on the word "might"). Energy costs being high can be well-solved by increasing the supply of energy (drilling, etc).
But energy supply (by many different metrics) have been increasing for years. So either (a) energy costs are not high (b) increasing energy supply will not affect energy costs (c) it takes way more energy supply than we've seen over the last several years to affect energy costs.
Affordable housing are the old "luxury" apartments people vacate to go live in the new luxury apartments. Fighting new luxury builds is dumb, because you need them to free up the old builds for lower income.
Right. And few people involved in real estate have any interest in lowering housing prices. Not builders. Not agents. Not the government. Not your neighbor. Etc.
"Affordable housing" is the brand name for state-subsidized housing in the US that usually has higher than market unit build price with the surplus captured by various affiliated organizations.
Novel industrial policy mechanisms like this are always proposed and underperform market solutions.
My understanding is that buildings that are built to rent out are of higher technical quality than comparable owned apartments, since the owner's plan isn't to take the money and run but is on the hook for upkeep. And thus more expensive to build.
I've definitely heard the horror stories about condo builders who halfass the work in a way that will last just long enough for it to get into the hands of ownwers and the owners' HOA. My own building had a lawsuit (before my time) that resulted in the original builders paying for a replacement of the entire building's plumbing.
Programs like San Francisco's Local Operating Subsidy Program would not be necessary if the affordable properties covered had higher technical quality (which should translate to lower maintenance costs). The fact that LOSP coverage can go up to 100% of affordable housing operating cost indicates that higher spending is not improving technical quality - if quality were superior, no subsidy to operate would be required.
We will probably never know what happened unless the CCP collapses and all the documents leak out.
China had a bit of a freak out when the pandemic first started and pretty much silenced everyone close to the lab and kept everyone else out. It was until months/years later that China started allowing chosen data out, and who knows the truthfulness of it.
Authoritarian regimes, especially China, are the furthest thing from transparent.
>The marketing materials AI produces are generic, soulless, and unremarkable. The public has already grown tired of them. What will stand out? The human touch.
My guess is that it is likely from Y caps filtering common-mode noise. They are great for filtering EMI to pass strict emissions regs, so it is easy to use them liberally.
I'll admit I don't know what kind of currents they see, but 5mA on transients wouldn't be particularly surprising. There is also the factor of how sensitive the GFCI is.
Over 100nF of unbalanced Y-caps is needed at 60Hz/120V to see 5mA gnd current. That would be a big Y capacitor (2.2n or lower is common) and they’re usually used in a balanced configuration which would reduce it further. If a Y capacitor fails on one leg, you have a problem. If a Y capacitor fails is only present on one leg, you have a problem.
And if it’s resistive leakage, that’s almost a watt of dissipation - also huge and probably a short in a cable and potentially a fire risk.
5mA is a massive threshold for well designed devices
It would be fast transients that are in the kHz to MHz and over 120V. Thats why I wonder about the sensitivity of the GCFI. Can it tolerate 5mA for 50ns?
In swatcoders defence, he likely loaded the comments between the time that karpathy posted the AI comment and the time he posted the follow-up (there is a 5 minute gap). The timings on the comments seem to suggest that possibility at least.
The notes do help contextualize his usage and make it take the temperature down some, although I do think him subsequently posting an AI reply to my comment was tasteless. (But I also get it. I used harsh words there and invited some ribbing in return.)
The takeaway from this should be "Never listen to anyone who offers a fix for high housing costs if they say anything other than "Build more housing""
reply