988 karma · joined August 12, 2019
So you lose some economy of scale when enrollment drops, you have to cut school funding, which makes the schools worse, which makes the town less desirable, which drops real estate prices, etc.
I think an 11% decrease is conservative if anything. None of our peer nations spend more than ~5% on administration as a share of overall healthcare spending.
1. On average, how healthy is your group of ACA plan holders? If the group has a bunch of chronic conditions, they get more subsidy money to offset the increased care costs. Going to the PCP allows them to have official medical evidence of those conditions.
2. The government gives these plans quality ratings to help people compare them to each other. These ratings are partially based on how often patients get their annual screenings and patient satisfaction. A gift card for a PCP visit accomplishes both aims.
There are also more practical concerns. Preventative care is cheaper than an acute incident for the company. You'd rather catch an arrhythmia at a PCP appointment than pay for the cost of a heart attack.
28% of the US is at Level 1 or below. That level roughly corresponds to reading a simple text and being able to answer a question about it that requires limited inference (i.e. if the question is "what color are pigs," the answer would be in the text as "pigs are pink in color.")
[0]: https://nces.ed.gov/surveys/piaac/2023/national_results.asp
> Adults at this level are able to understand simple problems and develop and implement solutions to solve them. Problems contain a limited number of elements and little to no irrelevant information. Solutions at this level are simple and consist of a limited number of steps. Problems are embedded in a context that includes one or two sources of information and presents a single, explicitly defined goal.
This test is administered on a tablet, so I think scores can be interpreted as a sort of combo of computer use and problem solving.
A full 40% of adults from 55-64 are at that level or lower. Wikipedia thinks that a novel online form would be a Level 2 task, provided it involves navigating across more than 1 page. Based on that framework and assuming your dad can use the sort function in his email, he represents a top 20th percentile adult for the 55-64 age group. It's probably even higher considering his age as the trend is towards older groups having a rougher time with the survey.
I also think it's ageist to assume that older people can't use a computer. But assessments like this indicate that a full quarter of the US regardless of age would have trouble with some of the basic tasks we associate with computer use. So designers should consider their intended audience when deciding what's simple or not simple enough.
[0]: https://nces.ed.gov/surveys/piaac/2023/national_results.asp
Does that make him infallible? Of course not. But it does mean I'm going to give him the benefit of the doubt here.
The small towns in my area are desirable not just because they're small, but because they're small and have the same amenities as a bigger town. These towns are also increasingly relying on debt to fund decades of deferred maintenance just to keep the amenities they already have alive.
This would become dire if (for example) property values fell, as residential real estate taxes make up the majority of tax revenue. This might happen if demand for the town fell (perhaps due to fewer people), which would quickly become a catastrophic cycle (revenues fall, roads and schools deteriorate, which makes the town less desirable, repeat ad infinitum).
Regardless, the S&P 500 also excludes a company like Microstrategy (the company that holds Bitcoin) from their index, had excluded Robinhood for a wile due to missing the profit requirements, and so on. It was never "meant" to cover the 500 largest companies by market cap, and has generally resisted pressure to change that.
Plenty of great books would make terrible movies for this reason, and plenty of pretty terrible books can actually make good movies.
1. Continuing to grow their share of the market.
2. Margins staying high.
3. Inference costs coming down.
4. A need for Anthropic's models specifically.
I buy 3. But 1, 2, and 4 rely on models continuing to improve at the same rate, such that you need the latest version to stay competitive. At the cut below frontier models, there's already robust competition between open source models, cheaper providers like Deepseek, more local AI alternatives, etc.
I think the case for the unit economics being fine starts to fall apart if you can't charge a large premium for your best in class model.
I could be wrong, but the margins are so good because there isn't a "substitute" for the frontier models. The performance difference between the latest Opus and a more open model provider is large enough to justify the extra cost. If that difference shrinks, I think the cost people are willing to pay will go way down.
There were some mistakes made but they were all pretty reasonable decisions. But the situation that caused this in order:
1. The support line for the pump won't suggest sending it to where you are because you say you have a backup plan.
2. The support line also isn't a doctor; they can't evaluate how good or bad your backup plan is, just how to deal with the pump.
3. You message your GP. But the message system doesn't connect you live to the on-call practitioner; that's a phone number that is not easily findable.
4. Your GP fills a prescription for the backup pen and sends it to the closest pharmacy.
5. The pharmacy takes the prescription, but they don't actually know if they can fill it until later. They also can't find a pharmacy farther than 25 miles away from the current location.
Each issue is reasonable by itself, but the end result is that the patient had to call 2 different people, go to a pharmacy, and still wasn't able to get what they needed. The gaps that could have (theoretically) been patched are:
1. Support could have known the possible workarounds. I realize that legally it could have been dicey, but the "fix" suggested on Reddit seems like something that could have been mentioned as an interim solution.
2. Your GP could make it more obvious which communication method reaches the office on a weekend.
3. The GP should be able to talk to the pharmacy to ensure they have the medication in the prescription.
4. The pharmacy should be able to search farther than 25 miles to find a medication for a prescription they've been sent.
It should be much more difficult for a patient to hit every edge case like this, especially when just being able to talk to each other would have stopped some of these.
Family businesses handle this by slowly getting the next generation involved, such that the intangible value has been transferred by the time the owner wants to retire. That works less well with a stranger. You really need high levels of trust to make that work fairly for both parties.
Family businesses have traditionally gotten around this by having their children involved for 10+ years prior to taking over. That way, you slowly transfer your "social capital" to the new owner (in this case your kid). This is understandably harder to manage with a stranger, because you can't transfer the value before they buy it, but they won't buy it if they can't guarantee you'll help them transfer the value.
To extend this, the eventual problem is that you eventually lose all of your new users, which means you lose the extremely valuable intermediate users (ones that know enough to ask complex questions, but not so much that they can figure out the answer).
That being said, I can see a few plausible biases (though none of them explain the scale of the increase IMO):
1. CRC risk is correlated with some previously fatal, but now curable disease. The mechanism would be that your high-risk CRC patients would die due to yellow fever or something in 1970, meaning they don't have the chance to get CRC. The important thing is that it would have to "artificially" remove high-risk patients from the age group, but not low risk patients.
2. CRC risk is noticeably higher at 24 than it is at 20, and all-other-cause mortality is significantly lower today. That would lead to a higher proportion of 24-year-old "years" in the calculation.
3. People used to die of CRC before it was caught, which caused it to not be recorded as a cancer incidence.
1 seems unlikely, and even if true shouldn't make a big difference. 2 seems the most possible, but still unlikely to make a huge difference. I don't know enough about how they determine cause of death to know if 3 is a possible outcome.
Regardless, digital file possession and ownership doesn't map cleanly to our language. I technically don't own any Kindle books I buy, I can't share them, yet I clearly have access to an ebook. So I both do and don't currently possess said book.
In my experience, the reason they're cheaper is because they offer fewer features (cut down on ongoing expenses) and because they aggressively enforce TOS (the margins are thinner, so you're less able to afford people using more resources than allocated).
Or maybe I'm wrong, and "we don't want sports gambling" is different legally from "we want to be the regulator of gambling." But it seems like they're just variations of the same question, which is: do states have the power to regulate this?
Not to strain the analogy, but it's more like not locking your bike when it's in your locked apartment (the apartment being your computer). The thought being that if someone puts the time and effort into breaking into your apartment, a bike lock isn't going to do anything to stop them.
There's also the synergy that GameStop now has access to more used gaming inventory, a category that I'm led to believe is high margin for the stores.
This sort of thing all the time when (for example) a movie lover opens a movie theater. Running a profitable movie theater is a lot less about movies and a lot more about maximizing concession revenue.
Whereas prediction markets (and some other stock market/commodity option contracts) are explicitly tied to an event. They're saying "I'll give you $1 if this clearly defined event occurs."
But its worth mentioning that there are fewer and fewer "other public spaces." My local library is just that, a library, and that means I can't:
1. Eat in it, perhaps while studying. 2. Talk above a whisper. 3. Rent anything but books that I might want/need. 4. Do anything on a computer but be on the internet (the computers run a locked down version of Windows XP)
That's not a "problem" exactly. This library is doing exactly what a library is supposed to do. But my town has one other "public" space, which is a combined community and senior center. That's not good for much outside of chair yoga for a kid in a high risk environment; it's largely designed for adults.
It's nice that my library is "just a library" because I don't need it to be anything else. But the fact is that the library is one of the few open, walk-in, free public spaces left. It being "just a library" in that case seems like a missed opportunity.
That's why I disagree with people that say "just give me whatever you gave the LLM." That's only useful if you, the writer of the prompt, have no intention of looking at the LLM output before sending it.
Zeno couldn't prove that there were no gaps; he showed that infinity was different from how we understood finite things, bit that's not the same as proving there are no gaps.
Later, mathematicians proved the existence of irrational numbers. These were "gaps" in the rational numbers, but they weren't all the "same" of that makes sense? The square root of 2 and Euler's number are both irrational, but it's not immediately clear how you'd make a set that includes all the numbers like that.
However, there are "gaps" in that number line. Between 1 and 2, there are values that aren't integers. So the integers make a number line that is infinite, but that has gaps.
Then we have something like the rational numbers. That's any number that can be expressed as a ratio of 2 integers (so 1/2, 123/620, etc.). Those ar3 different, because if you take any two rational numbers (say 1/2 and 1/3), we can always find a number in between them (in this case 5/12). So that's an improvement over the integers.
However, this still has "gaps." There is no fraction that can express the square root of 2; that number is not included in the set of rational numbers. So the rational numbers by definition have some gaps.
The problem for mathematicians was that for every infinite set of numbers they were defining, they could always find "gaps." So mathematicians, even though they had plenty of examples of infinite sets, kind of assumed that every set had these sorts of gaps. They couldn't define a set without them.
Cantor (and it seems Dedekind) were the first to be able to formally prove that there are sets without gaps.
More generally, though, the response can be as simple as "We have received this email; the request will take some time, here's roughly when you can expect an update."