Anthropic is making money from people who under-utilize their subscriptions, and presumably by sneaky throttling or not-sneaky throttling power users. Currently they are in an adoption race. Whether being first will actually let them "win" the market (and the market is a bit ill-defined) is unclear.
Completely agree. AI is here to stay, it's going to garner more and more use overtime. However, I'm skeptical that the investments being made right now are at of the right scale & at the right time. I completely agree that over time we'll rework more and more of our society around LLMs or their successors. However, like you say it's a slow process: we have to learn how to do it effectively, organizations need to change, people's attitudes and behaviors have to adapt. I just don't see is "getting there" fast enough to justify current spending levels.
I find myself rarely reaching for Opus nowadays, it's just too slow. I assume there are tricky use-cases where it's really useful though, just not super relevant for my day to day. I much prefer a faster, "weaker" model.
I also neglected to mention the cost of the house sitting there for months vacant waiting for buyer. I'd always price mine slightly under the market to get them sold.
As for accounting guidance, just what are you driving at? Are those expenses I listed not real? They certainly took real money out of my account!
I mean, it can be offset with money.
- Kids take time - Yes, so does working. If you cutout the 90h my partner and I spend working, that's a lot of time to put into raising children.
- Kids take effort - Yes, so does working. If I didn't need to work this becomes much easier.
- Opportunity cost - Yes, just pay me for the opportunity cost. Pay for my PhD after my kids are in grade school.
It's just that these policies are very expensive, and right now we allocate our money mostly to make rich people richer and maintain very high QoL for our elderly population. That's a choice we make in setting up our society.
Exiting the workforce for a decade (Replacement fertility rate is 2.1 implying some people will have 3 kids, spaced 2 years apart plus 5 years of child-rearing until kindergarten) has an opportunity cost that is potentially in the millions of dollars, depending on the industry, and the time costs of child rearing doesn't suddenly end at 5 either.
Depends on implementation. For example, a wealth tax that has a "cap" at some ludicrous amount of wealth, like $10M, would effect very few people and therefore be insignificant for the average worker. So 99% of people would continue saving with no change at all to their behavior. The externalities could be nice though, since it'd distribute capital more efficiently. Sort of a general stimulant to the economy.
This line of thinking though assumes it would have no impact on the largest players though. It hinges on a "calling their bluff", that high NW individuals won't change anything despite now being forced to annually liquidate assets to cover taxes. And this doesn't even touch on the immense impracticality of annually valuing assets. Or how to manage assets in illiquid markets, or how to sell 30% of a painting to cover 1% of it's mark-to-market value by year end.
The reason wealth taxes never go anywhere is because when you sit down and learn what wealth is, how it works, and what is practical, it makes the most sense by far to just tax things whenever they go back to cash.
Really the only genuine tax loop-hole is the step-up basis on inheritance. Everything else is just an elaborate deferral to pay taxes later.
> despite now being forced to annually liquidate assets to cover taxes
Allow paying the tax with assets. Put the assets into a black box sovereign wealth fund that's controlled by some mechanical algorithm which sells things at random as needed to fund the government budget. At scale this will be indistinguishable from a whole-economy index fund.
The best part about this is rich people can't beg off by saying "I have to liquidate stuff".
How do you pay with assets for real estate or boats or paintings? An IOU that can be cashed in when the asset is sold. Oh the boat is owned by an LLC so it never changes hands? No problem, the government has a share in the LLC too. (IANAL, IANAA so working out the loopholes is left as an exercise for the reader).
A second benefit is startup shares don't have to get hit with a capital gains tax before the startup goes public. Right now people sometimes pay taxes on shares that are eventually worth zero. Instead if this tax could be paid in startup shares, then there's no unfair tax bill.
As a condition of paying in assets, forbid the government from exercising any control over the assets. No shareholder voting, no board seats, not even choosing the paint color on the boat.
Additionally, this tax can't be on top of income tax. The whole point is to fix the worker-funded tax pyramid scheme. It has to be revenue-neutral with respect to income tax.
The bigger issue is, at least in the US, roughly 2/3 of assets of the wealthy have no meaningful liquidity. There is also no mark-to-market because in many cases these are idiosyncratic goods that may only find a buyer once over decades. Even some real estate markets only clear a single transaction on the scale of decades so any valuation is mostly fiction -- there are no comparables.
You could pay for these using the 30% of the assets that have some practical degree of liquidity but now you are putting massive downward pricing pressure on those because it is essentially a leveraged liquidation. Effectively, the total percentage of assets that are non-liquid would increase.
People tend to underestimate just how non-liquid the assets of the wealthy are. Most of that wealth isn't in stocks and bonds.
Real estate is a bad example because it's already subject to property taxes, which is a form of wealth tax. Maybe it doesn't need another wealth tax.
Private businesses are a better example. They don't trade on markets, sometimes don't have multiple shareholders. There already exist methods for valuing businesses (discounted cashflow, for example). Let the taxpayer pick one and make them stick to it.
> You could pay for these using the 30% of the assets that have some practical degree of liquidity
I already said "no liquidations, pay with assets". For non-liquid assets pay with IOUs on said assets. The government cashes in the IOU when the asset changes hands - whether it's by sale, gift, or inheritance. Yes that's an inheritance tax; who cares? If you want to add a wealth tax to real estate, this is the way to do it.
There are a surprising number of edge cases out there.
Quite a few assets can never clear a market — they have value in some abstract sense but no concrete sense. For example, assets that are legal to own and transfer but illegal to buy or sell.
Some commodity assets have value that it is nonetheless not always transferrable. A common example relevant to wealth taxes is intangible assets where value is bound in who owns it and not the asset per se. Most of the value vanishes the instant you transfer to e.g. the government.
Another common issue is that wealth taxes can directly conflict with existing load-bearing contracts. As a practical matter, these government can’t just void most contracts, including contracts the government is a party to, for the purposes of generating tax revenue.
All of which is why most real-world wealth taxes limit scope to a handful of liquid, legible securities and similar. But as a percentage of wealth, these are pretty small so you don’t collect much revenue.
> A common example relevant to wealth taxes is intangible assets where value is bound in who owns it and not the asset per se. Most of the value vanishes the instant you transfer to e.g. the government.
What's an actual example of this?
> Another common issue is that wealth taxes can directly conflict with existing load-bearing contracts.
Except for outsourcing, of course. The labor pool isn't limited domestically for many kinds of work. And for tech specifically, the labor pool is highly mobile. So if another country becomes the best place to go, then the labor pool will move there. The simple analysis probably results in, counter-intuitively, the opposite of what you want, which is a decline in the competitiveness of American tech
reply