America can feed every person in the country easily, we import luxury goods cause we like coffee and out of season vegetables not because we’ve lost the production capability. If anything we are over subsidizing calories.
We also have enough nukes to glass the surface of the planet, enough firearms for every man woman and child to dual would them and 7x the rest of the worlds military.
The US’s people don’t need protection from other nations we need protection from each other.
A significant chunk of the "defense" spending is VA. Interest kind of shouldn't count? It's the cost of borrowing for our past spending on, you guessed it, mostly social programs.
Defense spending used to be a higher percentage of our budget than now, so if anything you should assign more of that to the debt.
About 16% of the spending is straight up defense spending and about 5% is the VA. I personally think it’s crazy to severe them and wish the va was part of the DoD. But people can disagree.
You only get to a majority by doing two misleading things… blaming the national debt on entitlements instead of general fund deficits, and counting Social Security (which is funded by its own payroll taxes) as a welfare program.
The actual means-tested safety net (Medicaid, food assistance, SSI, housing) is only around 15% to 20% of federal spending.
Social security is a welfare program. As is Medicare. They are odd regressive forms of them but they are tax funded welfare programs full stop.
As for fund deficits there are basically 4 things the federal government pays for, social security, Medicare, Medicaid and the military. Everything else is noise or interest. So when accounting for the debt load of that interest it only makes sense to assign some part of it to those 4 things.
Stop repeatedly changing definitions and accounting rules until the desired conclusion becomes true. It's tiresome.
In U.S. policy terminology, Social Security and Medicare are social insurance programs. Medicaid, SNAP, SSI, and housing assistance are means tested safety net programs.
Yes, Social Security, Medicare, Medicaid, and defense make up a majority of federal spending. But your conclusion only works by relabeling Social Security and Medicare as welfare, dismissing the rest of government as noise, and then assigning an arbitrary share of decades of accumulated interest expense to those programs.
Your last move is especially egregious. Federal debt reflects the cumulative gap between all federal spending and all federal revenue over time. You can't figure out what caused it by looking at today's largest spending categories.
Social Security's current shortfall is primarily the result of a demographic imbalance Congress has known about for decades and repeatedly declined to fully address. That makes it particularly silly to retroactively assign some arbitrary share of the rest of the federal government's accumulated debt to Social Security.
The accurate way to talk about this is that Social Security and the major health programs are a very large share of federal spending. "The social safety net is most of the budget" is a conclusion manufactured by your definitions.
Social safety net isn’t rigorously defined in the first place so acting like the definitions are being twisted isn’t a reasonable counter argument.
Pick what labels you want to throw social security, Medicare and Medicaid into. That bucket is about 46% of federal spending and increasing fast. It’s quite easy to get to 50% without including debt payments on things that look like social safety nets (snap, aca subsidies etc).
More importantly as you point out social security, Medicare and Medicaid are growing as a percentage of the budget (and a percentage of gdp) and are not discretionary spending. Those (and defense) are the only places we have any lever to change US debt other than increasing collection. And it’s not politically viable to change them.
You’d have to tariff fruits and vegetables like crazy to make the American varieties more expensive than oceanic shipped produce. That seems a fairly steep price for the Canadian consumer.
Which highlights the problems with tariffs, they are self harming. You have to be willing to give up better alternatives for some other reason, and food is a hard one.
> You’d have to tariff fruits and vegetables like crazy to make the American varieties more expensive than oceanic shipped produce.
Bullshit. I already see in the supermarket (Canadian Superstore) oranges from South Africa, lemons from Morocco, broccoli from Mexico, mangoes from Malaysia, fish from the Philippines, coconut from Vietnam, kiwis from Italy, etc.
Yes, oceanic shipping adds to the cost. But most of those countries already have cheaper labor, and that partially offsets the extra cost.
Mangoes and South Asian fish don’t grow close to you no matter what. You don’t think bull shipping Mexican tomatoes via the ocean isn’t going to cost more than the same tomato shipped by the US land route?
If you were going to take that tariff money to build up a local tomato industry it might make sense, Canada already has a growing hot house industry. But don’t kid yourself, tariffs cost the consumer and for low margin highly seasonal and subject to going bad goods like produce is about the worst case scenario.
If you want people to buy stuff from you Americans then just try to stop pushing forward American imperialism down into everyone's throats. This isn't just about money, it is about my country's sovereignty.
Nope. Not how prices work. If you have a replacement good that would be cheaper but you artificially inflate it with a tax you are forcing that tax on every transaction.
Again, if you think that tax is valid, feel free but tariffs costs are born by the consumer in the tariff applying country, ask us how we know…
Sadly most of us do realize it. But this is why tariffs are so pernicious. Because you hate me, you are going to tax yourself. It’s a negative sum world view. It’s bad enough that Trump believes the nonsense, he’s been able to export it.
This is not what most public libraries do with donation books. They mostly just recycle them.
The very act of cataloging donation books takes a huge amount of resources that the collection managers don’t have.
Many public libraries don’t have a collection management department at all! They outsource that function entirely to vendors and those vendors mostly source new books and spend resources preparing them for the hard use of a library (changing bindings, uploading and cleaning metadata, tagging, etc).
Decommissioning is done by volunteers who just look at the condition of books and chuck the bad looking ones in a bin for recycling.
My wife worked in this industry on the vendor side, your model of public libraries is closer to a small subset of certain big city libraries narrowed to their rare and research departments. The median book bought by a library is a Danielle Steele romance novel packaged for library consumption and sent to a small town library system that will be lucky to have a single professional librarian for the whole system.
I work for a large enterprise and fable/mythos are specifically banned because of the difference in data retention agreements from those models to other Anthropic models.
I have found it extremely helpful to keep the “token prediction machine” metaphor front and center in my work with llms, not because it expresses some fundamental limit of the technology but because it allows me to have a mental model about where and how to use the technology in my process.
Contrast this to the “chat” mental model. If you are chatting with someone and they lie to you, that is a transgression. A bad prediction is just a normal outcome. These are _human_ judgements but you address transgressions differently than normal but incorrect outcomes. That’s a helpful distinction for me when building my tooling.
If anything, that we are getting so much traction out of text prediction is the most magical part of llms. But it doesn’t change that you shouldn’t treat what they output the same as what a human says.
Hmm, what specific differences does it make in practice?
For me, the huge one is that LLMs are currently bad at learning from experience. I don’t trust any kind of automatic MEMORY.md or whatnot; in fact I greatly prefer starting from a clean slate each time because the LLM’s baseline general knowledge is so good.
In terms of accuracy and “lying”, I don’t really see a huge difference. Most LLMs are unfortunately a bit sycophantic and over-confident, but you sometimes see that in people as well.
The biggest difference is I view any interaction with an llm that doesn’t produce an artifact that can be verified with skepticism.
That is I very rarely even use a “chat” interface anymore. Under the covers I’m still using the mainline agents but I’m treating them more like black box Unix tools that take text in and spit text out.
My interactions are evolving to have very small llm cores wrapped with old fashioned deterministic programming.
This makes the boundary/verification step more deterministic and less susceptible to bad predictions. But it also has the side effect that I rarely interact with llm output that has “tone” (for instance sycophantic output). That usually only happens when I’m debugging what went wrong, particularly around model upgrade cycles.
I can get extremely dumb models to get our code style correct because of those guard rails and a specific style document.
reply