Maybe LLM coding agents change the equation by making it much easier to adapt and use foreign and probably incomplete code. Getting you closer to competing with the original authors in a shorter amount of time than generating new code from scratch.
Nobody disputes that I own the copyright in a sound recording I made just by pushing the red button on my recorder. So it is a mystery to me that copyright to any sort of human conditioned machine generation is in dispute.
The sound recording analogy breaks down at the point where the recorder makes no creative decisions. Pressing record captures what is already there. Prompting Claude generates something that did not exist, through decisions the model makes about structure, naming, pattern, and implementation. The closer analogy is hiring a session musician and telling them the key and tempo. You own the recording under work-for-hire if they signed the right contract, but the creative expression in the performance is theirs unless explicitly assigned. The button you push to start the model is not the same button as the one on the recorder.
> Prompting Claude generates something that did not exist, through decisions the model makes about structure, naming, pattern, and implementation.
LLMs don't make decisions. Their output is completely determined by an algorithm using the human prompt, fixed weights, and a random seed. No different than the many effects humans use in image or audio editors. Nobody ever questioned whether art made using only those effects on a blank canvas was subject to copyright.
Fourier theory says that any sound, however complex, can be synthesized by summing sines and cosines. That's what an LLM does, if you twist the metaphor enough. It synthesizes complex outputs from simpler basis functions that are, or should be, uncopyrightable.
The fact that it inferred those basis functions from studying copyrighted works doesn't seem relevant. Nor does the fact that the "Fourier sums" sometimes coincide with larger fragments of works that are copyrighted. How weird would it be if that didn't happen?
Nobody is doing that, though. You might get a watermarked screenshot or stock photo now and then, or a couple of mostly-verbatim paragraphs from Harry Potter.
In any case, if the copyright mafia insists on butting heads with AI, they'll find that the fight doesn't quite play out the way it has in the past.
Here's a random high school in Northern California. Everyone on the team is beating 16.7 seconds in the 100m. For the 1600m there are six kids with times under 4m30s and another seven with times under 4m40s, all in the last month.
Not sure that disproves the point :) Most people have never been anywhere close to competing with the top 6 athletes at a high school with ~2k students.
OK, so let's do the math. There's about 25k high schools in the USA. Let's suppose they all have a track team, and let's assume that they all have 5 team members who can break 04:30 for 1600m. Sure, at some schools that's too few, but at others it is too many.
That gives us 125k high schoolers in the USA who can break 04:30 for 1600m. There are about 18M high school students. So of just the high school population alone, about 0.7% of them can do this.
Assuming there are the 4x as many adults that can do this as there are high school students, that gives us slightly less than 0.2% of the total US population capable of this.
We just have different ideas of what constitutes "mere mortals." 1 in 150 high school students or even 1 in 500 from the general population doesn't sound super human to me at all. Talented, yes but not god like.
> four wheels would double the rolling resistance and thus the effort required to push the cart
Rolling resistance = Crr * N where N = the normal force a.k.a. the weight bearing down on the wheel for a slow moving cart. If you use the same tires but double the axles you reduce the weight on each wheel by half leading to about the same rolling resistance for the vehicle as a whole.
> its narrow wheels minimize rolling resistance
This is a myth that has gripped the bicycle industry for decades but has been slowly dying over the last two. Fat tires have slightly lower rolling resistance at the same pressure and a smoother ride at the same rolling resistance. It takes more energy to lift a hard wheel over a small bump than for a softer wheel to deform and lift less so the difference becomes greater as soon as the road isn't perfectly smooth.
The main reason road racing tires only got a little fatter and aren't as fat as say e-bike tires is aerodynamics which is hardly applicable to hand carts.
> Unlike a van or a car, my handcart doesn’t need gasoline, electricity, or batteries, making it entirely independent from energy infrastructures.
TANSTAAFL - you're going to need more calories to compensate for the additional effort you're expending to move your handcart. Human food is far from the cheapest fuel and few people are able to obtain it "entirely independently from energy infrastructure."
Restaurant portion sizes have definitely increased - a lot - since the 1940s-50s. Maybe some minor pullback the last few years but still way larger than back then. A McDonald's Quarter-pounder was considered very large, that was in 1971, many sit-down restaurant burgers today are 5-8 oz.
One other thing to compare is business and health regulations. Compliance with that is certainly more involved and costly today than in 1940 and would account for part of the price.
The problem with this model is that the staff and insurance are essentially fixed costs, so if they sell 500 burgers on Saturday but only 250 on Tuesday, then the insurance cost-per-burger on Tues is double what it is on Sat. Staffing might increase by an extra body or two on the busy days but won't double, so it also has a much higher cost-per-burger on Tues.
I am not a restauranteur, just a customer (and observer) but I dont think many restaurant operators understand this concept either. Many seem to be raising prices to cover higher costs-per-item due to fewer customers to spread the fixed costs over. And then the higher prices turn more people off, now prices need to be raised again. Death spiraling themselves.
Insurance is not a fixed cost. Property and auto insurance are, but liability is a percentage of sales, its fixed for a year then adjusted for next years planned sales.
Unfortunately no mention of prices, so increase in portion sizes might be below inflation; and I suspect the former could be a strategy of compensation for inflation by making it seem less drastic ("yes it costs more, but we also made it bigger!")
One time when I was a kid and my dad and I were in line at Fuddruckers, we overhead someone else in line say "I don't think I could eat a third of a pound, so I'll have to get a half a pound instead. It's still a reference we laugh about over two decades later.
These prices adjusted for today's value seem off though. I'm guessing you'd be hard pressed to find a diner burger for $5.14 anywhere. No, fast food joints are not the same here and not part of this discussion.
Where is the discrepancy? I've never really trusted these "adjusted for inflation" type numbers. I'm not an economist so I have no idea how they are calculated, but they've always just felt off to me. Usually, the numbers are for something esoteric to me, but these are about something I have some familiarity. In my experience, the adjusted burger price is about half the actual cost of today.
A good rule of thumb is to ask "are you paying mostly for human labor or for machine labor"? The former is likely to be more expensive now than it was in the past and the latter is likely to be less expensive, all relative to general inflation prices.
A hot dog / hamburger at a diner is mostly human labor, so you'd expect it to be cheaper in the past.
Labor is typically around 30% of the final cost of prepared food in a restaurant.
Remaining 70% is 30% food costs (which has dropped drastically since the 50s), then 20-30% operations. Profit is whatever is left.
So a diner burger is not mostly labor but I honestly have no idea what these costs were 70 years ago. I'd love to know, seems like something is missing.
Food cost hasn't dropped because you can't even get the food they used to have. You have something that costs less now, but is worth even less than what it costs. And now that Sysco has completed it's eradication of all variety and competition, it doesn't even cost less any more.
Things just don’t really convert neatly because the shape of what people spend money on in life hasn’t evolved uniformly.
Food appears somewhat cheaper, housing much cheaper; but clothing and tools/appliances were much more expensive. Things like student debt and healthcare costs are also interesting to compare and wildly differ over time & place.
Also common for the average middle class person to spend a sizable percentage of their income on travel/vacation today; as I understand it that was quite uncommon before the mid 20th century.
Travel and vacation were much rarer. Many jobs gave only 2 weeks a year of vacation. Many jobs didn't include travel. That's changed with the invention of cheap airlines. Alas, some like SWA have changed their business model.
Well, the $5.14 figure is using the generalized inflation number derived by tracking the price of a specific basket of goods over time, across the entire country. This is a reasonable number to pick.
If you narrow down to Food for all Urban Consumers[1], it shifts to more like $5.24. If you look at "Food away from home in New York-Newark-Jersey City, NY-NJ-PA, urban wage earners and clerical workers, not seasonally adjusted" that number moves to $7.60. Which confirms your intuition: restaurant prices are way higher than the overall inflation rate predicts.
How do we explain the difference? A variety of ways. Maybe the burgers you get are "better" in some way. Bigger. Better cut of meat. More veggies and toppings. I wasn't around in 1959 and never ate at that specific diner, but it's a real possibility. In fact, this is explicitly called out in the FAQ[3]:
> Specifically, in constructing the "headline" CPI-U and CPI-W, the BLS is not assuming that consumers substitute hamburgers for steak. Substitution is only assumed to occur within basic CPI index categories, such as among types of ground beef in Chicago. Hamburger and steak are in different CPI item categories, so no substitution between them is built into the CPI-U or CPI-W.
There's also some other complicating factors to account for, like coupons and bundling. Like consider Applebee's Really Big Meal Deal deal. "NEW Big Bangin’ Burger with unlimited fries & soda, still just $9.99" Or you can order just the burger for... $15.99[4]. I don't even know how BLS copes with that and am sorta guessing they just take the a la carte prices for consistency, even though that likely overstates price levels consumers actually pay?
There are two diners near me (in NYC) where a burger is $5.25/$5.50 respectively.
(I don’t disagree with you directionally though; I think a nontrivial aspect of this is shifting expectations/norms around what passes for food service. Americans broadly want their food - even diner food - to be upclassed beyond a plain hamburger on a white bread bun.)
The Market Basket used to calculate the BLS CPI changes over time, which can make long range comparisons difficult.
I’ve read of political influence on the market basket to lower the reported rate of inflation by the incumbent party, but I’m not educated enough on the topic to give an opinion on if it happens.
Counter service family joints absolutely in the $5 area for standard ol' boring 1/4/lb. Maybe your definition of diner is different? There's a place by me with diner in the name that has a burger for $4.99.
> Simple inflation adjustment gives us x but the real price is more or less than x. Why is that?
Restaurant economics are a function of ingredient costs and labour. I suspect ingredient costs are close to OP's estimated multiples. But real wages are way up since the 1950s. Anything with a large labour component of costs will have tended to rise faster than inflation, which is an average of goods and services.
(There are specialised metrics if you actually wanted to dig into this question.)
Are you saying the prices listed were just for the ingredients and not the actual cost to the person ordering? They mentioned they saw the price in a photo which suggest it is what the person would be charged. I get that labor costs would cause an increase of raw ingredient price comparisons for total prices. But if you could pay buy a burger for a nickel but now need $10, there is a definite issue in just a "simple" adjustment that suggests you'd only need $5. If the numbers are that far off because the simple needs to be more advanced, what's the point of the simple numbers? Bad data is worse than no data.
That may be true, but I suspect that it’s also hard to compare apples to apples. A burger in 1959 is hard to compare to a burger today. Today’s burger almost certainly has twice as much meat. The invention of (and ubiquitous advertising of) the quarter–pounder means that everyone had to make their burgers larger to match. Sides are larger, drinks are larger, etc, etc.
Inflation is a measure of change in overall purchasing power.
What a specific purchase costs is highly dependant on the inputs, the cost of its labour (which might grow faster or slower than the average wage), and a lot of other factors.
Food is way more expensive today than it was 50 years ago. Airplane tickets are way cheaper. Everyone has a cellphone now, and middle class families have multiple cars, but a trip to the doctor will mean that ~15% of the population will be on the verge of not paying their bills. On the other hand, I have access to ~every major piece of music ever made for ~$15/month, so that's something.
I believe you, but in my own experience I've met more people who say this than who mean this.
Usually it's situational. People might genuinely like to be wrong when the novelty is fun or useful, for example in lab work or in low stakes classwork. However, they despise it with politics, their job, or anything else that might have actual consequences in their lives.
"In 1969, Schenck was largely overturned by Brandenburg v. Ohio, which limited the scope of speech that the government may ban to that directed to and likely to incite imminent lawless action (e.g. a riot)." - Wikipedia
Amazon employs around 900,000 people in logistics. The crude annual mortality rate in the USA is around 911/100,000. If there are 900,000 employees working eight hours a day then around seven people a day are dying of natural causes on their shift. This is without considering that they are being worked to the bone.
This only works if you assuming the mortality rates are evenly distributed. Most of the people who die are not working right until the end—and the conditions which lead to them dying usually aren’t compatible with a demanding job.
You are correct that it is a rough estimate but my point stands. While most of us will never experience the shock of someone dying at work, it is an every day occurrence at the scale of Amazon.
You have provided no evidence supporting that belief and brushing aside the obvious challenges makes it hard to believe you have done the math. I’d also note that if this was actually true, it would be more surprising that they didn’t have a policy for dealing with it and had to improvise on the fly.
The utter contempt you express for human life is abhorrent. Cloaking it in math only exacerbates your cruel disregard for, well, lacking shame in expressing such mental illness in public. I’d recommend therapy but you probably have a formula to justify not going to that either. Disgusting.
When I saw the article I recalled a doctor who worked at the sports stadium. Probably every stadium has a doctor on duty because there are medical emergencies any time you get 50,000 people together. Sometimes people die while they are still on the premises.
So I wanted to know how approximately how many people you would expect die of natural causes per day in a group of people as large as Amazon warehouse workers.
If you expect people to die every day while working in an Amazon warehouse and there was no cause of death disclosed for the unfortunate person referenced in TFA then the fact that he died is not news.
> spherical harmonics can have uses beyond lighting
This math is also used in Ambisonic surround sound though newer techniques use planewave expansion.
For games, the full-sphere encoding of Ambisonic B-format can be decoded for arbitrary speaker locations and the soundfield rotated around any axis. I'm not sure if its ever been used for a game though.
I noticed that (similarity between the graphs and the shapes of atomic orbitals), and assumed that was what the article was about. And it wasn’t, and never brought it up, so I was thinking maybe I was confused about the similarity. So thank you for showing me I was not.
A single Ambisonic B-format recording can be shipped and at runtime decoded into any coincident or near-coincident stereo pair pointing in any direction or into any surround sound format. It is a universal format that encodes the direction and intensity of arriving sound over a full sphere.
reply