I feel like you're ignoring the size of each of these impacts.
Sitting around a campfire once a month will likely increase your cancer risk. But it increases your cancer risk way, way less than smoking a pack of cigarettes every day.
Similarly, the World Health Organization is confident that smoked meats increase cancer risk. But that makes no statement about how much more likely you are to get cancer. I'm confident that eating undercooked chicken raises my risk for salmonella; I'm equally confident that eating a raw and expired piece of chicken also raises my risk of salmonella. But I would not say that each of those raises my salmonella risk by the same amount.
> This person probably had the best diet in the world and received top tier care, so how did he end up with something like that?
If you lower your chances of dying of every common thing, your relative chance of dying of a rare or incurable disease goes up. If I cured every disease but cancer, my odds of eventually dying of cancer would be very high.
>What else would this be but intelligence? Imagine if your AI didn’t know what words mean and couldn’t do math and you still tried to make the assertion that it was just as intelligent.
Counterpoint: If I don't teach a kid to read and then ask them to do a test that has reading as a requirement, I can say two contradictory things:
1. This kid isn't intelligent.
2. I don't know what this kid's intelligence is; I gave them a test on something they weren't taught.
What you're describing is in line with 1. That's completely fair and definitely part of what I'd consider intelligence.
This paper is making a statement more in line with 2. I've seen similar conclusions elsewhere; the theories are things like "education is less and less aligned with what IQ tests measure."
In the context of fitness, I feel like 2 is more in line with what we'd be thinking about?
Is there a reason you expect IQ to be unbounded? Feels like there are physical constraints, at which point you're sampling from a truncated normal distribution (or similar).
The abstract quite literary cautions against the way you're interpreting the data. The relevant quote:
> Notably, these gains do not uniformly translate to a rise in underlying GMA, suggesting the presence of domain-specific improvements and test characteristic changes over time. Conversely, the observed decline is primarily due to decreases in word comprehension and numerical reasoning tests, also reflecting specific abilities not attributable to changes in the latent GMA factor. Our findings further challenge the validity of claims that changes in the general factor drive the Flynn effect and its reversal. Furthermore, they caution against using these scores for longitudinal studies without accounting for changes in test characteristics.
My understanding is that Musk doesn't have a controlling share of Tesla. So by the letter of the law, a SpaceX acquisition would qualify as a change in control (because we've gone to a new majority owner).
>What would it mean if SpaceX buys Tesla though? Does the combined market cap count? That would be wrong.
I took a look at the proxy statement as they have it outlined in this [0] document (Proposal 4). As currently formulated, Musk has 12 operational goals (like ship 1 million robots or hit 50 billion EBITDA) and 12 market cap goals. These need to be paired together for the shares to vest; so, if he reaches the first market cap number, he also needs to fulfill an operating goal for the first set of shares to be earned. These earned shares vest in ~5 years.
This comes with a huge caveat. If Tesla changes control, those operational goals go out the window and his stock award is based solely on market cap (along with the shares immediately vesting). The share price for this is the greater of:
1. The last traded Tesla price prior to the acquisition, or
2. The per share price outlined in the acquisition.
The "easiest" way to take advantage is to IPO SpaceX (which he did), pump up the market cap, acquire Tesla for a sizable premium, and vest as much of the stock award as you can. It means you get to avoid the operational goals entirely and vest a bunch of Tesla (but soon to be SpaceX) stock.
True, but "a better apple" is unequivocally less impactful than "you no longer have to walk to the river any time you want to drink a glass of water."
I think it's easier to see if you think in 20 year increments. The difference between 1920 and 1940 is way larger than the difference between 2000 and 2020, and I don't think it's particularly close. Going from "antibiotics haven't been discovered" to "antibiotics become widely manufactured" in 1945 is a huge difference, just as one example.
Yeah, it gets weird when you start trying to compare human versus AI energy demands because you can turn a computer off, but you can't really turn a human off. Most studies indicate that humans can do 3-4 hours of high-level knowledge work in a day. An AI is not limited by these restrictions; it could do 24 straight hours of work if you wanted it to. So do we count all the energy the human burned for the remaining 20 hours of the day and allocate it to those 4 hours? Or are we just comparing "energy spent working on the task" for the AI and the human?
>The innovations aren't better versions of things we had in the past, they are more so unique new inventions. Plumbing is not electricity is not globalized food chains is not computers.
This is only sort of true. Plumbing is an improved version of a well pump, which is functionally (if not technically) an improvement over walking down to the river for a pot of water. Globalized food chains are mostly improvements on the supply chain we had (boats got faster, so things can travel from farther away places more quickly, refrigeration got added to existing modes of travel, etc.).
>Also we don't need to make more land to have more people. We can make more habitable living space within the same amount of land area, especially in countries like canada and the USA where we dumped a bunch of low-density housing absolutely everywhere, we just choose not to do that.
I didn't mean to imply that we can't increase housing density. But I think it's clear that having 2.3x more people than 1950 means that all else being equal, people will need to settle for less land. There's just more demand per square mile.
> in one assignment I remember comparing the energy outputs between the human and robot equivalents of different tasks, whether or not the robot was humanoid in how it was designed
So I think the point in this context is relevant, even if it's apples to oranges.
This point is valid. However, lifestyle improvement rate is something that's slowing over time because of physical constraints.
For example, the vehicle mortality rate is 1.44 per 100 million miles driven. That's down 17% from 2000 (so 25 years ago). However, the change from 1975 to 2000 was 53%. That's because as we get closer to 0, it gets harder and harder to improve those rates. On this metric at least, I don't think another 25 years will result in a noticeable amount of improvement?
In the other direction, some things will become scarcer (and therefore cost more). Real estate is the obvious one; we can't create more land, and we keep having more people. Easily accessible drinking water is another; desalination is getting cheaper, but it's still way more expensive than pumping aquifer water.
And some improvements are necessarily 1 time things. You can get tropical fruits year round, but that's been widely available since the 80-90's from what I can tell. So come 20 years from now, what will people be able to buy in a grocery store that I can't buy right now?
Studies on grandmaster chess players indicate that at most you burn 10% more calories when engaged in deep thought than when you're at rest. So the energy "attributable" to an hour of knowledge work is like 10 calories (average sedentary calorie burn is like 80-100 per hour; add a max of 10% for the thinking gets you 8-10 calories). A pound of potatoes is like a buck and is about 320 calories. So you're looking at like 3 cents an hour at most to cover that energy burn. It's definitely even less; I certainly don't think as hard as a grandmaster chess player.
Then, assume power costs 20 cents per kilowatt hour (US avwrage) To match the human 3 cents per hour, you need an average of 150 watts of power drawn per hour. That's in the range of a budget graphics card, but not much past there.
However, if you sleep instead of sitting around, you can probably make AI cost competitive. Sleeping drops your metabolic rate by more, and lying down in bed (as opposed to sitting) also reduces calorie burn. Combined, you can reduce your burn by like 30 calories an hour. At the new 9 cents per hour human cost, you can afford to run a higher end graphics card at ~450 watts per hour. That puts you in RTX 3090 range.
> but there were no privacy features before. so we’re actually still much better off than we were for hundreds of years before this.
One notable thing we have today that we didn't have 100 years ago is a computer. Before, you could reasonably assume that recreating individual records wasn't feasible, at least not on a large scale. You can't assume that now. A 4 digit password was safe for hundreds of years, but it would be a security lability today for the same reason.
I assume (and this is a big assumption) that the US government will be focused on limiting access to the latest model, not necessarily everything smarter than Fable 5. Having access to the frontier model from a year ago (Sonnet 4.7-ish) wouldn't really help you from a cybersecurity perspective.
I think there's a world where the US funds development of the next model in exchange for exclusivity, at which point they could "release" the previous version from that exclusivity.
Your intuition is largely correct. For any given level of adaptive problem solver, approximately 55% of those people share the same literacy level (i.e. a Level 1 adaptive problem solver is also a Level 1 for literacy around 55% of the time). 35% have a reading level that's 1 up from their problem solving ability, and around 10% have a literacy level 1 below their problem solving ability.
The balance here is that good schools are one of the amenities that keep property values high and they have pretty high fixed costs.
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.
The main cost benefits are that you don't have to pay for marketing, underwriting, plan administration for a bunch of different plan types, negotiating with thousands of employers on the premium they'll pay, 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.
The insurance company that provides your ACA plans gets money from the government for doing so. How much money they get is tied to a few things (not an exhaustive list):
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.
I'm going to link my favorite survey [0] to illustrate this point. Its called the PIAAC, and it's a internationally standardized assessment of adult skills in literacy, numeracy, and problem solving.
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.")
Sitting around a campfire once a month will likely increase your cancer risk. But it increases your cancer risk way, way less than smoking a pack of cigarettes every day.
Similarly, the World Health Organization is confident that smoked meats increase cancer risk. But that makes no statement about how much more likely you are to get cancer. I'm confident that eating undercooked chicken raises my risk for salmonella; I'm equally confident that eating a raw and expired piece of chicken also raises my risk of salmonella. But I would not say that each of those raises my salmonella risk by the same amount.
See the difference?