Each of these phenomena have a name: there’s Jevons Paradox, which means, “We’ll spend more on what gets more productive”, and there’s the Baumol Effect, which means, “We’ll spend more on what doesn’t get more productive.”
```
I don't think that's exactly right. Jevons says "we consume more on what gets more productive" and Baumol says "the unit cost increases for that which is less productive".
The typical example for Baumol is the orchestra (or live music) which is today much more expensive than in the 1800s. I don't think we spend more in aggregate than we did in the 1800s!
Edit as I continue reading:
```
Other goods and services, where AI has relatively less impact, will become more expensive - and we’ll consume more of them anyway.
```
This definitely NOT the case. Basically the author is saying we will consume more of everything, which is not true! We famously stopped using horses and all the relevant industries.
The unit cost for horses, however, did increase!
What the author should be stating is that the new production bottlenecks will command a higher price and probably play a bigger role in the economy, but not everything gets to be a new bottleneck.
The modern day example that really made Baumol click for me is child care, particularly day care. It’s a highly labor intensive with basically minimal opportunities for productivity enhancements (due both to regulation and parental preferences, as well as just baseline sheer human decency). As the rest of the economy becomes more productive, the relative cost of child care goes up and up and up - which is why we now see situations where two-earner households can an entire after-tax income consumed by child care costs once they need to put 2-3 kids into daycare.
In economics, the Jevons paradox occurs when technological advancements make a resource more efficient *to use*; however, as the cost of *using the resource drops*, if demand is highly price elastic [PED], this results in overall demand [for the resource] increasing.
The cost to use tokens is already zero or deeply negative, depending on your accounting.
And the author is mistaking the cost of 'creating/extracting the resource' (token generation economics) with cost of 'using the resource' (calling an API). I don't think we'll understand the true cost of, and demand for AI services for quite some time.
I commend you for your imagination. Can I ask how you crafted the object to match the dimensions? I’m brand new to 3D printing and currently climbing the learning curve of printing itself but will want to start learning about doing my own modeling soon.
The only criticism I’d make is that patching drywall is dead simple and cheap and so your solution seems possibly a bit overengineered (and, while I’m at it, that Andreesen’s observation is both facile and meaningless and is probably a reflection more of the bids Marc Andreesen’s house manager gets than anything insightful about labor costs in America).
Welll, if you were good at drywall already, drywall patches would be faster and better. But if you are good at printing and scanning, and you enjoy that process, then it’s fine.
The challenge with the example is that “success” is personal preference. With plenty of examples, the success criteria are external.
The biggest actual problem would be in a fire, the PLA will burn and let the fire into the wall cavity, where drywall would maintain a barrier for much longer - that is why we have drywall in the first place, it is a decent fire barrier.
I would simply do the normal thing of covering the hole with drywall patch screen, covering that over with drywall joint compound, letting it dry, sanding, and painting. This is an under $50 trip to Lowes, and certainly cheaper than a flatscreen TV.
Cosmetically it's probably fine. The downsides all have to do with predictability and the ability to reason about what the wall is made of in the future.
A person who goes to eg hang a picture frame or shelf there will encounter a different material with different load bearing properties than expected. Pushing into the center of that area with EG a drill bit will not have the same physical response or give, and depending on how it was braced/integrated with the surrounding wall, the patch itself may be pushed or pulled out of place. Similar for anyone that leans on that area if it's at such a height.
The solution I was taught as a child is to saw the hole square, put a section of 2×4 behind it spanning the hole, held in place with a drywall screw through the drywall on each side of the hole, cut a square chunk of drywall small enough to fit in the hole, put a drywall screw through the middle of it into the 2×4, and tape, mud, sand, and paint.
I suspect that this procedure is faster and easier than taking a 3-D scan of the hole, 3-D printing a PLA patch, and gluing it in, but it does require most of an hour and the appropriate materials on hand.
It's a solution. There are better solutions, and far worse solutions (anyone who has worked to get a deposit back on a college rental has probably developed a few of their own), and most of them are all still fine because drywall isn't (shouldn't be) structural.
Crucially, even if you are completely unwilling to take a stab at a fix yourself, hiring a local handyman to patch a hole via some good enough technique should still be far cheaper in most places than buying a nice new TV.
But nothing is gonna ever beat buying a 2nd-hand framed picture or plaque or movie poster or grabbing a flyer from the junkmail on your porch and tacking it over the hole... And if you're determined to fix holes with a TV, you can probably find one used for about as cheap / free as any of the other choices. Which is what makes this such a stupid example - the cost of TVs, like framed images or furniture, spans from $0 to "as much as you're willing to pay". Hiring someone can also be arbitrarily expensive, but can by definition never be 0. So the comparison is rhetorical trickery and demonstrates nothing.
...other than, apparently, Andreessen's dissatisfaction with paying tradespeople.
It works so :shrug: I did the same to replace a part of a door frame I had to remove to make space for a washing machine 4 mm too wide. Nobody sell 400 mm of door frame so i just copied the frame shape, printed in 3 parts, and that was it. Filament color matched the frame one so I didn't have to paint.
that example of the radiologist review cases touches on one worry i have about automation with human-in-the-loop for safety. specifically that a human in the loop wont work as a safeguard unless they are meaningfully engaged beyond being a simple reviewer.
how do you sustain attention and thoughtfully review radiological scans when 99% of the time you agree with the automated assessment? i'm pretty sure that no matter how well trained the doctor is they will end up just spamming "LGTM" after a while.
The likelihood is that models will "box" questionable stuff for radiologist review, and the boxing threshold will probably be set low enough that radiologists stay sharp (though we probably won't do this at first and skills may atrophy for a bit).
This is also a free source training data over time so market incentives are there.
I have the same question about minor legislative amendments a certain agency keeps requesting in relation to its own statutory instrument. Obviously they are going to be passed without much scrutiny, they all seem small and the agency is pretty trustworthy.
(this is an unsolved problem that exists in many domains from long before AI)
How to you sustain attention of the other big X-ray use: security scanning? Most scanners will never see a bomb, so how do you ensure that they'll actually see one when it does happen?
The answer they've come up with is periodic tests and audits.
It’s not like people are piling into AI at the expense of other jobs, tech hiring and wages on general seem down relative to a couple of years ago. Hard to see the link between either effect and AI
Also weird that Dutch disease wasn’t mentioned at all, it actually seems more relevant.
The blue-line/red-line diverging price graph is such a great litmus test for someone's ability to do even a single round of interrogation.
Here's the question to ask:
Which of these have economies of scale, are scale-neutral, or have diseconomies of scale?
Ta-da!
Three sectors that have diseconomies of scale: education, healthcare, housing.
Essential services aggregate in the red because non-essential services that have diseconomies of scale... wait for it... never achieve scale! Then you have government step in because the services are important (often with hard-to-capture upside, ergo limited incentive for private investment to begin with), and now you have an aggregation of essential, expensive, government-involved services.
> The last piece of this economic riddle, which we haven’t mentioned thus far, is that elected governments (who appoint and direct employment regulators) often believe they have a mandate to protect people’s employment and livelihoods. And the straightforward way that mandate gets applied, in the face of technological changes, is to protect human jobs by saying, “This safety function must be performed or signed off by a human.”
I mean the role of you to go jail if something goes wrong is pretty important.... It unfortunately often doesn't work but the lengths at which well paid people go to avoid and then shorten prison sentences really should demonstrate this is the only punishment that works.
We cannot have computers solely in charge of stuff because the computer cannot have responsibility. If you take the radiologist out of the loop then you should take the responsibility.
> If you can make $30 an hour as a digital freelance marketer (a job that did not exist a generation ago), then you won’t accept less than that from working in food service. And if you can make $150 an hour installing HVAC for data centers, you’re not going to accept less from doing home AC service.
Plenty of people work jobs for less money because they enjoy the work more. I’m not sure if it’s worth reading what follows if most of the argument is predicated on this claim.
That might be true but that doesn't mean the effect is false in the aggregate. The classic example given in Baumol's paper was a violinist[1]. Sure, playing a violin might be more enjoyable than working in HVAC or churning out enterprise CRUD apps, and some people might even accept a paycut[2] to be a violinist, but that doesn't mean the effect isn't real. Despite zero productivity growth in being a concert violinist, wages for it has still risen, thanks to productivity growth elsewhere. People might be willing to take a pay cut to be a violinist, but not a arbitrarily large paycut.
I would have made a different argument. Often the "instantaneous" wage is high, but you're never told the volume. The data center HVAC job will be done one day and you'll have to wait for another data center to be built and possibly even move there.
The basic premise here is that productivity growth in one sector, increaes wages in other sectors. But we already know that productivity and wage growth are increasingly disconnected. Thsi si not to say the affect doesnt exist but that it must therefore be small or limited in scope.
https://www.epi.org/productivity-pay-gap/
There's a spike and subsequent drop after the pandemic, but that was from service workers being forced out of the labor force by lockdowns. If you exclude that, it follows a steady trend up. And yes, the chart has already been adjusted for inflation.
> Whereas, after a century of productivity gains, the average American middle-class household can comfortably manage a new car lease every two years, but needs to split the cost of a single nanny with their neighbors.
...
> Why it costs $100 a week to walk your dog (but you can afford it)
Two signs this article is coming from too-privileged a point of view for its observations to be meaningful or useful to 99% of the population.
>If you live in the United States today, and you accidentally knock a hole in your wall, it’s probably cheaper to buy a flatscreen TV and stick it in front of the hole, compared to hiring a handyman to fix your drywall.
Is it cheaper than buying a TV, and then having someone come install it on your wall? More meaningless drivel from Marc Andreesen.
Drilling a few holes to mount a TV is so straightforward that nearly everyone can do it, in contrast to patching drywall, so excluding the cost of hiring a professional for that case is reasonable. Granted, you might be able to learn how to do it with some youtube videos, trips to home depot, and trial/error, but the cost of that will quickly exceed the cost of a handyman if you value your time at all.
[+] [-] huevosabio|3 months ago|reply
Each of these phenomena have a name: there’s Jevons Paradox, which means, “We’ll spend more on what gets more productive”, and there’s the Baumol Effect, which means, “We’ll spend more on what doesn’t get more productive.”
```
I don't think that's exactly right. Jevons says "we consume more on what gets more productive" and Baumol says "the unit cost increases for that which is less productive".
The typical example for Baumol is the orchestra (or live music) which is today much more expensive than in the 1800s. I don't think we spend more in aggregate than we did in the 1800s!
Edit as I continue reading: ```
Other goods and services, where AI has relatively less impact, will become more expensive - and we’ll consume more of them anyway. ```
This definitely NOT the case. Basically the author is saying we will consume more of everything, which is not true! We famously stopped using horses and all the relevant industries.
The unit cost for horses, however, did increase!
What the author should be stating is that the new production bottlenecks will command a higher price and probably play a bigger role in the economy, but not everything gets to be a new bottleneck.
[+] [-] nocoiner|3 months ago|reply
[+] [-] mushufasa|3 months ago|reply
When you buy 10,000 handbags you pay the wholesale price whereas buying a single handbag can be quite expensive.
If there is way lower hose demand (volume of sales), the horse producers will have to charge a higher price per horse.
Thus, society in aggregate spends way less on horses while the price of a horse goes up.
[+] [-] CGMthrowaway|3 months ago|reply
Jevons describes the supply curve moving out, resulting in increased quantity
Baumol describes the supply curve moving back, resulting in higher prices
[+] [-] lkey|3 months ago|reply
| 'With radiologists, I’m totally speculating and I have no idea what the actual workflow of a radiologist involves...'
Speculation from a place of ignorance is suitable for bar-side musings, not an article that wants us to take it seriously.
[+] [-] lkey|3 months ago|reply
In economics, the Jevons paradox occurs when technological advancements make a resource more efficient *to use*; however, as the cost of *using the resource drops*, if demand is highly price elastic [PED], this results in overall demand [for the resource] increasing.
The cost to use tokens is already zero or deeply negative, depending on your accounting. And the author is mistaking the cost of 'creating/extracting the resource' (token generation economics) with cost of 'using the resource' (calling an API). I don't think we'll understand the true cost of, and demand for AI services for quite some time.
[+] [-] MarkusQ|3 months ago|reply
[+] [-] TekMol|3 months ago|reply
What does the HN panel say? Is it a solution? Or does it have any downside?
[+] [-] nocoiner|3 months ago|reply
The only criticism I’d make is that patching drywall is dead simple and cheap and so your solution seems possibly a bit overengineered (and, while I’m at it, that Andreesen’s observation is both facile and meaningless and is probably a reflection more of the bids Marc Andreesen’s house manager gets than anything insightful about labor costs in America).
[+] [-] abakker|3 months ago|reply
The challenge with the example is that “success” is personal preference. With plenty of examples, the success criteria are external.
[+] [-] csours|3 months ago|reply
The biggest actual problem would be in a fire, the PLA will burn and let the fire into the wall cavity, where drywall would maintain a barrier for much longer - that is why we have drywall in the first place, it is a decent fire barrier.
[+] [-] nom|3 months ago|reply
[+] [-] NoGravitas|3 months ago|reply
[+] [-] RickS|3 months ago|reply
A person who goes to eg hang a picture frame or shelf there will encounter a different material with different load bearing properties than expected. Pushing into the center of that area with EG a drill bit will not have the same physical response or give, and depending on how it was braced/integrated with the surrounding wall, the patch itself may be pushed or pulled out of place. Similar for anyone that leans on that area if it's at such a height.
[+] [-] kragen|3 months ago|reply
I suspect that this procedure is faster and easier than taking a 3-D scan of the hole, 3-D printing a PLA patch, and gluing it in, but it does require most of an hour and the appropriate materials on hand.
[+] [-] qwertycrackers|3 months ago|reply
[+] [-] Shog9|3 months ago|reply
Crucially, even if you are completely unwilling to take a stab at a fix yourself, hiring a local handyman to patch a hole via some good enough technique should still be far cheaper in most places than buying a nice new TV.
But nothing is gonna ever beat buying a 2nd-hand framed picture or plaque or movie poster or grabbing a flyer from the junkmail on your porch and tacking it over the hole... And if you're determined to fix holes with a TV, you can probably find one used for about as cheap / free as any of the other choices. Which is what makes this such a stupid example - the cost of TVs, like framed images or furniture, spans from $0 to "as much as you're willing to pay". Hiring someone can also be arbitrarily expensive, but can by definition never be 0. So the comparison is rhetorical trickery and demonstrates nothing.
...other than, apparently, Andreessen's dissatisfaction with paying tradespeople.
[+] [-] aeve890|3 months ago|reply
[+] [-] parpfish|3 months ago|reply
how do you sustain attention and thoughtfully review radiological scans when 99% of the time you agree with the automated assessment? i'm pretty sure that no matter how well trained the doctor is they will end up just spamming "LGTM" after a while.
[+] [-] jwahba91|3 months ago|reply
[+] [-] CuriouslyC|3 months ago|reply
This is also a free source training data over time so market incentives are there.
[+] [-] IIAOPSW|3 months ago|reply
(this is an unsolved problem that exists in many domains from long before AI)
[+] [-] bryanlarsen|3 months ago|reply
The answer they've come up with is periodic tests and audits.
[+] [-] Ar-Curunir|3 months ago|reply
https://www.a16z.news/p/why-ac-is-cheap-but-ac-repair-is/com...
[+] [-] wonnage|3 months ago|reply
Also weird that Dutch disease wasn’t mentioned at all, it actually seems more relevant.
[+] [-] estearum|3 months ago|reply
Here's the question to ask:
Which of these have economies of scale, are scale-neutral, or have diseconomies of scale?
Ta-da!
Three sectors that have diseconomies of scale: education, healthcare, housing.
Essential services aggregate in the red because non-essential services that have diseconomies of scale... wait for it... never achieve scale! Then you have government step in because the services are important (often with hard-to-capture upside, ergo limited incentive for private investment to begin with), and now you have an aggregation of essential, expensive, government-involved services.
[+] [-] littlestymaar|3 months ago|reply
[+] [-] lesuorac|3 months ago|reply
I mean the role of you to go jail if something goes wrong is pretty important.... It unfortunately often doesn't work but the lengths at which well paid people go to avoid and then shorten prison sentences really should demonstrate this is the only punishment that works.
We cannot have computers solely in charge of stuff because the computer cannot have responsibility. If you take the radiologist out of the loop then you should take the responsibility.
[+] [-] meken|3 months ago|reply
Plenty of people work jobs for less money because they enjoy the work more. I’m not sure if it’s worth reading what follows if most of the argument is predicated on this claim.
[+] [-] gruez|3 months ago|reply
[1] https://en.wikipedia.org/wiki/Baumol_effect
[2] https://en.wikipedia.org/wiki/Compensating_differential
[+] [-] imtringued|3 months ago|reply
[+] [-] 01HNNWZ0MV43FF|3 months ago|reply
Says a lot about American politics, figuratively speaking
[+] [-] josefritzishere|3 months ago|reply
[+] [-] techblueberry|3 months ago|reply
[+] [-] gruez|3 months ago|reply
https://fred.stlouisfed.org/series/LES1252881600Q
There's a spike and subsequent drop after the pandemic, but that was from service workers being forced out of the labor force by lockdowns. If you exclude that, it follows a steady trend up. And yes, the chart has already been adjusted for inflation.
[+] [-] NoGravitas|3 months ago|reply
...
> Why it costs $100 a week to walk your dog (but you can afford it)
Two signs this article is coming from too-privileged a point of view for its observations to be meaningful or useful to 99% of the population.
[+] [-] biophysboy|3 months ago|reply
Is it cheaper than buying a TV, and then having someone come install it on your wall? More meaningless drivel from Marc Andreesen.
[+] [-] gruez|3 months ago|reply