6 ms·

They discuss the fact that we are seeing linear gains from exponential investment in several fields (healthcare, semiconductors to name a few). This is a fascin

They discuss the fact that we are seeing linear gains from exponential investment in several fields (healthcare, semiconductors to name a few). This is a fascinating problem.

martin_a 7y ago
Sounds like pareto principle in final stages.
- netcan 7y ago
  Wouldn't that be declining returns from linear investments?
  - martin_a 7y ago
    Hm, you're probably right. Not sure though if it looks like linear returns when you choose the timeslot short enough.
  - mytailorisrich 7y ago
    Linear gains from exponential investments means declining returns.
netcan 7y ago
They discuss healthcare & higher education in parallel, and I do think the cost explosion in both is worth noting... But the differences are as important as the similarities. Some big healthcare cost increases come from increased consumption. It's not just incentives, bureaucracies and inefficiencies that caused cost increases. Once upon a time we had less medical care to consume, fewer procedures, cancer treatments and other such costly stuff. Medical care costs were mostly clinical, doctors salaries and such. Prenatal care once meant occasional visits to a doctor or nurse. These days, it also means ultrasounds, genetic testing, more blood tests & such. That costs more. There are just more available treatments to treat patients, especially the already high-cost stuff like cancer treatment, surgery.... These are new things that the medical industry are doing, and cost money. It's at least partly a "we buy more" situation, not just price inflation. A sociology degree, OTOH, is not different in a than it was in the 1950s. There are nearly no non-efficiency related cost increases in higher-ed. No new, expensive stuff is being used to teach these students. Prices just increased, and most of that got observed into administration. In terms of ROI at a societal level... Idk if it's usefully measurable, for medicine at least. There definitely aren't massive, healthcare-related increases I gross life expectancy and such. Does that mean bad ROI? This gets philosophical.
- throwawaymanbot 7y ago
  fascinating, but pity Marks Company is allowing the cancer of fascist content to be delivered with impunity to everyone.
- jeremyw 7y ago
  > It's not just incentives, bureaucracies and inefficiencies that caused cost increases. This might be true if there weren't stark counterexamples, e.g. Singapore at 1/4 the cost, with better outcomes. Notably there, government dollars flow through to consumers who choose among competitors by price and quality.
  - pmohun 7y ago
    There is also a lower focus on end of life care in Singapore, which makes up the vast majority of spending in the American system. These issues are cultural, not just political or economic. I am looking forward to seeing what PC and TC can come up with for a common framework. This seems to be a tough problem.
    - keenmaster 7y ago
      The healthcare sector constitutes 18% of our GDP, a truly insane amount. That is 7-8% (~1.5 trillion) more than it should be. Excessive end-of-life care explains almost none of that. The two biggest issues in American healthcare are: - Zero price transparency - Payment through intermediaries like insurance companies and your employer, who often pays half of your insurance premiums and other healthcare expenditures With no price transparency, prices are just made up with little connection to actual value delivered. Because of the middlemen, most patients don’t feel the full cost of healthcare. Patients in aggregate still pay for the full negotiated cost, just indirectly in the form of higher premiums and forgone wages (marginal premiums paid by the employer). At this point, Medicare for All is our best hope. A true market-based healthcare economy, as great as it may be, is a pipe dream. With Medicare for All, the government would set reimbursement rates in stone. Most doctors who want a substantial amount of patients would have to accept those rates. Unlike insurance companies, the government would not be able to aggressively increase its “premiums,” or taxes, every year and it won’t get kickbacks from hospitals for high cost care. The government would also set empirically backed standards of care that lower their cost burden with no impact on average lifespan. We could potentially allow for parallel “luxury care” purchased by the patient if they want to be able to expedite their non-emergency knee replacement from 6 months to 2 weeks. The government would be incentivized to launch health initiatives and laws that help lower the rates of obesity and diabetes in the population. Lastly, the government would also be incentivized to increase the supply of doctors. There are multiple instances in the past when medical organizations lobbied or misled Congress into over-limiting the supply of medical residencies. They also had a hand in restricting the number and size of medical schools. So yes, overpaid, scarce medical professionals are a big part of the cost burden, and their representative institutions are making sure they stay overpaid. Only recently have medical organizations lobbied to increase the number of residencies again. That’s just damage control. The American people should not accept compromise on healthcare any longer. That’s what the ACA was, and it did not do enough. Medicare for All will bring us in line with the rest of the developed world which has better health outcomes at a much lower cost. We can save >$1 trillion and put that money towards things like a national high speed rail network, anti-poverty efforts, government-subsidized corporate internships, and an accredited national online education and retraining platform with the best professors on every subject imaginable.
hanniabu 7y ago
I'd think it's more of a problem to assume they should continue to rise with exponential gains....
PunchTornado 7y ago
to me this is a meaningless question. it's so broad with so many variables, so many assumptions that you might as well ask if God exists and you'd have the same chance of answering it.
Barrin92 7y ago
in semiconductors, it's a natural result of hard physical limits driving up costs for innovation. In the economy overall it's institutional neglect for deep innovation, basic R&D and instead allocation of capital to eCommerce, finance, retail or the gig economy, with ironically companies like facebook contributing to the misery by keeping engineering talent occupied with fiinding better ways to get people to click on ads.
chubot 7y ago
That reminds me of deep learning. Investment in computing resources is definitely increasing exponentially over time: https://openai.com/blog/ai-and-compute/ https://openai.com/blog/ai-and-compute/ https://techcrunch.com/2019/11/10/the-post-exponential-era-of-ai-and-moores-law/ https://techcrunch.com/2019/11/10/the-post-exponential-era-o... Tthere doesn't seem to by any evidence of exponential gains in performance, even on their own benchmarks (which importantly don't measure "intelligence", but something more narrow): https://blog.piekniewski.info/2019/11/18/late2019-the-wizards-of-oz/ https://blog.piekniewski.info/2019/11/18/late2019-the-wizard... Even though we are doubling the compute available to AI every 3.5 months, so a factor 10x per year, since say 2015, so some 10.000x, the classification performance as measured by top-1 accuracy on ImageNet had barely moved. Now arguably top-1 accuracy on ImageNet is probably not the best measure, but still that seems somewhat striking. The diminishing returns are visible elsewhere too, even that Open AI Rubic's cube is somewhat indicative. From an outside perspective, it seems that everyone is coming around to the fact that self-driving cars is harder than expected. There are no "exponential gains" in that field -- to the contrary, it seems like 90% of the problem was solved several years ago, and you can now take undergraduate courses that teach the basics. But going from 90%, to 99%, to 99.9%, etc. is taking decades, not months or years. There seems to be something fundamentally wrong with the current approach. I believe at the time the OpenAI blog post was supposed to be "impressive". Certainly I felt a bit of that -- "hey something important is going on there!" I believe that in 10 years it will look like a waste of effort. There will be a different, more efficient approach to solving these types of problems. ---- Peter Thiel is one of the only "major tech figures" speaking out against AI, which I appreciate. And the "AI is communist" meme is pretty memorable, and I don't think wholly inaccurate. I listen to him for his contrarian opinions. I just wish he didn't have these random "tribal conservative" outbursts. The most recent one was a video on here where he was rambling about Obama saying something far worse than George Bush and the Iraq war. I guess it was about sending their daughter to Harvard, but saying that lots of schools are just as good. He was trying to get some points from his conservative audience. I guess it gets attention, but that kind of tribalism undermines the rational points that he generally makes. (I realize the OP doesn't feature Thiel but he was brought up elsewhere in this thread, and is one of the people also lamenting the lack of technological progress despite investment, although maybe due to lack of investment in certain areas too)
nugget 7y ago
One purely linear constraint is time. In many health care fields, it simply takes time for clinical studies to collect data, from which the next generation of clinical studies can be planned, and there's no good way to meaningfully speed up that process.

They discuss the fact that we are seeing linear gains from exponential investment in several fields (healthcare, semiconductors to name a few). This is a fascinating problem.

martin_a 7y ago

Sounds like pareto principle in final stages.

netcan 7y ago
Wouldn't that be declining returns from linear investments?
- martin_a 7y ago
  Hm, you're probably right. Not sure though if it looks like linear returns when you choose the timeslot short enough.
- mytailorisrich 7y ago
  Linear gains from exponential investments means declining returns.

netcan 7y ago

They discuss healthcare & higher education in parallel, and I do think the cost explosion in both is worth noting... But the differences are as important as the similarities. Some big healthcare cost increases come from increased consumption. It's not just incentives, bureaucracies and inefficiencies that caused cost increases. Once upon a time we had less medical care to consume, fewer procedures, cancer treatments and other such costly stuff. Medical care costs were mostly clinical, doctors salaries and such. Prenatal care once meant occasional visits to a doctor or nurse. These days, it also means ultrasounds, genetic testing, more blood tests & such. That costs more. There are just more available treatments to treat patients, especially the already high-cost stuff like cancer treatment, surgery.... These are new things that the medical industry are doing, and cost money. It's at least partly a "we buy more" situation, not just price inflation. A sociology degree, OTOH, is not different in a than it was in the 1950s. There are nearly no non-efficiency related cost increases in higher-ed. No new, expensive stuff is being used to teach these students. Prices just increased, and most of that got observed into administration. In terms of ROI at a societal level... Idk if it's usefully measurable, for medicine at least. There definitely aren't massive, healthcare-related increases I gross life expectancy and such. Does that mean bad ROI? This gets philosophical.

throwawaymanbot 7y ago
fascinating, but pity Marks Company is allowing the cancer of fascist content to be delivered with impunity to everyone.
jeremyw 7y ago
> It's not just incentives, bureaucracies and inefficiencies that caused cost increases. This might be true if there weren't stark counterexamples, e.g. Singapore at 1/4 the cost, with better outcomes. Notably there, government dollars flow through to consumers who choose among competitors by price and quality.
- pmohun 7y ago
  There is also a lower focus on end of life care in Singapore, which makes up the vast majority of spending in the American system. These issues are cultural, not just political or economic. I am looking forward to seeing what PC and TC can come up with for a common framework. This seems to be a tough problem.
  - keenmaster 7y ago
    The healthcare sector constitutes 18% of our GDP, a truly insane amount. That is 7-8% (~1.5 trillion) more than it should be. Excessive end-of-life care explains almost none of that. The two biggest issues in American healthcare are: - Zero price transparency - Payment through intermediaries like insurance companies and your employer, who often pays half of your insurance premiums and other healthcare expenditures With no price transparency, prices are just made up with little connection to actual value delivered. Because of the middlemen, most patients don’t feel the full cost of healthcare. Patients in aggregate still pay for the full negotiated cost, just indirectly in the form of higher premiums and forgone wages (marginal premiums paid by the employer). At this point, Medicare for All is our best hope. A true market-based healthcare economy, as great as it may be, is a pipe dream. With Medicare for All, the government would set reimbursement rates in stone. Most doctors who want a substantial amount of patients would have to accept those rates. Unlike insurance companies, the government would not be able to aggressively increase its “premiums,” or taxes, every year and it won’t get kickbacks from hospitals for high cost care. The government would also set empirically backed standards of care that lower their cost burden with no impact on average lifespan. We could potentially allow for parallel “luxury care” purchased by the patient if they want to be able to expedite their non-emergency knee replacement from 6 months to 2 weeks. The government would be incentivized to launch health initiatives and laws that help lower the rates of obesity and diabetes in the population. Lastly, the government would also be incentivized to increase the supply of doctors. There are multiple instances in the past when medical organizations lobbied or misled Congress into over-limiting the supply of medical residencies. They also had a hand in restricting the number and size of medical schools. So yes, overpaid, scarce medical professionals are a big part of the cost burden, and their representative institutions are making sure they stay overpaid. Only recently have medical organizations lobbied to increase the number of residencies again. That’s just damage control. The American people should not accept compromise on healthcare any longer. That’s what the ACA was, and it did not do enough. Medicare for All will bring us in line with the rest of the developed world which has better health outcomes at a much lower cost. We can save >$1 trillion and put that money towards things like a national high speed rail network, anti-poverty efforts, government-subsidized corporate internships, and an accredited national online education and retraining platform with the best professors on every subject imaginable.

hanniabu 7y ago

I'd think it's more of a problem to assume they should continue to rise with exponential gains....

PunchTornado 7y ago

to me this is a meaningless question. it's so broad with so many variables, so many assumptions that you might as well ask if God exists and you'd have the same chance of answering it.

Barrin92 7y ago

in semiconductors, it's a natural result of hard physical limits driving up costs for innovation. In the economy overall it's institutional neglect for deep innovation, basic R&D and instead allocation of capital to eCommerce, finance, retail or the gig economy, with ironically companies like facebook contributing to the misery by keeping engineering talent occupied with fiinding better ways to get people to click on ads.

chubot 7y ago

That reminds me of deep learning. Investment in computing resources is definitely increasing exponentially over time: https://openai.com/blog/ai-and-compute/ https://openai.com/blog/ai-and-compute/ https://techcrunch.com/2019/11/10/the-post-exponential-era-of-ai-and-moores-law/ https://techcrunch.com/2019/11/10/the-post-exponential-era-o... Tthere doesn't seem to by any evidence of exponential gains in performance, even on their own benchmarks (which importantly don't measure "intelligence", but something more narrow): https://blog.piekniewski.info/2019/11/18/late2019-the-wizards-of-oz/ https://blog.piekniewski.info/2019/11/18/late2019-the-wizard... Even though we are doubling the compute available to AI every 3.5 months, so a factor 10x per year, since say 2015, so some 10.000x, the classification performance as measured by top-1 accuracy on ImageNet had barely moved. Now arguably top-1 accuracy on ImageNet is probably not the best measure, but still that seems somewhat striking. The diminishing returns are visible elsewhere too, even that Open AI Rubic's cube is somewhat indicative. From an outside perspective, it seems that everyone is coming around to the fact that self-driving cars is harder than expected. There are no "exponential gains" in that field -- to the contrary, it seems like 90% of the problem was solved several years ago, and you can now take undergraduate courses that teach the basics. But going from 90%, to 99%, to 99.9%, etc. is taking decades, not months or years. There seems to be something fundamentally wrong with the current approach. I believe at the time the OpenAI blog post was supposed to be "impressive". Certainly I felt a bit of that -- "hey something important is going on there!" I believe that in 10 years it will look like a waste of effort. There will be a different, more efficient approach to solving these types of problems. ---- Peter Thiel is one of the only "major tech figures" speaking out against AI, which I appreciate. And the "AI is communist" meme is pretty memorable, and I don't think wholly inaccurate. I listen to him for his contrarian opinions. I just wish he didn't have these random "tribal conservative" outbursts. The most recent one was a video on here where he was rambling about Obama saying something far worse than George Bush and the Iraq war. I guess it was about sending their daughter to Harvard, but saying that lots of schools are just as good. He was trying to get some points from his conservative audience. I guess it gets attention, but that kind of tribalism undermines the rational points that he generally makes. (I realize the OP doesn't feature Thiel but he was brought up elsewhere in this thread, and is one of the people also lamenting the lack of technological progress despite investment, although maybe due to lack of investment in certain areas too)

nugget 7y ago

One purely linear constraint is time. In many health care fields, it simply takes time for clinical studies to collect data, from which the next generation of clinical studies can be planned, and there's no good way to meaningfully speed up that process.