In my commentary this week, I look at the insightful research note from Goldman Sachs, titled “The Potentially Large Effect of Artificial Intelligence on Economic Growth”, in which the authors explore the labour market's future.
Healthcare is becoming a dual care model of Clinical Care (restoring health) and Lifestyle Care (preserving health). Lifestyle Care includes programs like the National Diabetes Prevention Program (DPP), Medicare DPP, and social determinants of health programs with a Lifestyle Coach workforce delivering the lifestyle change programs in-person at convenient locations in their communities. This will expand into other areas of lifestyle related chronic disease prevention and management programs which is responsible for 85% of the health care costs nationally and these programs are now being reimbursed by healthcare payers for sustainable delivery. Mississippi is the first state to launch a Lifestyle Care ecosystem establishing a Lifestyle Coach workforce statewide.
Interesting. So the new roles emerge in the preventative space. And it might be the case that the level of formal professional training and certification that is required is lower than being a doctor who prescribes to, cuts into and sews up patients....
Exactly right, this is about delivering programs for lifestyle change where access to the program through the Lifestyle Coach is most important for participant enrollment. For example, the National Diabetes Prevention Program is by a certified Lifestyle Coach for in-person delivery of this year long program.
This reminds me of the 70s when we were told not to fear automation because somebody has to fix the machines. Even my husband, an 8th grade dropout, could see through the logic of that one.
So I realize with all my chatting that I did not answer your fundamental question: what can be put in place to prevent even more devastating social consequences?
De-internationalize the economy.
Have every locale produce much of its own food, energy and healthcare. Create standalone systems independent of the international flow of money. Forbid outsiders from owning local land. Reverse the financialization of human living space.
Large Language Models require some getting used to in how to use LLM to generate accurate and useable output. The human users will need experience and continuous learning to engage productively with LLM. Consequently, there will will a period of exploration, testing, and sorting out before the full impact of LLM is realized in a continuous growth capitalist model of business. Meanwhile, the increasing violent impacts of reaching and perhaps exceeding 1.5C in global warming in 2024 (as El Niño births itself) will work against the known playbook of dense energy (oil, coal, etc.) capitalism that is dependent on continuous growth. Past paradigms of business will be challenged and this could be a significant wrench in the works for LLMs that draw conclusions and make business projections from historical data.
LLM is a difficult technology with many ways to go sideways. It’s my personal opinion that a globalized economy is no longer viable for the times coming upon us. LLM is a product of complex systems only made possible by unsustainable use of dense energy (oil, coal. Etc). It’s not a paradigm for working out survival of local human populations in the rapidly devolving infrastructure of our current unknowable complexity. Survival over this century and beyond (maybe thousands of years) will require small-scale structures of human interaction and economy as the planet goes into crazy times of climatic instability. Other people have the few that we can only solve the downsides of dense-energy complexity with even more complexity assisted by artificial intelligence. The dense energy side of the argument holds the levers of power and media propaganda. Nature will be left to force people into modes of reduced complexity and localized survival.
Hi everyone. This is a fascinating topic that my team and I have been bouncing around for a couple weeks now. Whenever I consider specific interventions or programs that could work at the scale that we might need to upskill/reskill or to compensate for potential job losses in the short and medium term I tend to look at it in two ways.
First, and Azeem you alluded to this - the productivity gains caused by the influence of generative AI will certainly accelerate “production”. Tasks will be completed much faster than they would have otherwise and “work outputs” will be ready earlier than they would otherwise. At a project and an organizational scale, it will take some time for this new mode of production to stabilize across all functions of the business. When it does stabilize, that entire project, organization or potentially even nation will now be producing outputs much faster than they would have otherwise. This dynamic will accelerate the creation of new opportunities and new jobs, but it will also cause an increase in the inflow (or “demand”) for supporting services and materials to the given nation, organization, project or individual on the ‘supply’ side. For example, at the individual level, I can create great content now very quickly, so I’m going to take on more side projects of creating more courses, writing books, creating art, building apps, etc. Therefore I will consume more subscriptions, licenses, tools, and services as I set about these new projects. Therefore, workforce could be stable, or potentially more work will be undertaken than would otherwise.
The second thing I think about is less about the process of production and more about the objectives or risks that exist in our economic and social environment that actually *necessitate* this acceleration discussed in my first point. I think we can all agree that there is no shortage of big, hairy, audacious problems out there that are quite literally a matter of human survival. I think this is where Azeem’s point about public sector leadership should come into play. Particularly in the context of Gen AI as “the great equalizer”, there should be plenty of people that could, in theory, make important contributions toward solving these big problems rather than, say, new shows about cats or the latest and most advanced first person shooter. Government programs like tax credits, or even basic income programs could incentivize people to work on contributing to solving the big problems (in collaboration, of course, with our new robot assistants). This could also be further enabled by web3 tech like metaverse and blockchain/DAO technologies. Within my team here in Canada we’ve recently taken to the idea of FOW = AI + Web3 + ESG. (there’s probably a better mathematical representation of that, beyond my humble arts degree😊)
In sum, I am actually of the abundance mindset and I look ahead to an sharply accelerated future with many opportunities to create breakthroughs to some of humanity’s toughest problems. Would love to hear what others think of these ideas as well, this is all very recent thinking as things are moving so fast!
1) Why should AI boost productivity given that digitization hasn‘t led to the Productivity gains we have hoped in the past? De facto we see a Productivity decline in OECD member states. So what‘‘s different with GAI? Seeing the slow adoption rotes auf advanced technologies in the EU in particular amongst SMEs there is a lot to be done to make sure not only large enterprises boost their competitiveness. https://www.oecd.org/economy/growth/digitalisation-productivity-and-inclusiveness/
2) Given the petrified education system in Europe, what skills does the current workforce as well as the next gen of workers actually need? My takeaways from having been involved in co-creation of educational digital skills programs for the past 20 years: It’s much less technical, digital literacy itself, but more tied to „learning“ skills - critical thinking, problem solving, creativity and innovation etc. And it starts in kindergarden. So what’s the best way to teach kids, students, workers and educators themselves to develop the 21st century skills mindset and capacity! Let them solve real world problems with new easy accessible tools such as GAI apps. It is nothing we can solve with upskilling programs alone. We need a whole new attitude to approach what is nicely framed as „21st century skills“ - at least in Europe. https://en.m.wikipedia.org/wiki/21st_century_skills
Goldman Sachs? You mean the vampire squid on the face of the US economy? These thoughts on labor haven't included labor at all.
Growth in productivity hasn’t meant shit for the American worker. There was a similar productivity shift in the Nineties. And what did we hear? “Learn to code!” How did that work out? Look out the window: decimated towns, rampant drug addiction, hollowed-out manufacturing base, collapsing families.
Any economic growth related to jobs being replaced by AI would be only a temporary bump. You know, like the ‘90s. Remember that rise in productivity? Has it prevented our current economic doldrums? No.
“Raise demand in complementary services.” Maybe there is a demand for a new type of job - but if the overall wage cost is the same as before the AI implementation, there’d be no economic growth.
The AI model spurs growth by reducing overall wage costs, not just for an individual firm. So there are going to be “losers” as you so nicely put it. It is going to be people who can’t feed their families, who turn to drugs as a solace or who kill their families. Loser, indeed.
The Goldman Sachs paper seems to be here:
https://d110erj175o600.cloudfront.net/wp-content/uploads/2023/03/29112641/Global-Economics-Analyst_-The-Potentially-Large-Effects-of0D0A-Artificial-Intelligence-on-Economic-Growth-Briggs_Kodnani.pdf
Healthcare is becoming a dual care model of Clinical Care (restoring health) and Lifestyle Care (preserving health). Lifestyle Care includes programs like the National Diabetes Prevention Program (DPP), Medicare DPP, and social determinants of health programs with a Lifestyle Coach workforce delivering the lifestyle change programs in-person at convenient locations in their communities. This will expand into other areas of lifestyle related chronic disease prevention and management programs which is responsible for 85% of the health care costs nationally and these programs are now being reimbursed by healthcare payers for sustainable delivery. Mississippi is the first state to launch a Lifestyle Care ecosystem establishing a Lifestyle Coach workforce statewide.
Interesting. So the new roles emerge in the preventative space. And it might be the case that the level of formal professional training and certification that is required is lower than being a doctor who prescribes to, cuts into and sews up patients....
Exactly right, this is about delivering programs for lifestyle change where access to the program through the Lifestyle Coach is most important for participant enrollment. For example, the National Diabetes Prevention Program is by a certified Lifestyle Coach for in-person delivery of this year long program.
One of the Apps I use already has a Lifestyle Coach function. I just use the daily reflection part of it. That is already using an small "AI" model,
This reminds me of the 70s when we were told not to fear automation because somebody has to fix the machines. Even my husband, an 8th grade dropout, could see through the logic of that one.
separately you might find this interesting https://twitter.com/jburnmurdoch/status/1641799667942866944?s=46&t=x7VrGz6qT5yK5HkIaWJssA
The Lifestyle Coach is the perfect example of AI as a co-pilot to help guide their cohorts through the healthy lifestyle change process.
You're funny.
UGH. Depressing but true.
The reason we need "dual care" is because people are killing themselves with poor diet and unhealthy lifestyles. Why?
So I realize with all my chatting that I did not answer your fundamental question: what can be put in place to prevent even more devastating social consequences?
De-internationalize the economy.
Have every locale produce much of its own food, energy and healthcare. Create standalone systems independent of the international flow of money. Forbid outsiders from owning local land. Reverse the financialization of human living space.
Large Language Models require some getting used to in how to use LLM to generate accurate and useable output. The human users will need experience and continuous learning to engage productively with LLM. Consequently, there will will a period of exploration, testing, and sorting out before the full impact of LLM is realized in a continuous growth capitalist model of business. Meanwhile, the increasing violent impacts of reaching and perhaps exceeding 1.5C in global warming in 2024 (as El Niño births itself) will work against the known playbook of dense energy (oil, coal, etc.) capitalism that is dependent on continuous growth. Past paradigms of business will be challenged and this could be a significant wrench in the works for LLMs that draw conclusions and make business projections from historical data.
Ok, Ed. I'm sure you have enough experience with the internet to realize the trolls are going to hijack this LLM learning.
LLM is a difficult technology with many ways to go sideways. It’s my personal opinion that a globalized economy is no longer viable for the times coming upon us. LLM is a product of complex systems only made possible by unsustainable use of dense energy (oil, coal. Etc). It’s not a paradigm for working out survival of local human populations in the rapidly devolving infrastructure of our current unknowable complexity. Survival over this century and beyond (maybe thousands of years) will require small-scale structures of human interaction and economy as the planet goes into crazy times of climatic instability. Other people have the few that we can only solve the downsides of dense-energy complexity with even more complexity assisted by artificial intelligence. The dense energy side of the argument holds the levers of power and media propaganda. Nature will be left to force people into modes of reduced complexity and localized survival.
Good points.
Alpaca the new Stanford model that piggybacks of Meta's LLaMa 7b costs $600 to train allegedly on consumer hardware.
The path that requires "nuanced public policy" is very unlikely to be available.
Would you consider Luxembourg's Digital Skills Bridge to be unnuanced? https://digital-luxembourg.public.lu/initiatives/luxembourg-digital-skills-bridge
Hi everyone. This is a fascinating topic that my team and I have been bouncing around for a couple weeks now. Whenever I consider specific interventions or programs that could work at the scale that we might need to upskill/reskill or to compensate for potential job losses in the short and medium term I tend to look at it in two ways.
First, and Azeem you alluded to this - the productivity gains caused by the influence of generative AI will certainly accelerate “production”. Tasks will be completed much faster than they would have otherwise and “work outputs” will be ready earlier than they would otherwise. At a project and an organizational scale, it will take some time for this new mode of production to stabilize across all functions of the business. When it does stabilize, that entire project, organization or potentially even nation will now be producing outputs much faster than they would have otherwise. This dynamic will accelerate the creation of new opportunities and new jobs, but it will also cause an increase in the inflow (or “demand”) for supporting services and materials to the given nation, organization, project or individual on the ‘supply’ side. For example, at the individual level, I can create great content now very quickly, so I’m going to take on more side projects of creating more courses, writing books, creating art, building apps, etc. Therefore I will consume more subscriptions, licenses, tools, and services as I set about these new projects. Therefore, workforce could be stable, or potentially more work will be undertaken than would otherwise.
The second thing I think about is less about the process of production and more about the objectives or risks that exist in our economic and social environment that actually *necessitate* this acceleration discussed in my first point. I think we can all agree that there is no shortage of big, hairy, audacious problems out there that are quite literally a matter of human survival. I think this is where Azeem’s point about public sector leadership should come into play. Particularly in the context of Gen AI as “the great equalizer”, there should be plenty of people that could, in theory, make important contributions toward solving these big problems rather than, say, new shows about cats or the latest and most advanced first person shooter. Government programs like tax credits, or even basic income programs could incentivize people to work on contributing to solving the big problems (in collaboration, of course, with our new robot assistants). This could also be further enabled by web3 tech like metaverse and blockchain/DAO technologies. Within my team here in Canada we’ve recently taken to the idea of FOW = AI + Web3 + ESG. (there’s probably a better mathematical representation of that, beyond my humble arts degree😊)
In sum, I am actually of the abundance mindset and I look ahead to an sharply accelerated future with many opportunities to create breakthroughs to some of humanity’s toughest problems. Would love to hear what others think of these ideas as well, this is all very recent thinking as things are moving so fast!
A few questions:
1) Why should AI boost productivity given that digitization hasn‘t led to the Productivity gains we have hoped in the past? De facto we see a Productivity decline in OECD member states. So what‘‘s different with GAI? Seeing the slow adoption rotes auf advanced technologies in the EU in particular amongst SMEs there is a lot to be done to make sure not only large enterprises boost their competitiveness. https://www.oecd.org/economy/growth/digitalisation-productivity-and-inclusiveness/
2) Given the petrified education system in Europe, what skills does the current workforce as well as the next gen of workers actually need? My takeaways from having been involved in co-creation of educational digital skills programs for the past 20 years: It’s much less technical, digital literacy itself, but more tied to „learning“ skills - critical thinking, problem solving, creativity and innovation etc. And it starts in kindergarden. So what’s the best way to teach kids, students, workers and educators themselves to develop the 21st century skills mindset and capacity! Let them solve real world problems with new easy accessible tools such as GAI apps. It is nothing we can solve with upskilling programs alone. We need a whole new attitude to approach what is nicely framed as „21st century skills“ - at least in Europe. https://en.m.wikipedia.org/wiki/21st_century_skills
Goldman Sachs? You mean the vampire squid on the face of the US economy? These thoughts on labor haven't included labor at all.
Growth in productivity hasn’t meant shit for the American worker. There was a similar productivity shift in the Nineties. And what did we hear? “Learn to code!” How did that work out? Look out the window: decimated towns, rampant drug addiction, hollowed-out manufacturing base, collapsing families.
Any economic growth related to jobs being replaced by AI would be only a temporary bump. You know, like the ‘90s. Remember that rise in productivity? Has it prevented our current economic doldrums? No.
“Raise demand in complementary services.” Maybe there is a demand for a new type of job - but if the overall wage cost is the same as before the AI implementation, there’d be no economic growth.
The AI model spurs growth by reducing overall wage costs, not just for an individual firm. So there are going to be “losers” as you so nicely put it. It is going to be people who can’t feed their families, who turn to drugs as a solace or who kill their families. Loser, indeed.