Seems like the Fed’s goal is slowly but surely shifting into manipulating financial markets to keep the government funded as long, and at lowest rates possible. In this new paradigm, inflation is good, especially asset price inflation that pumps GDP but not pleb consumer goods that make their way into CPI. These deficits aren’t going anywhere!
Re: AI turning this all around, seems unlikely. AI seems like the classic bubble story atm where we need to see a crash/dis-enlightenment before reaching the “plateau of productivity” where these gains actually make it into the real economy and not just speculative gains. Also, if anything AI is going to massively benefit capital over labor, increasing the haves/nots dichotomy.
I have to agree that hopes for AI being the magical solution to improving productivity enough to matter at the economy wide level are, if not misplace, remarkably early. perhaps in 20 years, when companies have learned what its true benefits are and can incorporate those benefits into their processes effectively, at the margin things will be better. But it is classic bubble thinking that AI is going to enhance productivity significantly prior to the end of the decade. At least in my view. We have seen many technological breakthroughs in the past and it simply takes time for them to reach a point of significant impact.
I work in AI, and do research (but am weak on the finance side). And I'll add that: no one should be confident, because even the people at the forefront, when polled, have a range of projections. This show there is fundamental uncertainty, which we should take on board.
thanks for the links, very interesting and not surprising that there is a wide range of views. as I have said before, I guess the real question is how one defines "intelligence". is it the ability to regurgitate data in a slightly different format? or is the ability to consume data, understand the interactions and make new and meaningful progress on the subject. as an odd example, I watched the movie "Air" last night and was struck by the scene where Matt Damon (as Sonny Vaccarro) was watching the tape of Michael Jordan making the winning shot in the NCAA championship game as a freshman, and making the leap that Jordan was the first choice for the shot because he had the intangibles that create greatness, and that the coach and the rest of the team knew it as well. that leap, taking the same video that had been seen thousands of times and understanding the ultimate benefit is the type of intelligence that I believe remains many years in the future. winning at chess or go on a brute force calculation basis is not the same as seeing the possibilities of something based on your body of knowledge.
I suppose we should stick to a functional definition: the ability to replace jobs or perform useful work. That's the dream—having us all retire happily with friendly and aligned superintelligence.
The intelligence of GPT-3 and GPT-4 surprised even the experts who were doubtful, like you. Very few people predicted that mere scaling would bring us to this level—I certainly didn't (https://old.reddit.com/r/slatestarcodex/comments/12yu784/did_anyone_correctly_predict_ais_ability_to_code/). One of the few who did was Gwern, but he says that the labs are so locked down that he no longer has the information to make predictions. We're not quite there yet, but I see more uncertainty! The rational way to respond to uncertainty is probabilistically, e.g., by considering a range of options, at least in your investments.
"The Highway to BA" might have been missed on some, but rhymes fantastically! Agree that the transmission of lower rates into higher rents is the strongest argument against the 'high rates are inflationary' arguments that are emerging. Great piece as ever Florian.
Well, hang on a sec, how can you say this, “yet in a bizarre combination do not care about pumping the prices of assets held by only a very minor share of the population”?
Is it accurate to say that homes are only owned by a very small share of the population? True my 10-yr old doesn’t technically own our home, but it’s pretty much a family asset and I think that’s a fair description that applies to most Americans, or am I way off?
I’d assume the Fed is pretty concerned about home prices getting too high, as I think 2006 is fresh in everyone’s mind. Maybe they need to have more of a sense of urgency, but can’t this be a case where supply should take care of this before things get dangerous?
Or do you not trust that we’ll be able to build enough homes quickly enough to avoid a catastrophe?
Enjoyed the read, looking forward to reading more from you.
Seems like the Fed’s goal is slowly but surely shifting into manipulating financial markets to keep the government funded as long, and at lowest rates possible. In this new paradigm, inflation is good, especially asset price inflation that pumps GDP but not pleb consumer goods that make their way into CPI. These deficits aren’t going anywhere!
Re: AI turning this all around, seems unlikely. AI seems like the classic bubble story atm where we need to see a crash/dis-enlightenment before reaching the “plateau of productivity” where these gains actually make it into the real economy and not just speculative gains. Also, if anything AI is going to massively benefit capital over labor, increasing the haves/nots dichotomy.
I have to agree that hopes for AI being the magical solution to improving productivity enough to matter at the economy wide level are, if not misplace, remarkably early. perhaps in 20 years, when companies have learned what its true benefits are and can incorporate those benefits into their processes effectively, at the margin things will be better. But it is classic bubble thinking that AI is going to enhance productivity significantly prior to the end of the decade. At least in my view. We have seen many technological breakthroughs in the past and it simply takes time for them to reach a point of significant impact.
I work in AI, and do research (but am weak on the finance side). And I'll add that: no one should be confident, because even the people at the forefront, when polled, have a range of projections. This show there is fundamental uncertainty, which we should take on board.
But those projects are WIDE, they include "in the next 2 years" and "30 years away". Check out the links in this prediction markets to see surveys of experts https://www.metaculus.com/questions/3479/when-will-the-first-artificial-general-intelligence-system-be-devised-tested-and-publicly-known-of/
thanks for the links, very interesting and not surprising that there is a wide range of views. as I have said before, I guess the real question is how one defines "intelligence". is it the ability to regurgitate data in a slightly different format? or is the ability to consume data, understand the interactions and make new and meaningful progress on the subject. as an odd example, I watched the movie "Air" last night and was struck by the scene where Matt Damon (as Sonny Vaccarro) was watching the tape of Michael Jordan making the winning shot in the NCAA championship game as a freshman, and making the leap that Jordan was the first choice for the shot because he had the intangibles that create greatness, and that the coach and the rest of the team knew it as well. that leap, taking the same video that had been seen thousands of times and understanding the ultimate benefit is the type of intelligence that I believe remains many years in the future. winning at chess or go on a brute force calculation basis is not the same as seeing the possibilities of something based on your body of knowledge.
I suppose we should stick to a functional definition: the ability to replace jobs or perform useful work. That's the dream—having us all retire happily with friendly and aligned superintelligence.
The intelligence of GPT-3 and GPT-4 surprised even the experts who were doubtful, like you. Very few people predicted that mere scaling would bring us to this level—I certainly didn't (https://old.reddit.com/r/slatestarcodex/comments/12yu784/did_anyone_correctly_predict_ais_ability_to_code/). One of the few who did was Gwern, but he says that the labs are so locked down that he no longer has the information to make predictions. We're not quite there yet, but I see more uncertainty! The rational way to respond to uncertainty is probabilistically, e.g., by considering a range of options, at least in your investments.
I know one thing for sure: even if progress stalls, there are already published approaches that will allow us to apply GPT-4-level intelligence to robotics (https://eureka-research.github.io/, https://www.microsoft.com/en-us/research/group/autonomous-systems-group-robotics/articles/chatgpt-for-robotics/) or to add planning, which will increase capabilities. It's early days, but I reckon some of it will pan out. I try to learn from field-wide surprises by being appropriately unsure.
When my message is "uncertainty," I'm afraid no one likes it 😜. People want a crystal ball, not more options 😂. It's an unpopular take.
"The Highway to BA" might have been missed on some, but rhymes fantastically! Agree that the transmission of lower rates into higher rents is the strongest argument against the 'high rates are inflationary' arguments that are emerging. Great piece as ever Florian.
Buenos Aires surely a good relative bet now!
In fairness, given what Milei is trying to do down there, BA may be a better bet going forward!
Thanks, Florian! I truly enjoy reading your research. When is your next article coming?
Well, hang on a sec, how can you say this, “yet in a bizarre combination do not care about pumping the prices of assets held by only a very minor share of the population”?
Is it accurate to say that homes are only owned by a very small share of the population? True my 10-yr old doesn’t technically own our home, but it’s pretty much a family asset and I think that’s a fair description that applies to most Americans, or am I way off?
I’d assume the Fed is pretty concerned about home prices getting too high, as I think 2006 is fresh in everyone’s mind. Maybe they need to have more of a sense of urgency, but can’t this be a case where supply should take care of this before things get dangerous?
Or do you not trust that we’ll be able to build enough homes quickly enough to avoid a catastrophe?
Enjoyed the read, looking forward to reading more from you.
thank you Florian, have a great day!