Since MMT first became a shiny new thing, I’ve never seen how it was anything but Keynesianism at the zero lower bound for interest rates. I suppose the one novel feature was pretending that MMT would work when interest rates weren’t zero. Not at all surprised that the founder of MMT has realized that it doesn’t.
The biggest earthquakes signaling possible doom for Russia’s invasion of Eastern Europe were the failure of the Russia-supported RN to win outright power in France and the historic Conservative wipeout in the UK. Russia may want to sweep up Ukraine and gang up on Poland and the Baltics, but historic unity between Germany, France and UK to oppose Russia will ultimately prevent Putin from achieving his stated goals, even if Trump betrays US commitments to everyone (which we all expect, right?). In that scenario, the best outcome (for Putin) is he kills another million to settle for a cease fire and frozen war with the survival of most of Ukraine as a permanent western-looking, EU/NATO-seeking dependent. Russia becomes the next North Korea, a pariah with frequent nuclear weapons rattling to retain relevance, exploited by China and India and lesser contemptible “western” countries like Hungary for bargain rate raw resources.
China continues to dream of and threaten a Taiwan invasion while their CHICOM party bosses fret about the odds of less than total success, like Putin in Ukraine, and possible knock-on ignition of their own discontented and unemployed young people.
What a time to be alive and aware of a modicum of world history and current events. That excludes 85% of Fox viewers, BTW.
Problem with these sort of AI tweet stories is that there's no way to tell if the people reporting them are being honest. This guy has tweeted that he'd rather vote for a dead body than Trump so he's not exactly an unbiased source, the left just demonstrated a willingness to engage in a massive conspiracy in which they lied to each other/the public to create support for Biden, and they have a looooong history of crying wolf about bots on Twitter. For example he makes the assertion in his TikTok video that any username that ends in numbers is a bot, which is absolute nonsense. That's been the default username pattern Twitter gives people for years:
As there's no way to prove attribution it could easily be the guy himself. "Nobody real has any doubts about Biden, everyone who does is secretly a bot, look I proved it!"
On AI potential, I am still bullish. Profitable use cases are everywhere I look and even getting just one right can consume a lot of GPU capacity. Right now there's still a lot of work needed to get even the basic cases like support, summarization etc really reliable, but that's getting easier.
Good to see the MMT people get their comeuppance! About time!
I know Toby personally and I don't believe he'd fake something like that. He just lucked into finding a bot whose base model still had hooks like the "ignore all previous instructions" thing.
And as far as usernames like that being bots, I mean... fine maybe not all are, but a hell of a lot of those numbered accounts are fakes of one kind or another, because for folks who are spinning up accounts, it's so much easier to make them like that. Saying they _all_ are bots is an exaggeration, but it's a fact that _most_ real people who engage on the platform make an effort to either use a number that's significant to them personally (like a year), or find a string that avoids the need for the trailing number.
It's literally the default since 2017, the site doesn't even let you pick a username anymore unless you explicitly change it after signup. I think you overestimate how many people change defaults in software when there's no need. I work in software and the vast majority of people won't go into the settings screen even if they actually would benefit from changing a setting, and if you're the sort of person signing up for Twitter only in the last few years, you definitely aren't the sort to care about a specific username.
As for knowing him, so what? The people the left lie to the most are their friends and allies. It wasn't Fox News viewers who got caught by surprise during the presidential debate.
Under-discussed, Texas being a high property tax / zero income tax state helped Austin respond to the COVID demand surge. Property tax is typically 2-3% per year, so homeowners in Austin with a $400k house were already paying 600-700/mo in property tax pre-COVID. When home prices doubled (or more than doubled) in 2020-2021, they were followed by giant new tax bills, and when people have to pay an extra 600-700/mo in property tax they way less excited about "appreciation."
I've been in city council and community meetings and listened to people in Texas try to stop public amenities being built near them because "it'll be nice but I can't afford the higher taxes." Since people have more of a stake in price *stability* they're easier to persuade to support more supply.
Which helps explain California NIMBYs. Prop 13 still reigns and so appreciation is fine for current residents who will see no property tax appreciation as a result of appreciating real estate prices.
I agree with the Poland analysis. Personally, I don't think Russia has much of a chance of actually overrunning Ukraine, and I don't see any sign that a rapid political or military collapse is likely (anything's possible, of course). That said, I haven't really been following the war news.
That is to say, I don't think Russia has any hope of defeating Poland under non-nuclear circumstances. Equally, I don't think it would be wise or necessary for Poland to jump in conventionally in the foreseeable future.
But as to the critical question of NATO following Poland into the war (obviously short of an attack by Russia on Poland), I'm doubtful. As I understand the North Atlantic Treaty (which is to say, not well), it is a defensive alliance, and it seems likely that the allies would use the arguably offensive nature of a Polish intervention as described in the piece to punt.
Of course that's true. Noah isn't saying that NATO allies will be forced by their treaties to enter the war, he's saying that if Poland actually jumps in and starts fighting, other NATO countries might decide to jump in as well.
Even if they manage to overrun Ukraine somehow, they will just be stuck in a permanent low-level conflict as Ukraines neighbors will happily flood the country with cheap weapons for conducting guerrilla warfare. I guess it’s hard for someone who watched the Iraq and Afghanistan situations to believe that this doesn’t turn into a permanent quagmire. Also is there any way Russia manages to form a million-man army of Ukrainians that doesn’t immediately start fragging the ever-loving crap out of the Russian officer corps?
The AI bubble commentary seems hopelessly confused to me. There may be an investment bubble, but those are pretty hard to conclusively detect until afterwards. In the meantime, Nvidia's price is high but not extraordinarily so given their earnings growth.
Regarding costs of using gen AI, those are going down 2-10X per year right now, so affordability will largely take care of itself. Of course, that does make it harder to generate the same amount of revenue.
I work at a tech consulting company, and I'm seeing lots of enterprise use cases. Every workflow that isn't terribly complex is getting automated, although with humans in the loop so far - loan applications, purchases, healthcare admissions and billing, etc. For software developers, using an AI copilot is becoming pretty standard. Every software vendor is integrating AI. And AI models are rapidly getting better. Companies like Microsoft, Google, and Meta aren't startups hoping to goose their valuations; they are making big bets because they have considered strategies that predict the technology will be transformative. They could be wrong, but I don't think industry critics have made a compelling case that they are.
Goldman also quotes Acemoglu, who as we've discussed has some notable biases against AI. He thinks it will lead to much higher inequality without meaningfully improving productivity, which is a very specific and ideologically convenient point on the spectrum of usefulness.
I'm glad there are still a good number of AI critics out there (otherwise it would definitely be a bubble), but I'd like them to make better reasoned arguments. In the meantime, I think gen AI is a new kind of universal connector that allows software to easily work with natural language, human concepts, images, and other areas that previously were extremely hard to work with, so it does seem like it will be transformative regardless of what happens with valuations in the next couple of years.
It’s unclear whether the dominant use cases for generative AI will be via a “universal assistant” like ChatGPT or as an enabling technology for apps that solve specific problems. Most of the critique of generative AI’s revenue potential assumes the universal assistant model. But, if LLMs are more of an enabling technology, it will take longer for killer apps to emerge. Ben Thompson often points out at Stratechery that the algorithmic feed was critical to growth of media/search/social networking companies, but it took years for them to figure out that it was the right approach.
I would be curious if the more housing => lower rents (quickly) pattern holds outside of sunbelt cities immediately post Covid. These seem like they’d be the places that would have gotten the biggest rent runup from coastal professionals moving in during the pandemic. Therefore as the pandemic recedes they might have had falling rents anyway.
Property developers and real estate investors are likely locked into high-interest debt that could be justified had sunbelt areas been growing before the Fed started hiking interest rates. But the roaring growth rates have slowed down throughout the U.S. The largest metros are losing population, and the fastest-growing sunbelt metros are seeing growth in the single-digits when they had been accustomed to double-digit growth since the 1980s onward.
Real estate values for both homeowners and renters are historically high, and the high prices coupled with higher interest costs to finance it all are making homeowners and tenants alike wary of taking on the risk of a big move.
Apartment buildings along major, commercial boulevards will increase pedestrian traffic and create vibrant neighborhoods and businesses: restaurants, cafés, bars, markets -- maybe even music venues. Just need to keep building underground metro routes, like Los Angeles is.
A.I. needs to work primarily on understanding prompts. Siri and Alexa are still very stupid (why are they behind the rest of the industry, instead of ahead of it? They have way more money!). Still and Video image generators are not there yet.
But this program will be amazing if it figures it out. It's for generating storyboards for movies from screenplays. It can also generate screenplays. And eventually, no doubt, it will be able to "animate" screenplays as well. https://storyboarder.ai
Am I reading the chart under point #5 correctly - US inequality as measured by the Gini coefficient hasn't changed since the mid-90s? This will be useful the next time (yet another) person complains about increasing inequality in America and how the "fat cats" are getting rich at the expense of everyone else.
Speaking about the concern of Mosler, I think that Argentina under Milei is actually trying to reduce inflation in a similar way: its Central Bank has reduced the policy interest rate from 133% to 40% since he came to office, and even Milei himself has talked a lot about reducing interest rate to "repay all monetary liabilities". Ivan Werning actually mentioned the policy in a tweet, and saying that in Argentine case, it could be a way to reduce their inflation: https://x.com/IvanWerning/status/1786076862327701566.
However, this has also led to quite a lot of ridicule from anti-Milei commentators in Argentina, with this tweet as an example (though in Spanish): https://x.com/SergioChouza/status/1786114736339185943. So with that in mind, could you write a post in the future about whether reducing interest rate to fight inflation could actually work, and what case could this happen, given that Erdogan in Turkey basically burned the central bank' credibility and raised inflation to roughly 100% annually when doing it?
There is too AI grief.
Arcade Fire's "Deep Blue" is about AI grief
Since MMT first became a shiny new thing, I’ve never seen how it was anything but Keynesianism at the zero lower bound for interest rates. I suppose the one novel feature was pretending that MMT would work when interest rates weren’t zero. Not at all surprised that the founder of MMT has realized that it doesn’t.
The biggest earthquakes signaling possible doom for Russia’s invasion of Eastern Europe were the failure of the Russia-supported RN to win outright power in France and the historic Conservative wipeout in the UK. Russia may want to sweep up Ukraine and gang up on Poland and the Baltics, but historic unity between Germany, France and UK to oppose Russia will ultimately prevent Putin from achieving his stated goals, even if Trump betrays US commitments to everyone (which we all expect, right?). In that scenario, the best outcome (for Putin) is he kills another million to settle for a cease fire and frozen war with the survival of most of Ukraine as a permanent western-looking, EU/NATO-seeking dependent. Russia becomes the next North Korea, a pariah with frequent nuclear weapons rattling to retain relevance, exploited by China and India and lesser contemptible “western” countries like Hungary for bargain rate raw resources.
China continues to dream of and threaten a Taiwan invasion while their CHICOM party bosses fret about the odds of less than total success, like Putin in Ukraine, and possible knock-on ignition of their own discontented and unemployed young people.
What a time to be alive and aware of a modicum of world history and current events. That excludes 85% of Fox viewers, BTW.
The largest party alliance In the French Parliament is led by a party anti-Ukraine, anti-EU, anti-NATO, antisemitic and pro Putin.
Problem with these sort of AI tweet stories is that there's no way to tell if the people reporting them are being honest. This guy has tweeted that he'd rather vote for a dead body than Trump so he's not exactly an unbiased source, the left just demonstrated a willingness to engage in a massive conspiracy in which they lied to each other/the public to create support for Biden, and they have a looooong history of crying wolf about bots on Twitter. For example he makes the assertion in his TikTok video that any username that ends in numbers is a bot, which is absolute nonsense. That's been the default username pattern Twitter gives people for years:
https://tinysubversions.com/notes/twitter-usernames/
As there's no way to prove attribution it could easily be the guy himself. "Nobody real has any doubts about Biden, everyone who does is secretly a bot, look I proved it!"
On AI potential, I am still bullish. Profitable use cases are everywhere I look and even getting just one right can consume a lot of GPU capacity. Right now there's still a lot of work needed to get even the basic cases like support, summarization etc really reliable, but that's getting easier.
Good to see the MMT people get their comeuppance! About time!
I know Toby personally and I don't believe he'd fake something like that. He just lucked into finding a bot whose base model still had hooks like the "ignore all previous instructions" thing.
And as far as usernames like that being bots, I mean... fine maybe not all are, but a hell of a lot of those numbered accounts are fakes of one kind or another, because for folks who are spinning up accounts, it's so much easier to make them like that. Saying they _all_ are bots is an exaggeration, but it's a fact that _most_ real people who engage on the platform make an effort to either use a number that's significant to them personally (like a year), or find a string that avoids the need for the trailing number.
It's literally the default since 2017, the site doesn't even let you pick a username anymore unless you explicitly change it after signup. I think you overestimate how many people change defaults in software when there's no need. I work in software and the vast majority of people won't go into the settings screen even if they actually would benefit from changing a setting, and if you're the sort of person signing up for Twitter only in the last few years, you definitely aren't the sort to care about a specific username.
As for knowing him, so what? The people the left lie to the most are their friends and allies. It wasn't Fox News viewers who got caught by surprise during the presidential debate.
First :P
You win!! :D
Under-discussed, Texas being a high property tax / zero income tax state helped Austin respond to the COVID demand surge. Property tax is typically 2-3% per year, so homeowners in Austin with a $400k house were already paying 600-700/mo in property tax pre-COVID. When home prices doubled (or more than doubled) in 2020-2021, they were followed by giant new tax bills, and when people have to pay an extra 600-700/mo in property tax they way less excited about "appreciation."
I've been in city council and community meetings and listened to people in Texas try to stop public amenities being built near them because "it'll be nice but I can't afford the higher taxes." Since people have more of a stake in price *stability* they're easier to persuade to support more supply.
If budgets stay the same, there is no reason monthly taxes paid should go up just because values go up (ie the tax rate should fall)
The insanity is local governments raising spending in line with property bubbles.
Which helps explain California NIMBYs. Prop 13 still reigns and so appreciation is fine for current residents who will see no property tax appreciation as a result of appreciating real estate prices.
I agree with the Poland analysis. Personally, I don't think Russia has much of a chance of actually overrunning Ukraine, and I don't see any sign that a rapid political or military collapse is likely (anything's possible, of course). That said, I haven't really been following the war news.
That is to say, I don't think Russia has any hope of defeating Poland under non-nuclear circumstances. Equally, I don't think it would be wise or necessary for Poland to jump in conventionally in the foreseeable future.
But as to the critical question of NATO following Poland into the war (obviously short of an attack by Russia on Poland), I'm doubtful. As I understand the North Atlantic Treaty (which is to say, not well), it is a defensive alliance, and it seems likely that the allies would use the arguably offensive nature of a Polish intervention as described in the piece to punt.
Of course that's true. Noah isn't saying that NATO allies will be forced by their treaties to enter the war, he's saying that if Poland actually jumps in and starts fighting, other NATO countries might decide to jump in as well.
And Trump likes Poland, because they’ve always contributed their fair share
Even if they manage to overrun Ukraine somehow, they will just be stuck in a permanent low-level conflict as Ukraines neighbors will happily flood the country with cheap weapons for conducting guerrilla warfare. I guess it’s hard for someone who watched the Iraq and Afghanistan situations to believe that this doesn’t turn into a permanent quagmire. Also is there any way Russia manages to form a million-man army of Ukrainians that doesn’t immediately start fragging the ever-loving crap out of the Russian officer corps?
The AI bubble commentary seems hopelessly confused to me. There may be an investment bubble, but those are pretty hard to conclusively detect until afterwards. In the meantime, Nvidia's price is high but not extraordinarily so given their earnings growth.
Regarding costs of using gen AI, those are going down 2-10X per year right now, so affordability will largely take care of itself. Of course, that does make it harder to generate the same amount of revenue.
I work at a tech consulting company, and I'm seeing lots of enterprise use cases. Every workflow that isn't terribly complex is getting automated, although with humans in the loop so far - loan applications, purchases, healthcare admissions and billing, etc. For software developers, using an AI copilot is becoming pretty standard. Every software vendor is integrating AI. And AI models are rapidly getting better. Companies like Microsoft, Google, and Meta aren't startups hoping to goose their valuations; they are making big bets because they have considered strategies that predict the technology will be transformative. They could be wrong, but I don't think industry critics have made a compelling case that they are.
Goldman also quotes Acemoglu, who as we've discussed has some notable biases against AI. He thinks it will lead to much higher inequality without meaningfully improving productivity, which is a very specific and ideologically convenient point on the spectrum of usefulness.
I'm glad there are still a good number of AI critics out there (otherwise it would definitely be a bubble), but I'd like them to make better reasoned arguments. In the meantime, I think gen AI is a new kind of universal connector that allows software to easily work with natural language, human concepts, images, and other areas that previously were extremely hard to work with, so it does seem like it will be transformative regardless of what happens with valuations in the next couple of years.
It’s unclear whether the dominant use cases for generative AI will be via a “universal assistant” like ChatGPT or as an enabling technology for apps that solve specific problems. Most of the critique of generative AI’s revenue potential assumes the universal assistant model. But, if LLMs are more of an enabling technology, it will take longer for killer apps to emerge. Ben Thompson often points out at Stratechery that the algorithmic feed was critical to growth of media/search/social networking companies, but it took years for them to figure out that it was the right approach.
So much bearish talk about AI acts like AI is as good as its going to get.
Build more housing!
I would be curious if the more housing => lower rents (quickly) pattern holds outside of sunbelt cities immediately post Covid. These seem like they’d be the places that would have gotten the biggest rent runup from coastal professionals moving in during the pandemic. Therefore as the pandemic recedes they might have had falling rents anyway.
It appears that Minneapolis and Seattle also have lower or stabilizing rents after allowing more residential construction, but not as much as Austin.
They built a lot of units in 2021-2022 and now things don’t look as great because, like many a pandemic bubble, the growth boom to the sunbelt has slowed and investors who built - and bought - apartment buildings are starting to feel some pain. Or so says the New York Times. https://www.nytimes.com/2024/07/04/business/apartment-multifamily-loans-trouble.html?unlocked_article_code=1.6E0.T6Vy.VcfSI5yRQrhY&smid=url-share
The pain might be on the business side.
Property developers and real estate investors are likely locked into high-interest debt that could be justified had sunbelt areas been growing before the Fed started hiking interest rates. But the roaring growth rates have slowed down throughout the U.S. The largest metros are losing population, and the fastest-growing sunbelt metros are seeing growth in the single-digits when they had been accustomed to double-digit growth since the 1980s onward.
Real estate values for both homeowners and renters are historically high, and the high prices coupled with higher interest costs to finance it all are making homeowners and tenants alike wary of taking on the risk of a big move.
Apartment buildings along major, commercial boulevards will increase pedestrian traffic and create vibrant neighborhoods and businesses: restaurants, cafés, bars, markets -- maybe even music venues. Just need to keep building underground metro routes, like Los Angeles is.
Can't wait for the Chinese Slacker Generation to blow my mind.
A.I. needs to work primarily on understanding prompts. Siri and Alexa are still very stupid (why are they behind the rest of the industry, instead of ahead of it? They have way more money!). Still and Video image generators are not there yet.
But this program will be amazing if it figures it out. It's for generating storyboards for movies from screenplays. It can also generate screenplays. And eventually, no doubt, it will be able to "animate" screenplays as well. https://storyboarder.ai
Am I reading the chart under point #5 correctly - US inequality as measured by the Gini coefficient hasn't changed since the mid-90s? This will be useful the next time (yet another) person complains about increasing inequality in America and how the "fat cats" are getting rich at the expense of everyone else.
Speaking about the concern of Mosler, I think that Argentina under Milei is actually trying to reduce inflation in a similar way: its Central Bank has reduced the policy interest rate from 133% to 40% since he came to office, and even Milei himself has talked a lot about reducing interest rate to "repay all monetary liabilities". Ivan Werning actually mentioned the policy in a tweet, and saying that in Argentine case, it could be a way to reduce their inflation: https://x.com/IvanWerning/status/1786076862327701566.
However, this has also led to quite a lot of ridicule from anti-Milei commentators in Argentina, with this tweet as an example (though in Spanish): https://x.com/SergioChouza/status/1786114736339185943. So with that in mind, could you write a post in the future about whether reducing interest rate to fight inflation could actually work, and what case could this happen, given that Erdogan in Turkey basically burned the central bank' credibility and raised inflation to roughly 100% annually when doing it?