I don't think Silver understands risk. He has a romanticized view limited to those who have a choice as to whether to accept certain risks or not. Failure for venture capitalists means not making the big score, not losing everything. There's a reason venture capitalists tend to come from relatively wealthy families. Compulsive gamblers may risk losing everything, but professional gamblers are more like venture capitalists. It may be a negative sum business, but it's still a business.
A woman having sex with a guy she met at a bar takes a much bigger risk than any of these. Ditto for a fireman or peace officer responding to a 911 call. Immigrants to the US took on a broad variety of risks, and despite overcoming many of them are now being clobbered by a long shot outcome. These risks are much harder to romanticize. They are definitely not flowing in Silver's river.
One more note on how I wish we had the early-days Nate Silver of <http://fivethirtyeight.com> back, and how the story of the evolution here is rather tragic. Silver understood poll aggregation. But now he writes about things he does not understand. Kaleberg had a nice comment on this:
> **Kaleberg**: 'I don't think Silver understands [real] risk. He has a romanticized view limited to those who have a choice as to whether to accept certain risks or not. Failure for venture capitalists means not making the big score, not losing everything...
And let me chime in and add to this:
There is a very very very very very important distinction between people who are working with their counterparties, and people who are working against them. Or, indeed—going further into the psychopath realm—people who do not think that their counterparties are real people of any concern to them at all.
Case in point: Silver's extraordinary and persistent rage against Congressman Jim Leach.
IMHO Jim Leach did very few good things with his time in the U.S. Congress. But one of those good things was placing roadblocks to try protect people who did not understand the risks they were running from so easily becoming marks grifted by the likes of Nate Silver.
And Silver's rage against Jim Leach still persists. It looks as though Jim Leach will live rent-free in Nate Silver's brain until the day Silver dies, if not after:
> **Nate Silver**: On the Edge <>: 'I’d been a professional poker player for three years between 2004 and 2007, during the so-called Poker Boom. The Poker Boom began
because of the increasing availability of online poker and because of Chris Moneymaker, an accountant from Nashville who won an online qualifying tournament for a seat at the $10,000 Main Event at the 2003 World Series of Poker and then parlayed that into winning the Main Event for $2.5 million.
> If you’d asked ChatGPT to design a person who would most increase the amount of interest in poker by winning the WSOP, it might have spat out Moneymaker. An affable, pudgy, late twenty-something dude with a boring corporate job, he was exactly the customer the online poker sites were targeting, an archetype for every office drone who wanted to break out of his cubicle and win the big jackpot.
> The number of participants in the World Series of Poker Main Event exploded from 839 in Moneymaker’s 2003 to 8,773 just three years later in 2006, largely fueled by people who had won their seats online. I was one of those people who lived the dream. I soon found myself on a nocturnal schedule. Poker games are usually best late at night, when your opponents are some combination of drunk, sleep-deprived, or delirious from winning or losing a bunch of money. So I’d come home from my cubicle,
take a nap, and then play poker online, sometimes straight through until the morning, when I’d straggle into work and struggle through the day.
> Needless to say, this wasn’t sustainable, and—making considerably more money as a poker player than as a consultant—I quit my corporate job within about six months to play poker and work for the baseball statistics startup Baseball Prospectus. It was a good living for a couple of years—but like most edges in gambling, it wouldn’t last. Some of this was the natural evolution of the game: the Poker Boom sputtered into more of a poker plateau as losing players either went broke, quit, or got better, removing one sucker from the table at a time.
> But it was also partly the doing of the U.S. Congress. In late 2006, the GOP-led Congress, hungry for a victory with “moral majority” voters ahead of the midterms as Republican congressman Mark Foley resigned from office for having sent sexually explicit messages to underage male pages, passed a bill called the Unlawful Internet Gambling Enforcement Act (UIGEA).
> The UIGEA didn’t ban online poker per se, but it established regulations that choked off payment processors: it’s hard to play poker if you can’t exchange cash for chips. Some sites closed to U.S. players while others remained open, but between the shadow of illegality and the increased friction of getting your money in and out, inexperienced new players avoided the games, making them much tougher to beat.
> There was one silver lining: the UIGEA piqued my interest in politics. The bill had been tucked into an unrelated piece of homeland security legislation and passed during the last session before Congress recessed for the midterms. It was a shifty workaround, and having essentially lost my job, I wanted the people responsible for it to lose their jobs, too. And they did: Republicans lost both the House and the Senate, including the seat of Representative Jim Leach of Iowa, the chief sponsor of the UIGEA, whose thirty-year tenure in office ended partly because of poker players who had contributed money to his opponent.
> Struggling to win money as the games were drying up, I quit poker about six months later. With my newfound interest in politics and the extra time on my hands, I wound up starting FiveThirtyEight in 2008...
And I really do not think Silver has any idea what kind of a look he is giving into himself here.
---
The full Kaleberg comment:
> **Kaleberg**: 'I don't think Silver understands risk. He has a romanticized view limited to those who have a choice as to whether to accept certain risks or not. Failure for venture capitalists means not making the big score, not losing everything. There's a reason venture capitalists tend to come from relatively wealthy families. Compulsive gamblers may risk losing everything, but professional gamblers are more like venture capitalists. It may be a negative sum business, but it's still a business.
> A woman having sex with a guy she met at a bar takes a much bigger risk than any of these. Ditto for a fireman or peace officer responding to a 911 call. Immigrants to the US took on a broad variety of risks, and despite overcoming many of them are now being clobbered by a long shot outcome. These risks are much harder to romanticize. They are definitely not flowing in Silver's river...
======
**DELONG’S GRASPING REALITY: Trying to make my readers—and myself—smarter. I think I am a go-to source to understand things economic in the past and in the present. Too online since 1995. Subscribe! <`https://graspingreality.substack.com/subscribe`> Currently featuring**:
Testing for AI-assisted/generated text should be close to a minimum test for LLM AI demonstrating skill in pattern matching. In the all-data-encompassing big data center AI limit in ignoring the still NP-too-hard state of technology the primary question is simple: does this exact text sequence exist in the all encompassing data store? If the new text fed to the AI-assisted predicate filter does not exist in the store then it's a coin flip if the new unknown text is human or AI-generated and not in the data store. In the non-complete-store search reality it can be estimated by many approaches but for some range of multi-word phrases or better semantic labeled token sequences in the test input my intuition is the statistical prevalence in the training data store is different (depending on how you create your metric) than in AI-generated input which you can generate (bootstrap). Also in the full data store limit, but also applicable here, is using in the same analysis the corpus of Brad DeLong known writings and recorded talks, which, if you have all the data and enough compute, you can analyze using standard statistical analysis techniques, and at a legally robust level be able to argue in court that Brad wrote this or AI-assist was used. I don't read science fiction but it's easy to think of a lot of ways that store of all knowledge and the compute to use "classic" statistical analysis tools and our existing suite of laws with upstanding people like Todd Blanche at DOJ to be sub-optimal for many of us. I have reasons to be skeptical of many of the claims the AI-evangelicals and how data-center scale impacts the true knowledge generation ability of LLM-AI for the real domains where knowledge is generated (not entertainment) but in any society control maximization (political, economic) is definitely optimized if you hold all the data (including/especially surveillance) and have the compute to manipulate it.
I don't think Silver understands risk. He has a romanticized view limited to those who have a choice as to whether to accept certain risks or not. Failure for venture capitalists means not making the big score, not losing everything. There's a reason venture capitalists tend to come from relatively wealthy families. Compulsive gamblers may risk losing everything, but professional gamblers are more like venture capitalists. It may be a negative sum business, but it's still a business.
A woman having sex with a guy she met at a bar takes a much bigger risk than any of these. Ditto for a fireman or peace officer responding to a 911 call. Immigrants to the US took on a broad variety of risks, and despite overcoming many of them are now being clobbered by a long shot outcome. These risks are much harder to romanticize. They are definitely not flowing in Silver's river.
One more note on how I wish we had the early-days Nate Silver of <http://fivethirtyeight.com> back, and how the story of the evolution here is rather tragic. Silver understood poll aggregation. But now he writes about things he does not understand. Kaleberg had a nice comment on this:
> **Kaleberg**: 'I don't think Silver understands [real] risk. He has a romanticized view limited to those who have a choice as to whether to accept certain risks or not. Failure for venture capitalists means not making the big score, not losing everything...
And let me chime in and add to this:
There is a very very very very very important distinction between people who are working with their counterparties, and people who are working against them. Or, indeed—going further into the psychopath realm—people who do not think that their counterparties are real people of any concern to them at all.
Case in point: Silver's extraordinary and persistent rage against Congressman Jim Leach.
IMHO Jim Leach did very few good things with his time in the U.S. Congress. But one of those good things was placing roadblocks to try protect people who did not understand the risks they were running from so easily becoming marks grifted by the likes of Nate Silver.
And Silver's rage against Jim Leach still persists. It looks as though Jim Leach will live rent-free in Nate Silver's brain until the day Silver dies, if not after:
> **Nate Silver**: On the Edge <>: 'I’d been a professional poker player for three years between 2004 and 2007, during the so-called Poker Boom. The Poker Boom began
because of the increasing availability of online poker and because of Chris Moneymaker, an accountant from Nashville who won an online qualifying tournament for a seat at the $10,000 Main Event at the 2003 World Series of Poker and then parlayed that into winning the Main Event for $2.5 million.
> If you’d asked ChatGPT to design a person who would most increase the amount of interest in poker by winning the WSOP, it might have spat out Moneymaker. An affable, pudgy, late twenty-something dude with a boring corporate job, he was exactly the customer the online poker sites were targeting, an archetype for every office drone who wanted to break out of his cubicle and win the big jackpot.
> The number of participants in the World Series of Poker Main Event exploded from 839 in Moneymaker’s 2003 to 8,773 just three years later in 2006, largely fueled by people who had won their seats online. I was one of those people who lived the dream. I soon found myself on a nocturnal schedule. Poker games are usually best late at night, when your opponents are some combination of drunk, sleep-deprived, or delirious from winning or losing a bunch of money. So I’d come home from my cubicle,
take a nap, and then play poker online, sometimes straight through until the morning, when I’d straggle into work and struggle through the day.
> Needless to say, this wasn’t sustainable, and—making considerably more money as a poker player than as a consultant—I quit my corporate job within about six months to play poker and work for the baseball statistics startup Baseball Prospectus. It was a good living for a couple of years—but like most edges in gambling, it wouldn’t last. Some of this was the natural evolution of the game: the Poker Boom sputtered into more of a poker plateau as losing players either went broke, quit, or got better, removing one sucker from the table at a time.
> But it was also partly the doing of the U.S. Congress. In late 2006, the GOP-led Congress, hungry for a victory with “moral majority” voters ahead of the midterms as Republican congressman Mark Foley resigned from office for having sent sexually explicit messages to underage male pages, passed a bill called the Unlawful Internet Gambling Enforcement Act (UIGEA).
> The UIGEA didn’t ban online poker per se, but it established regulations that choked off payment processors: it’s hard to play poker if you can’t exchange cash for chips. Some sites closed to U.S. players while others remained open, but between the shadow of illegality and the increased friction of getting your money in and out, inexperienced new players avoided the games, making them much tougher to beat.
> There was one silver lining: the UIGEA piqued my interest in politics. The bill had been tucked into an unrelated piece of homeland security legislation and passed during the last session before Congress recessed for the midterms. It was a shifty workaround, and having essentially lost my job, I wanted the people responsible for it to lose their jobs, too. And they did: Republicans lost both the House and the Senate, including the seat of Representative Jim Leach of Iowa, the chief sponsor of the UIGEA, whose thirty-year tenure in office ended partly because of poker players who had contributed money to his opponent.
> Struggling to win money as the games were drying up, I quit poker about six months later. With my newfound interest in politics and the extra time on my hands, I wound up starting FiveThirtyEight in 2008...
And I really do not think Silver has any idea what kind of a look he is giving into himself here.
---
The full Kaleberg comment:
> **Kaleberg**: 'I don't think Silver understands risk. He has a romanticized view limited to those who have a choice as to whether to accept certain risks or not. Failure for venture capitalists means not making the big score, not losing everything. There's a reason venture capitalists tend to come from relatively wealthy families. Compulsive gamblers may risk losing everything, but professional gamblers are more like venture capitalists. It may be a negative sum business, but it's still a business.
> A woman having sex with a guy she met at a bar takes a much bigger risk than any of these. Ditto for a fireman or peace officer responding to a 911 call. Immigrants to the US took on a broad variety of risks, and despite overcoming many of them are now being clobbered by a long shot outcome. These risks are much harder to romanticize. They are definitely not flowing in Silver's river...
======
**DELONG’S GRASPING REALITY: Trying to make my readers—and myself—smarter. I think I am a go-to source to understand things economic in the past and in the present. Too online since 1995. Subscribe! <`https://graspingreality.substack.com/subscribe`> Currently featuring**:
* The Fourteen-Lion Parade <https://braddelong.substack.com/p/fourteen-lion-parade>
* Lessons for Debt Control from Clinton's Success in the 1990s <https://braddelong.substack.com/p/lessons-for-debt-control-from-clintons>
* Fun, Not Fear: Rebuilding Universities for the Machine-Learning Era <https://braddelong.substack.com/p/university-assessment-in-the-ai-age>
* "Star Trek: The Wrath of Khan" as Franchise Hinge <https://braddelong.substack.com/p/2025-12-24-star-trek-the-wrath-of>
<https://braddelong.substack.com/p/donald-trump-is-too-old-to-be-president>
======
<https://braddelong.substack.com/p/substackpangram-is-scanning-the-substack>
I found it. For a wonderfully unromantic take on the prize fighting world: "Why I Fixed Fights" over at https://deadspin.com/why-i-fixed-fights-1535114232/
Did you check for false positives? Is the fraction of posts written by the actual Brad DeLong tagged as AI-assisted zero percent?
So far the flow is. I have not gone back and checked the stock...
Testing for AI-assisted/generated text should be close to a minimum test for LLM AI demonstrating skill in pattern matching. In the all-data-encompassing big data center AI limit in ignoring the still NP-too-hard state of technology the primary question is simple: does this exact text sequence exist in the all encompassing data store? If the new text fed to the AI-assisted predicate filter does not exist in the store then it's a coin flip if the new unknown text is human or AI-generated and not in the data store. In the non-complete-store search reality it can be estimated by many approaches but for some range of multi-word phrases or better semantic labeled token sequences in the test input my intuition is the statistical prevalence in the training data store is different (depending on how you create your metric) than in AI-generated input which you can generate (bootstrap). Also in the full data store limit, but also applicable here, is using in the same analysis the corpus of Brad DeLong known writings and recorded talks, which, if you have all the data and enough compute, you can analyze using standard statistical analysis techniques, and at a legally robust level be able to argue in court that Brad wrote this or AI-assist was used. I don't read science fiction but it's easy to think of a lot of ways that store of all knowledge and the compute to use "classic" statistical analysis tools and our existing suite of laws with upstanding people like Todd Blanche at DOJ to be sub-optimal for many of us. I have reasons to be skeptical of many of the claims the AI-evangelicals and how data-center scale impacts the true knowledge generation ability of LLM-AI for the real domains where knowledge is generated (not entertainment) but in any society control maximization (political, economic) is definitely optimized if you hold all the data (including/especially surveillance) and have the compute to manipulate it.