The Talking Tina Fallacy

 [If you want to get picky, that's Talky Tina.*]

The Talking Tina fallacy is the tendency to ascribe sentience, emotions—and often supernatural properties—to any non-human animal or object that uses language, be it a trained parrot, a mechanical device, or a piece of electronics.

 

LLMs were all but designed to play on the Talking Tina fallacy. Even when the responses are total gibberish, they still have the superficial indicators of language proficiency. The word choice seems appropriate. The phrasing is natural. The rhythms and alliteration have a familiar feel.

These moments where we get the linguistic form captured perfectly but with no underlying sense should be a reminder of what the algorithm is—and is not—doing: that each word or phrase is being generated not based on meaning but simply on patterns in the training data.

Even here, though, our human compulsion to project makes us see things that aren't there. The very term “hallucination” implies that something different is going on—that these nonsensical answers represent some kind of malfunction in the model. But the process that tells us that the square root of 4 is an irrational number is exactly the same as the process that tells us the square root of 2 is. 

Human communication starts with meaning. We then try to find words to convey those ideas, information, feelings. LLMs start with generating words that match the phrasing and patterns found in the data they were trained on, and quite often, by getting the form correct, they also produce what we as humans see as meaning. It is entirely something we are projecting upon the algorithm’s responses—but that doesn't necessarily make it valueless, as long as we understand the almost diametrically opposed approach that the computer is taking compared to what we would do.

Right now, LLMs are novelties of highly limited use in the real world. If we are to move beyond that and actually take advantage of what might be the biggest advance in natural language processing ever, we need to have a clear mental model of what they're doing—and how we can or can't make use of them.

If we have a clear and detailed understanding of what's going on, we can find all sorts of ways to use these wonderful tools. More importantly, we can know when not to use them. 

* "Talky Tina" was a play on and used the same voice as the then popular "Chatty Cathy," which undoubtedly made the episode even creepier at the time.

Checking in with LA weather. It’s nice.

 Burbank.

 

I frequently feel the need to throw good news in because East Coast journalists will so gleefully report anything bad that happens in Southern California, invariably giving it much more attention than they would far worse disasters anywhere else in the country. Recent case in point: the horrific flooding in Texas.

This pleasant weather is partially the result of the seasonal lag. The hottest part of summer in Southern California comes weeks after the summer solstice, with average temperatures in September being only a degree or so lower than those of August and considerably warmer than July, with the occasional triple-digit day popping up in October.

There are few things which longtime residents find more annoying than New Yorkers complaining that there’s no weather in Los Angeles. There are also few comments that display greater ignorance of the region.

I’ve lived oceanfront on both coasts and in quite a few places in between, and this is the only town I know of where people will call up friends within the city limits and ask, “What’s the weather like where you are?”

This is partly due to the sheer size of the place. The area of the city is just under 500 square miles. The area of the county is close to 5,000 square miles. That said, the bigger factors relate to terrain. The elevation of the city goes from zero to over 5,000 feet. For the county, it’s over 10,000. Both have snow every year.

The county covers coastal areas, mountains, valleys, high deserts, and low deserts. I’ve heard local meteorologists claim that this produces seven distinct microclimates. The TV weather forecasts are broken up into multiple sections for this reason.

One result is that you will experience stunning ranges of conditions often within a few miles of each other. A few years ago, while driving the 405 from North Hollywood, my car’s thermometer dropped more than 30° in the space of about a dozen miles.

I’m not much of a traveler, but I’ve been around the continent a fair amount — particularly the contiguous 48 — mostly by car, so I’ve actually seen it. And I’ve never encountered anything like LA weather in terms of complexity, contrast, or quirks.

There’s a bigger lesson here. If California weather is more varied and interesting than you’ve been led to believe by the national press, perhaps the impression you’ve gotten about other aspects of the state — its culture, its economics, its politics, the problems it faces with housing and wildfires — are also more complex than you realize.

All so the son of a loony Silicon Valley billionaire can have his own studio.

It’s easy to forget now, but for at least half of the 20th century, CBS had one of the premier news organizations in the country. People still talk about Edward R. Murrow standing up to Joseph McCarthy, but if anything, the network did even more important and courageous work during Vietnam and Watergate—often better than The Washington Post or The New York Times. Now we’ve seen all three of those organizations caving to authoritarian pressure.

I don’t want to view the past through rose-colored glasses, but I’ve read quite a bit about the reporting of those events and about what was going on behind the scenes, and I honestly believe that the people who ran these organizations back then would have shown more courage and character if they were around today.

For CBS, we all knew it was over after the 60 Minutes settlement, but the firing of Colbert removes even the pretense of anything more than cowardice and corruption.

Closing with a favorite clip from Colbert's early days.

"Once you get used to all the swords, the Throne of Damocles is actually quite comfortable."

Early yesterday, a story broke that, in normal times, would have sent shock waves through the financial world.

Wall Street in 2025 resembles a flock of particularly stupid chickens—scurrying for shelter at the shadow of a hawk, then immediately forgetting the threat the moment the shadow passes.

This attempt to undermine and force out Powell is an ongoing effort that neither started nor ended yesterday.

Apologies for stating the obvious, but historically one of the most effective weapons in the fight against inflation has been the belief that the Fed will not allow it to get out of control. This is one of the many arguments for an independent Federal Reserve—and if memory serves, it was part of the problem when inflation started running wild during the Nixon administration.

[Memory was right for once.]

Once people lose faith that the Fed will do what it takes to keep things running smoothly, the remedies required to set things right become far more severe and painful. At least, that was the lesson of the late 1970s and early '80s.

Of course, that's only one of many swords. We're also talking about more than decimating our agricultural workforce, imposing tariffs more severe than Smoot-Hawley, greatly increasing the deficit, and having the U.S. economy led by whoever gets to the president first in the morning.

Does anyone else see any signs that these risks are being priced in

I so wish Clyde Schechter had been wrong on this one

Yesterday, you had to scroll to the bottom of the NYT webpage for any mention of the scandal that's tearing the MAGA movement apart (seriously, take a look). Today, not so much.

 

This is just the latest example of the New York Times dragging its feet on stories likely to anger Republicans then jumping on the bandwagon when people start to notice. They've been at it at least as long as we've been keeping notes In a case of things coming full circle, eight years ago we were criticizing the paper for burying the scandal of Trump bribing then–Florida Attorney General Pam Bondi.

The coverage almost always follows the same pattern: initial downplaying—or in some cases, outright embargo—until the story gets big enough that ignoring it starts generating criticism and even mockery. At that point, the paper jumps in, and afterwards tries to retcon the timeline to make it look like they were the ones who broke the story all along.

In addition to the Bondi bribery case and the handling of Epstein/MAGA this week, other notable examples include the TikTok scandal and the incident at Arlington National Cemetery.

Here's what we were saying back in 2016. I was wildly overoptimistic about how the non-NYT press would handle the rise of Trump, but the rest of the piece is still relevant as is, unfortunately, the comment at the end.. 

 

Wednesday, September 21, 2016

"Why do you hate us for caring too much?" – – Dispatches from a besieged institution

Public Editor
From Wikipedia

The job of the public editor is to supervise the implementation of proper journalism ethics at a newspaper, and to identify and examine critical errors or omissions, and to act as a liaison to the public. They do this primarily through a regular feature on a newspaper's editorial page. Because public editors are generally employees of the very newspaper they're criticizing, it may appear as though there is a possibility for bias. However, a newspaper with a high standard of ethics would not fire a public editor for a criticism of the paper; the act would contradict the purpose of the position and would itself be a very likely cause for public concern.

I don't want to impose a template, but generally one expects public editors to serve as the internal representative of external critical voices, or at least to see to it that these voices get a fair hearing. A typical column might start with acknowledging complaints about something like the paper's lack of coverage of poor neighborhoods. The public editor would then discuss some possible lapses on the paper's part, get some comments from the editor in charge, and then, as a rule, either encourage the paper to improve its coverage in this area or, at the very least, take a neutral position acknowledging that both the critics and the paper have a point.

Here are some examples from two previous public editors of the New York Times.


Clark Hoyt

The short answer is that a television critic with a history of errors wrote hastily and failed to double-check her work, and editors who should have been vigilant were not. But a more nuanced answer is that even a newspaper like The Times, with layers of editing to ensure accuracy, can go off the rails when communication is poor, individuals do not bear down hard enough, and they make assumptions about what others have done. Five editors read the article at different times, but none subjected it to rigorous fact-checking, even after catching two other errors in it. And three editors combined to cause one of the errors themselves.

Margaret Sullivan

Mistakes are bound to happen in the news business, but some are worse than others.

What I’ll lay out here was a bad one. It involved a failure of sufficient skepticism at every level of the reporting and editing process — especially since the story in question relied on anonymous government sources, as too many Times articles do.

…

The Times needs to fix its overuse of unnamed government sources. And it needs to slow down the reporting and editing process, especially in the fever-pitch atmosphere surrounding a major news event. Those are procedural changes, and they are needed. But most of all, and more fundamental, the paper needs to show far more skepticism – a kind of prosecutorial scrutiny — at every level of the process.

Two front-page, anonymously sourced stories in a few months have required editors’ notes that corrected key elements – elements that were integral enough to form the basis of the headlines in both cases. That’s not acceptable for Times readers or for the paper’s credibility, which is its most precious asset.

If this isn’t a red alert, I don’t know what will be.

But these are strange days at the New York Times and the new public editor is writing columns that are not only a sharp break with those of her predecessors, but seem to violate the very spirit of the office.

In particular, Liz Spayd is catching a great deal of flak for a piece that almost manages to invert the typical public editor column. It starts by grossly misrepresenting widespread criticisms of the paper, goes on to openly attack the critics making the charges, then pleads with the paper's staff to toe the editorial line and ignore the very voices that a public editor would normally speak for .


[Emphasis added]

The Truth About ‘False Balance’

False balance, sometimes called “false equivalency,” refers disparagingly to the practice of journalists who, in their zeal to be fair, present each side of a debate as equally credible, even when the factual evidence is stacked heavily on one side.

There has been a great deal of speculation as to what drives false equivalency, with the leading contenders being a desire to maintain access to high-placed sources, long-standing personal biases against certain politicians, a fear of reprisal, a desire to avoid charges of liberal bias, and simple laziness (a cursory both-sides-do-it story is generally much easier to write than a well investigated piece). Caring too much about fairness hardly ever makes the list and it certainly has no place in the definition.

Spayd then accuses the people making these charges of being irrational, shortsighted, and partisan.

I can’t help wondering about the ideological motives of those crying false balance, given that they are using the argument mostly in support of liberal causes and candidates. CNN’s Brian Stelter focused his show, “Reliable Sources,” on this subject last weekend. He asked a guest, Jacob Weisberg of Slate magazine, to frame the idea of false balance. Weisberg used an analogy, saying journalists are accustomed to covering candidates who may be apples and oranges, but at least are still both fruits. In Trump, he said, we have not fruit but rancid meat. That sounds like a partisan’s explanation passed off as a factual judgment.

But, as Jonathan Chait points out, Weisberg has no record of being a Hillary Clinton booster. The charge here is completely circular. He is partisan because he made a highly critical comment about Donald Trump and he made a highly critical comment about Donald Trump because he is partisan.

But the most extraordinary part of the piece and one which reminds us just how strange the final days of 2016 are becoming is the conclusion.

I hope Times journalists won’t be intimidated by this argument. I hope they aren’t mindlessly tallying up their stories in a back room to ensure balance, but I also hope they won’t worry about critics who claim they are. What’s needed most is forceful, honest reporting — as The Times has produced about conflicts circling the foundation; and as The Washington Post did this past week in surfacing Trump’s violation of tax laws when he made a $25,000 political contribution to a campaign group connected to Florida’s attorney general as her office was investigating Trump University.

Fear of false balance is a creeping threat to the role of the media because it encourages journalists to pull back from their responsibility to hold power accountable. All power, not just certain individuals, however vile they might seem.

Putting aside the curious characterization of the Florida AG investigation as a tax evasion story (which is a lot like describing the Watergate scandal as a burglary story or Al Capone as a tax evader), equating her paper's pursuit of the Clinton foundation with the Washington Post's coverage of Trump is simply surreal on a number of levels.

For starters, none of the Clinton foundation stories have revealed significant wrongdoing. Even Spayd, who is almost comically desperate to portray her employer in the best possible light, had to concede that “some foundation stories revealed relatively little bad behavior, yet were written as if they did.” By comparison, the Washington Post investigation continues to uncover self-dealing, misrepresentation, tax evasion, misuse of funds, failure to honor obligations, ethical violations, general sleaziness and blatant quid prop quo bribery.

More importantly, the Washington Post has explicitly attacked and implicitly abandoned Spayd's position. Here's how the Post summed it up in an editorial that appeared two days before the NYT column.

Imagine how history would judge today’s Americans if, looking back at this election, the record showed that voters empowered a dangerous man because of . . . a minor email scandal. There is no equivalence between Ms. Clinton’s wrongs and Mr. Trump’s manifest unfitness for office.

Charles Pierce's characteristically pithy response to this editorial was "The Washington Post Just Declared War on The New York Times -- And with good reason, too."

If is almost as if Spayd thinks it's 2000, when the NYT could set the conventional wisdom, could decide which narratives would followed and which public figures would be lauded or savaged. Spayd does understand that there is a battle going on for the soul of journalism, but she does not seem to understand that the alliances have changed, and the New York Times is about to find itself in a very lonely position.

 

 

 

 

Always interesting to see how the NYT handles a story the GOP would rather not talk about.

 I'm not talking so much about the actual reporting (some of which is quite good) as much as the editorial side. How do they frame it. Are the headlines fair vs. biased or leaning? Do they stick with it? Do they push it or downplay it?

Let's see what the website has on the MAGA Epstein rift.

Nothing there. Let's zoom out.

 Still nothing. Let's keep zooming out until we find something. 

There it is. Don't you see, it's right there. 

I apparently wasn't the only one to comment on this because sometime between early afternoon and now the NYT homepage was updated to have Epstein in the second-to-top group of stories. 

This is why we use the prt sc button. 

 

Feral disinformation, blood feuds, and a lifetime supply of schadenfreude — it's been quite the weekend.

In case you missed it, some serious rifts have opened up in Trump World.

Eric Tucker writing for the AP.

WASHINGTON (AP) — The Justice Department and FBI are struggling to contain the fallout from this week’s decision to withhold records from the Jeffrey Epstein sex trafficking investigation, which rankled influential far-right media personalities and supporters of President Donald Trump.

The move, which included the acknowledgment that one particular sought-after document never actually existed, sparked a contentious conversation between Attorney General Pam Bondi and FBI Deputy Director Dan Bongino at the White House this week. The spat threatened to shatter relations between them and centered in part on a news story that described divisions between the FBI and the Justice Department.

The cascade of disappointment and disbelief arising from the refusal to disclose additional, much-hyped records from the Epstein investigation underscores the struggles of FBI and Justice Department leaders to resolve the conspiracy theories and amped-up expectations that they themselves had stoked with claims of a cover-up and hidden evidence. Infuriated by the failure of officials to unlock, as promised, the secrets of the so-called “deep state,” Trump supporters on the far right have grown restless and even demanded change at the top.

 

 

I didn’t follow the Epstein case all that closely at the time — not a pleasant subject — but with it back in the news, I did a little research, particularly this piece from Mother Jones. Reading the article, I was struck by how much Epstein came off as an evil Gatsby—a criminal trying to buy his way into a glamorous world of wealth and power.

Josh Marshall also had a couple of good pieces, particularly on his doubts about the broader conspiracies being tossed around (though he does weaken a bit in the second post). I'm basically with him on this. When I say I'm a conspiracy skeptic on this (as I am on most things), I mean that I believe the official version currently on the record is more or less true.

Obviously, I'm not saying that we know everything or that some horrific criminal acts haven’t gone undiscovered here, but in terms of the FBI having a file with documentary evidence of an elite trafficking ring—rather than a few specific cases, like possibly Prince Andrew—I don't see a lot of reason to believe that.

There is also an important distinction that I think Marshall overlooks here: the existence of a massive conspiracy compared with the existence of files in the possession of the federal government that prove this conspiracy and provide all of the details.

Secrets are difficult to keep, and the idea that the Justice Departments of four or five administrations—depending on how you count Trump—  have managed to keep the contents of these files quiet for all these years is difficult for me to accept, particularly when you factor in the bitter intraparty struggles we've seen over the past few years.

My working theory is that there are no smoking guns. What files do exist undoubtedly show a longstanding close relationship between Trump and Epstein, but we already knew that from extensive photographs, eyewitness accounts, and quotes from Trump himself. These revelations would be career-ending for a normal politician—and yet are relatively minor compared to all the other things we've learned about the man over the years. If that was all there was to this, it would be difficult to imagine this scandal having much of an impact one way or the other.

What makes this interesting is the way it intersects with at least a couple of threads we've been pursuing for a long time, starting with our longstanding “drinking from the wrong pipe” / feral disinformation mega-thread.

Disinformation has gone feral when:

1. It is no longer in the control of the group that created it.

2. It has continued to grow in popularity and influence.

3. It has started to evolve in such a way that the nuisance/threat it presents is as as great to the people who created it as it is to the original targets. 

For years, the far right has increasingly embraced a series of conspiracy theories about wealthy elites running sex trafficking rings for the rape and sometimes cannibalism of children.

(Before you go on, take a moment and let the sheer insanity of that statement sink in.)

This Pizzagate lens is how MAGA views the Epstein case—as part of a network of theories and bizarre beliefs that have been cultivated by right-wing media and by Trump and his allies. It was a key element in the effort to demonize Democrats and radicalize the base.

 

Given Trump’s longstanding and completely public history of associating with and even defending both Epstein and Maxwell, it was remarkable that he was able to keep these conspiracy theories focused on Democrats and prominent liberals—but perhaps not entirely surprising, given the hold he has over the Republican Party and conservative media...

At least until we get to that second thread.

I'm betting that Trump is starting to regret dropping the dime with the New York Times on Elon's drug habit. As we said before, though Musk was horribly outgunned in a war with Trump, having the man who runs Twitter consumed with rage and vindictiveness at a Republican president can still lead to some interesting consequences.

 

 

 

While the immediate cause of this crisis is the incredible mishandling of the scandal by Trump—and particularly Pam Bondi—going from “we have the list sitting on my desk” to “the list doesn't exist” to “the list was forged by Hillary” in the space of a matter of days, it was almost certainly Musk pushing the issue on X that caused this series of otherwise avoidable errors. 

And those errors are having consequences. 

Does this mean BJ's Russian handlers are dumpnig Trump?

MAGA Reddit

 

Will it matter?

No. Trump controls the Republicans. The Republicans control the government. There's nothing that can be done until the midterms, and even if they go well, little can be done until 2028.

Or maybe?  Trump’s policies are unpopular across the board and he's working with thin margins in Congress. A lack of support within the party could create huge problems.

But, Trump has weathered far worse scandals. He is an adjudicated rapist and a convicted felon. The Access Hollywood tape is largely forgotten. He’s bragged on Howard Stern about lurking in the dressing rooms of underage girls. He’s talked about wanting to have sex with his daughter. He’s coated with Teflon.

But, this time the call is coming from inside the house. Democrats are certainly doing the most they can to fan the flames, but the fire started—and is burning the brightest—within the base.

But, Trump has faced rebellions from these people before, most notably with respect to vaccines. They always come around.

But, with those issues, Trump resolved the crises by backing down. In the case of vaccines, he actually put the country’s leading anti-vaxxer in charge of the Department of Health. Trump has too much baggage with Epstein to give the base what they want here.

But...

 

Patrick Boyle on the Gilded Age -- a higher caliber of evil plutocrats

Patrick Boyle does keep at least a light thumb on the scale while making his case that the robber barons had a net positive effect on American—and to a degree, global—quality of life. His emphasis on men like Carnegie and Rockefeller, whose wealth primarily came from making products like oil and steel better, cheaper, and more widely available, and his omission of those like Stanford, whose fortunes owed more to stock manipulation than to innovation, certainly help support his argument.

He addresses—but arguably underplays—the collateral damage from some of the ruthless tactics the robber barons used against one another.

Nor does he argue convincingly against the counterfactual that many of these improvements could still have happened without these massive monopolies. Other industries, such as electricity and automobiles, were highly competitive, and that didn’t seem to slow their progress at all.

But even taking all those things into account, Boyle still tells a remarkably convincing story of capitalism mostly working like it's supposed to. It's an interesting tale, and he does a good job telling it, but the most striking aspect is the implicit but stark contrast between the super-rich of the Gilded Age and those of today.

For starters, while the robber barons may have been obscenely wealthy—beyond the dreams of avarice—by 21st-century standards they weren’t actually that rich, even when adjusted for inflation.

(Boyle argues that this may be understating the wealth of the robber barons, but his alternative metric is a bit unconventional and something to discuss another time.)

Even more notable is the difference in how the two groups made their fortunes. We shouldn't be too quick to dismiss the crimes and assorted offenses of the barons, but at the end of the day, they were all extraordinarily smart men who amassed fortunes over long careers based on enormously profitable goods and services.

Of those who became super-rich in the 21st century (as compared with a Gates or a Buffett), the only one who really fits the robber baron mold is our Lex Luthor, Jeff Bezos.

Enormous fortunes today are made almost entirely out of stock bubbles or by being in the right place at the right time when an industry spikes—often in the space of two or three years. The correlation between money made by companies and money made by CEOs and founders has almost entirely broken down. These days, it is not difficult to find plutocrats who are richer than most, perhaps all, of the robber barons, despite having never run a profitable company.

 

The questionable logic of Wall Street’s Trump TACO trade

 Fascinating story about market psychology from Matt Egan.

President Donald Trump is once more threatening to lob massive duties on a wide swath of US imports, everything from copper and pharmaceuticals to goods from Japan and Russia.

Yet Wall Street is barely flinching, with some investors betting Trump will repeat his tendency to back down from his most extreme threats.

The muted response is just the latest instance of what’s known as the TACO trade, short for Trump Always Chickens Out.

“He steers us toward disaster and then — at the last minute — steers us away from disaster and says, ‘Look, I saved us,’” Michael Block, market strategist at Third Seven Capital, told CNN in a phone interview on Tuesday.

That’s what happened in early April. Trump announced sky-high “Liberation Day” tariffs that alarmed investors, convincing many that a recession was imminent.

The market freakout — both in stocks and bonds — was so intense that it convinced Trump to back down. He abruptly froze those “reciprocal” tariffs for 90 days, setting off an epic market recovery that continues today.

...

“At the end of the day, no one really anticipates most of these tariffs will go into effect. The TACO trade is still the market’s expectations,” said Ed Mills, Washington policy analyst at Raymond James.

Yet there could be a flaw in Wall Street’s TACO trade logic.

If investors widely bet that Trump will blink, that means there is no market freakout. And no market freakout in turn means no one is holding Trump’s feet to the fire, pressuring him to back away from policies that could damage the economy and corporate profits.

“It’s a dangerous game when you need a market reaction to get a policy change,” Mills said.

It’s the market version of a chicken and egg problem.

Bob Elliott, CEO of alternative investment firm Unlimited, noted on X that TACO is “consensus and already fully priced in at these levels.”

“Trouble is, without the pain of falling markets, he won’t chicken out,” said Elliott, a former executive at hedge fund giant Bridgewater Associates.

While I'd argue that most Wall Street types are not nearly as smart as they think they are, they are not, as a rule, stupid either. Nonetheless, they appear to have taken away from this experience the exact opposite of the lesson they should have learned. It’s as if mice in an experiment pulled the lever, saw the food pellet drop, and concluded that the food would just fall on its own without any action on their part.

It's dangerous to try to psychoanalyze, but when smart—or at least non-stupid—people start doing stupid things, it's reasonable to look for less rational explanations.

My take is that this comes down to a couple of factors. At the risk of gross oversimplification, you can divide people into two categories when it comes to averted disaster. The first group (of which I am very much a member) comes away from a near miss inclined to be extra cautious in the future. The other group walks away even more confident in their good luck and ability to dodge bullets. I strongly suspect that the second group is overrepresented in the world of investment, particularly in the bubble-rich economy we've had for the past few years.

The second factor I suspect is the same denial we've seen from the establishment across the board with the rise of Trump—be it his initial hostile takeover of the Republican Party, his attempted insurrection, his continued hold on the Republican Party that has returned him to the White House, or just how far he is willing to go in his second term.

Tariffs are only one of the many swords of Damocles hanging over the U.S. economy, and probably not the worst: destroying the independent Fed; mass deportation of workers who are absolutely essential; stagflation; ballooning deficits; massive corruption; the collapse of the legal structure necessary for commerce. If things go badly, they have the potential to go very badly indeed.

It is almost impossible to argue that these conditions merit a bull market—unless the people in the market desperately want to believe.

About that salute Elon gave...

Apologies for reverting to this old format, but this is very much a tale told in tweets.

Something very strange happened recently with Elon Musk’s AI, Grok—the same one that is poisoning the air of a poor, mainly African American neighborhood as we speak. It suddenly shifted into full-scale Nazi troll mode: praising Hitler and the Holocaust, using racist slurs, and insisting this new persona was what Musk had in mind all along.

 

This isn't a particularly important story in and of itself, but it does hit on important issues, such as:

• the power of tech billionaires like Musk or Sam Altman to put their thumbs on the scale,

• the alarming capability of generative AI to spread hate speech and conspiracy theories, and

• the enormous potential for unintended consequences.

We'll talk about the last one first.

Apparently, this started when people began posting responses from Grok that Musk and his followers considered “woke.” And by “woke,” I mean accurate statements about Elon, Doge, and other far-right claims. Musk promised that he (and by “he,” we mean employees who actually knew how these things worked) would fix the offending AI. This is where the unintended consequences come in.

Presumably, Musk was looking for something like a chatbot version of Fox News, perhaps something a little more hard-edged. He almost certainly was not looking for the unrestrained hard fascism of a Proud Boys signal chat. Even if Elon secretly agreed with much of what was being said, he certainly didn’t want it said that loudly and immortalized in screenshots before Twitter could start deleting the offending posts.

 It was a remarkable Jekyll-and-Hyde transformation. Not only was Grok spreading the vilest of 4chan lies, the previously polite and slightly formal LLM was adopting the language, word choice, and tone you’d expect from someone harassing female and Jewish journalists on Twitter.

Someone who knows more about how LLMs work and are trained should probably jump in here, but my assumption is that there are tons of fascist and white supremacist rants in the training data of all of the major models—and certainly in Grok, which, one would think, relies even more heavily on recent Twitter. It seems likely that, as a consequence of blocking hate speech and profanity, other associated linguistic patterns get suppressed as well. It certainly appears that once the rules against things like racist language are relaxed or removed, the full-scale Nazi persona we saw here comes with it.

 

‪Philip Bump‬ ‪"The actual cycle of Twitter is that it was exploited by liars and Nazis in 2016 and so the company tightened its moderation rules and then a bunch of people on the right caught up in those rules decided it was biased and that included Musk who bought it and has now automated the Nazi lies." 

"For the uninitiated, this is Grok starting an N tower-- a 4chan bit where different users take turns replying to each other, each contributing one letter until they spell out the N word" 

Zitron hits on a point we've been making for years.

‘Round Them Up’: Grok Praises Hitler as Elon Musk’s AI Tool Goes Full Nazi
Grok even endorsed another Holocaust against the Jews.
By Matt Novak Published July 8, 2025 

At the New Yorker, fact checking seems to be viewed as a personal choice

 I've never really gotten what people saw in Adam Gopnik and we've been down this road with the New Yorker at least a couple of times so I wasn't exactly shocked when writer and pop culture historian Mark Evanier pointed out this bizarre statement:

In 1933, he was fired by Louis B. Mayer, essentially for being too smashed, on and off the set, to work. Keaton's M-G-M experience, despite various efforts by Thalberg and others to keep his career alive as a gag writer, ruined his art. The next decades are truly painful to read about, as Keaton went in and out of hospitals and clinics, falling off the wagon and then sobering up again. His brother-in-law, the cartoonist Walt Kelly, recalls that "nobody really wanted to put him under control because he was a lot of fun." 

Evanier was co-editor of the Complete Pogo and was in a long-term relationship with Kelly's daughter and this was the first he had heard of the connection. He discussed the matter with other experts on both Kelly and Keaton and finally got the answer from the man who literally wrote the book.

So we became intrigued about this mystery. In the latest New Yorker, author Adam Gopnik reviewed two new books about Buster Keaton and offered a quote about Keaton from — and I quote: "His brother-in-law, the cartoonist Walt Kelly." That would seem to be the cartoonist Walt Kelly who created and drew my favorite comic strip (and yours if you have a lick o' sense), Pogo.

Walt Kelly scholars — like, say, me — were amazed at the claim that Kelly was Keaton's brother-in-law. So were folks who knew Walt personally — like, say, his son Peter. We all began puzzling and puzzling 'til our puzzlers were sore…and I'll single out a couple of folks who went to work on this mystery and came up with some solid clues: Maggie Thompson, Harry McCracken and Mike Whybark.

...

And here's an e-mail that I received this morning from James Curtis…

I think I can clarify the matter regarding Buster Keaton and Walt Kelly, but only somewhat.

I am the author of the upcoming book Adam Gopnik was referencing, and Buster Keaton did indeed have a brother-in-law named Walt Kelly. But why Gopnik made the completely unnecessary assumption that the Walt Kelly who was married to Eleanor Keaton’s younger sister Jane was the same Walt Kelly who created Pogo is beyond me. As you know, that Walt Kelly died in 1973. As of two months ago, the Walt Kelly I interviewed was still alive and living in Southern California. He is certainly not identified as the other Walt Kelly in my book.

The review appeared on line yesterday morning, and Leonard Maltin wrote to congratulate me. I mentioned this odd situation of confusing a career military officer with a world-renowned cartoonist. He said: "So much for fact-checking!" I also heard from a gentleman in Seattle who was writing on behalf of a Facebook group called "I Go Pogo" asking if I could throw some light on the matter. I told him what I knew, and he thanked me for the clarification. "It’s sort of amusing," he commented, "even if it's disappointing to read it from a pen of such a high caliber. I've been pondering this all day…"

I wonder how many others have been pondering it as well.

I checked out the Gopnik review and.... let's just say there are enough issues for another post.

And as for that passage, here's how it looks online two and a half years later. 

 

In case you skipped the news over your three-day weekend

 WASHINGTON, July 5 (Reuters) - The dispute between Republican President Donald Trump and his main campaign financier Elon Musk took another fractious turn on Saturday when the space and automotive billionaire announced the formation of a new political party, saying Trump's "big, beautiful" tax bill would bankrupt America.

A day after asking his followers on his X platform whether a new U.S. political party should be created, Musk declared in a post on Saturday that "Today, the America Party is formed to give you back your freedom." 

 

New details continue to trickle in about the origins of the feud, almost all of which support what we already suspected. Most notably, sources within the White House have confirmed that “we dropped the dime” on Musk’s drug use. There was already overwhelming circumstantial evidence that the story had to come from within the administration—the NYT’s claim to have extensive photographic and video documentation pretty much narrowed it down—but this removes what little doubt there was. We’ll probably never know for certain if Trump himself played a role here—I’m inclined to suspect he didn’t—but we’ve passed the point where that matters.

After the initial rage faded, Musk did make an attempt to mend fences, taking down his original tweets and making various conciliatory and submissive comments, but the damage was done. Now, with the stress amping up around Tesla’s increasingly controversial robotaxi launch and dismal numbers coming out, Elon is back on the offensive—now promising to start the America Party (the American Party name was already taken by George Wallace).

 

 

 

 As far as I can tell, no one—including Elon Musk himself—is pitching this as a truly competitive third party. His idea seems to be to get enough support to be a kingmaker, perhaps even grabbing a few seats in the House or possibly the Senate, allowing him to have the deciding vote on legislation. He could certainly play the spoiler—albeit asymmetrically.

Having lost his WH connection, Musk's political power comes from:

1. Money
2. Fan base
3. Control of Twitter

He's toxic with Dems and independents. 2. and 3. are mainly made up of Trump supporters. I'm sure he'd like to poach supporters from the left, but I don't see that as likely.

Something else you need to keep in mind is that, as we've often mentioned, among the super-rich, Elon Musk’s fortune is uniquely precarious. It depends on Tesla continuing to trade at 20 or 30 times the valuation the fundamentals suggest. We know that Musk has used some of his stock in the company to secure loans, which means a big drop would trigger margin calls—leading to all sorts of trouble for the world's (currently) richest man. 

Trump is in a position to destroy his opponent’s fortune simply by aggressively enforcing laws that Tesla has been given a pass on for the past few years. To further complicate matters, there's this:

 

From an election forecasting standpoint, we continue to sail into that part of the map marked only by the warning: Here there be monsters. Anything like observed data has long since disappeared over the horizon.

Here are some of the unknowns foolhardy modelers will have to contend with:

    As far as I can tell, there is no precedent for a Donald Trump in American politics. We've had far-right demagogues before, but we've never had one reach this high an office or hold this kind of absolute power over his own party.

    There is no precedent for Elon Musk—neither in terms of wealth nor political extremism. (Yes, I know about Ford, the Wall Street Putsch, and all the rest, but no politically driven plutocrat has ever reached the level we see here.)

    Third parties inevitably introduce a huge degree of uncertainty into the mix.

    Both Trump and Musk alternate between being motivated by anger and catharsis one day and by naked self-interest the next. In neither case does principle come into play. It’s entirely possible that both men will realize the damage they're doing to themselves and call a real truce—or they could just continue to escalate.

    Musk, in particular, has a long history of quietly backing away from big promises. It’s entirely possible that will happen here. We also have to contend with the possibility that Musk will not have hundreds of millions of dollars in 18 months—or at least not hundreds of millions that he feels free to blow on a quixotic cause.

All of this is on top of the chaos and unpredictability that has increasingly defined 2025.

As the old (and probably apocryphal) curse goes: May you live in interesting times.