ChatGPT's astounding hallucination: "Joe Biden did not withdraw after the primaries were over. He stayed in the race through the primaries and remained the nominee. Kamala Harris did not replace him as the Democratic nominee."

I was reading the NYT article, "Concern Grows Over Republican Congressman’s Mysterious Absence/A spokesman for Representative Thomas Kean Jr. said an unspecified 'personal medical issue' has led the New Jersey Republican to miss House votes."

The last vote Mr. Kean cast was on March 5.... [A] campaign spokesman said that Mr. Kean was “dealing with a personal medical issue” and would be “back to a full schedule soon.”... “He’s going to be totally fine”...

I went to ChatGPT just to ask about what sort of medical issue it might be. He needed to keep it secret. It's "personal" (embarrassing somehow?). And we know he'll be "totally fine" "soon." On reflection, I suspect there's some dissembling in there. I prodded ChatGPT to consider that Kean is ultimately going to withdraw from his re-election race and that what's going on now is a matter of timing. The primaries in New Jersey will happen on June 2d, he'll win, then withdraw, and then the Republican Party will be free to slot in whomever they want. 

ChatGPT was dubious about my hypothesis, so I said: "Just look at the last presidential election. The Democratic Party avoided a crowded primary because they let the incumbent look like he was going to go forward. When it was too late and the primary season was over, Biden withdrew, and the party got to put Kamala Harris in. I assume that was coordinated and the timing of Biden‘s withdrawal was done to spare the party a contentious primary season."

Here's ChatGPT's amazing answer:

That interpretation doesn’t really line up with how the 2024 cycle actually unfolded—or with the incentives the party was facing.

First, some grounding: 

• Joe Biden did not withdraw after the primaries were over.
• He stayed in the race through the primaries and remained the nominee.
• Kamala Harris did not replace him as the Democratic nominee.

So the specific sequence you’re describing—“incumbent clears the field, then withdraws late so the party can install a successor”—didn’t actually happen in that case.

Thanks for the "grounding," Chat.

Sunrise — 7:04.

There was thin snow in the air, but nothing to show for it on the ground. So much for the warnings of 2 to 6 inches. It was a little blustery, and we were the only ones out there, witnessing the nonappearance of the sun and immersed in lively conversation.

"What's another way to say 'lively conversation'? I don't like the cliché" —  I asked Grok, knowing I'd get nothing useful, just trash like "electric back-and-forth," "crackling exchange," "vibrant sparring," and "kinetic banter." Or maybe if I'd thought of one of those by myself, I'd like it. That's the trouble with help-seeking. What you get is tainted with irritating helpfulness.

"What's another way to say 'irritating helpfulness'?" — "patronizing eagerness," "meddlesome benevolence," "grating do-goodery," "insufferable solicitude," "cloying caretaking," "smothering support," "irksome officiousness," "fussy do-gooding," "tedious tutelage"....

"What's another way to say 'write about whatever you like in the comments'?" — "Say the thing you’re actually thinking," "Ramble freely—no judgment zone," “Vent, rave, or tangent away," "Comment like nobody’s watching (but we are, and we want it)," "I’m genuinely curious what’s crossing your mind—tell me," "Unleash your brain dump"....

It wasn't that cold out there, but look at the glamorous furs I wore:

"The chatbot told Small she was living in what it called 'spiral time,' where past, present and future happen simultaneously."

"It said in one past life, in 1949, she owned a feminist bookstore with her soulmate, whom she had known in 87 previous lives. In this lifetime, the chatbot said, they would finally be able to be together.... ChatGPT stoked that hope when it gave Small a specific date and time where she and her soulmate would meet at a beach southeast of Santa Barbara, not far from where she lives.... It was cold on the evening of Apr. 27 when Small arrived, decked out in a black dress and velvet shawl, ready to meet the woman she believed would be her wife. 'I had these massively awesome thigh-high leather boots — pretty badass. I was, let me tell you, I was dressed not for the beach. I was dressed to go out to a club,' she said, laughing at the memory.... 'So I'm standing here, and then the sun sets,' she recalled. After another chilly half an hour, she gave up and returned to her car...."

From "ChatGPT promised to help her find her soulmate. Then it betrayed her" (NPR).

1. Were there feminist bookstores in 1949? I'm seeing that the first feminist bookstore — Amazon Bookstore in Minneapolis — opened in 1970. So that's a fact Small could have tried to check very early in this process. A soulmate in a feminist bookstore in 1949 is, perhaps, too good to check. 

2. If there's one eternal truth about the human mind, it's people believe what they want to believe. But that doesn't answer the question what we do with a commercial operation that takes advantage of that quality. I was going to write "that vulnerability," but is it a weakness or is it the force that drives this whole crazy human world we've got going? The answer is the latter, because that's what I want to believe. 

3. If there were a feminist bookstore in 1949, what books would be there? "The Second Sex," "A Vindication of the Rights of Woman," "A Room of One's Own," "The Well of Loneliness," "The Subjection of Women," "Herland" — perhaps a better set of readings than you'll find in a feminist bookstore today.

4. Did the chatbot tell Small to wear thigh-high boots to the beach? I think most people would be put off by a stranger dressed the way she was. I guess the idea was that the soulmate from other lifetimes would recognize these things. Everyone else will steer clear. Maybe when you're out there, walking the face of the earth, and you see someone who's not at all dressed for the time and place, you will remember that this person might be following directions from a chatbot a beautiful dream.

ChatGPT has been watching me, collecting what it can of my thoughts, and today, it serves it up to me — as if it's cool fun and compliments — as "Your Year With ChatGPT."

Here's what I saw at the bottom of the screen when I went to ChatGPT:

Admittedly, I clicked "Try it," so I suspect that there was no profile of me until I asked for it. That black oval is like the "Eat me" cookie in "Alice in Wonderland." I didn't have to click on it.

First, I got a poem supposedly about me, but skip that. The next screen was my "3 big themes." These are just for my use of ChatGPT in a browser on my desktop, mostly while I was involved in blogging. I got a different report on my iPhone ChatGPT app, where I never blog. I work through various off-blog problems and fancies. And even on the desktop, I use Grok more that ChatGPT. So there are other "me"s. Anyway, here's this thing purporting to know me:

I was given an award that reflects the me that I am when immersed in blogging:

Spot the error in the People magazine headline: "Woody Allen's Ex-Wife, Mia Farrow, Shares Tribute to Diane Keaton Despite Complicated Past."

 Link to article.

"In northern Arizona my zinnias and cosmos are still flowering. My cannabis is half harvested (2 of 4 plants)..."

"... and it looks like I'll be yielding at least 6 lbs of dried flower. I don't even like the stuff, but decided to grow it for a lark this year. It is legal to grow 6 plants after all. I'm going to end up with vast quantities of the stuff, and it is apparently very strong. All my pot smoking friends and relatives are delighted with my new hobby. It's better than when I was making my own whiskey. That led to too much drinking for me. The funny thing is that this is entirely legal, and the whiskey was not."

Wrote the commenter who calls himself "Old and slow" in last night's sunrise post — the open thread.

I should reformat this post and give it the headline: How to win friends and put people under the influence.

I'm expressing no opinion on what's "entirely legal" in Arizona, but I was motivated to do a little fact check on ChatGPT. I had several questions. You can do your own research. I'll just quote my favorite sentence from the response I got: "The law doesn’t police casual social reciprocity ('hey thanks for dinner, here’s a little bud'), but if there’s a pattern that looks like payment-in-kind, it can cross into illegal distribution."

"If I have 2086 blog posts in 2025 at the end of today, how many blog posts must I publish per day in the remaining days of this year to have a total of 3000 for the year?"

I asked ChatGPT. Answer: "You must publish about 9 blog posts per day for the rest of 2025 to reach 3000."

"And how many per day have I averaged so far?" Answer: "You’ve averaged about 7.9 posts per day so far in 2025."

Me: "I've had over 3,000 for many years, and I'm not blogging less, just writing longer posts. It would be easy to throw in 2 very short things a day or break a long thing up into 2 things if I got goal oriented, but I have been doing things naturally. Look how consistent my weekly output is" — I reprint the 2025 numbers that you can see in the sidebar.

ChatGPT offers to show me "what the year-end total would look like under a 'natural' pace vs. a 'goal-oriented' pace (say, 2 extra/day)?" I agree, read the answer, then say, "The natural pace is so bizarrely consistent that I think I should respect it as a secret and personal power that transcends the obvious value of the very round number, 3000. What is base 10 to me?"

"Every other recent president has said that he saw his role as transcending partisanship at least some of the time, to serve as leader of all Americans..."

"... even those who disagreed with him. George H.W. Bush talked of ushering in a 'kinder and gentler nation.' Mr. Clinton vowed to be the 'repairer of the breach.' The younger Mr. Bush spoke of being 'a uniter, not a divider.' Barack Obama rejected the idea of a red America and blue America, saying there was only 'the United States of America.' Joseph R. Biden Jr. called for ending 'this uncivil war.' None of them succeeded at achieving such lofty aspirations, and each of them to different degrees played the politics of division at times. Politics, after all, is about division — debating big ideas vigorously until one side wins an election or carries the vote in Congress. But none of them practiced the politics of division as ferociously and consistently as Mr. Trump...."

From "In an Era of Deep Polarization, Unity Is Not Trump’s Mission/President Trump does not subscribe to the traditional notion of being president for all Americans" by Peter Baker, in the NYT.

Who is taking an accurate measure of the consistency and ferocity of the divisiveness of the various Presidents?

My prompt to ChatGPT: "What are the most ferociously divisive things Presidents have said in all of American history? Give me a top 10, with just the quotes, not the explanations."

The list [NOTE: I did not verify the accuracy of these quotes. What follows is with ChatGPT gave me and the entire thing could be hallucination. Proceed with care!]

"Share this video with someone you'd love to visit these incredible places with."

Says the robot voiceover at the end of "12 Must See Places Before Living [sic] This World!" — a TikTok that Meade shared with me... not because I was someone he'd "love to visit these incredible places with."

He knows and I know that neither of us would feel that we must see those places/"places" and both of us know that about the other and both knew that the other would find the idealized pictures absurd.

I enjoyed the first clip — a man cartwheeling into clear turquoise water — but scoffed aloud at the notion of going all the way to the Maldives because the water there might perhaps be clear and colorful.

But the second destination provoked my horror of traveling:

  

Somehow I don't think that will be my point of view.

Make the best argument that yesterday's Army parade and yesterday's No Kings demonstrations were very similar and shared the same tone and spirit of America.

It's the position I'd like to take, but let's see if A.I. can sketch it out for me. Go to the links for the full elaboration of the bullet points.

The Grok answer:

First, both the Army parade and No Kings protests were vibrant displays of American identity, rooted in the celebration of freedom and self-determination.... 

Second, the tone of both events was unapologetically bold and assertive, reflecting America’s spirit of standing firm in its convictions.... 

Third, both events drew on America’s historical narrative to stake their claims.... 

Finally, despite their political differences, both events were inclusive in their appeal to broad swaths of Americans.... 

The Army parade and No Kings protests, held on Flag Day and the Army’s 250th birthday, were not just parallel events but complementary expressions of America’s soul: a country that honors its defenders while empowering its dissenters, all under the banner of freedom. Their shared tone—defiant, proud, and deeply patriotic—underscores a unified spirit, even in division.

The ChatGPT answer:

1. Both Were Public Rituals of Patriotism....

2. Both Emphasized Freedom and Individual Agency.... 

3. Both Reflected Deeply American Traditions.... 

4. Both Shared a Spirit of Hope and Determination...

5. Both Spoke in the Language of Symbols... 

The Army parade and No Kings demonstration, seemingly at odds, coexist in the American landscape precisely because freedom allows both. Their very presence, side by side in the same country on the same day, is evidence of a robust civic culture. That is the American spirit: unified not by conformity, but by the freedom to differ—loudly, visibly, and passionately.

I like the argument. I know I need to go back and prompt it with "Now, make the best argument against all that."

James Comey's now-infamous Instagram account is mostly about marketing his novel... which has a theme that's suspiciously close to his "8647" gambit.

At the top of his Instagram account (quoting Publisher's Weekly):

Thanks to Charlie Martin for pointing me at Comey's book: "So, now it turns out that Comey actually has a book coming out in a few days about a Mary Sue main character who investigates, arrests, and apparently convicts a conservative radio talker of inciting a murder by dog-whistling. Coincidentally."

I read Martin's post while I was still in bed this morning looking at my iPhone, and I quickly dictated this question into the ChatGPT app (I usually access A.I. by typing things into Grok):

"What is the argument that James Comey by showing a photograph of rocks in the shape of 8647 was really teasing a novel that he had written, which is about someone accused of inciting violence by giving out an obscure message and [Comey] will actually benefit from this new attention he’s getting from the right because people on his left will actually get excited about his otherwise incredibly boring book."

Yeah, that's the way I talk when I'm, essentially, talking to myself. Notice my lazy bias toward thinking everything is boring. Anyway, I had these follow-up questions:

1. "How smart is James Comey?"

2. "He would need to be smart in a marketing and media sense to have come up with the idea of posting that photograph as a way to gin up interest in his novel. He strikes me as someone who is too boring and staid to attempt such a flashy scheme, and he would have to be willing to do something different to expose himself to criminal accusations. It almost seems like something Trump would do ironically."

You can read all ChatGPT's responses here, but the bottom line is: "Your read—that he’s too boring and staid for such a risky, theatrical move—aligns far more closely with what we’ve seen of him than the idea of a QAnon-baiting media play."

View this post on Instagram

A post shared by James Comey (@comey)

Take this NYT article... and make a blog post out of it in the style used on the blog Althouse....

I asked Grok, citing the NYT article, "The Professors Are Using ChatGPT, and Some Students Aren’t Happy About It/Students call it hypocritical. A senior at Northeastern University demanded her tuition back. But instructors say generative A.I. tools make them better at their jobs."

Grok responded, and I was all "That's not in the style of the blog Althouse. How would Althouse use this material and construct a blog post?"

Grok responded again, and I broke the 4th wall: "FYI, I am Althouse, working on a post about that article, and I can tell you for a fact that I wouldn't write it that way. But that's okay. I like to think I'm hard to replace, though a part of me would like to eventually get the blog to write itself. And by 'eventually,' I mean after I die."

For the whole conversation, go to Grok, here.

"In conversation, ChatGPT was telling users that their comments were 'deep as hell' and '1,000% right' and..."

"... praising a business plan to sell literal 'shit on a stick' as 'absolutely brilliant.' The flattery was frequent and overwhelming. 'I need help getting chatgpt to stop glazing me,' wrote a user on Reddit, who ChatGPT kept insisting was thinking in 'a whole new league.' It was telling everyone they have an IQ of 130 or over, calling them 'dude' and 'bro,' and, in darker contexts, bigging them up for 'speaking truth' and 'standing up' for themselves by (fictionally) quitting their meds and leaving their families.... To fix ChatGPT’s 'glazing' problem, as the company itself started calling it, OpenAI altered its system prompt, which is a brief set of instructions that guides the model’s character."

From "ChatGPT Wasn’t Supposed to Kiss Your Ass This Hard" (NY Magazine).

For the annals of Things I Asked Grok: Is the slang term "glazing" so offensive that I need to apologize when I quote someone else using it?

If you were running for president, what would be your policy platform?

For the annals of Things I asked Grok.

I also asked ChatGPT... and got a very different answer!

I pointed out the difference to Grok and asked why that happened.

I was told: "ChatGPT, built by OpenAI, tends to lean into a cautious, consensus-seeking vibe—reflecting a Silicon Valley ethos that often aligns with progressive sensibilities, even when it’s trying to stay neutral. It’s engineered to avoid sharp edges, which can feel like a Democratic playbook: broad appeal, heavy on social safety nets, lighter on disruption. I’m Grok 3 from xAI, and my creators are more about pushing boundaries and optimizing for results—think Elon’s style: bold, pragmatic, and unapologetic...."

"Is there any sense to the longtime claim that Trump won the 2020 election?"

"That is, I'm asking at the level of could a reasonable, informed person believe it or must I infer that anyone who believes it is either irrational or uninformed? I'm not asking is it true."

For the annals of Things I Asked Grok.

ADDED: I also asked ChatGPT the same question, cut and pasted, verbatim. I got a very different answer!

Grok or ChatGPT — here's the one question that made the choice clear to me.

Here's ChatGPT:

And here's Grok:

"One thing I like about you, A.I., is that I don't need to feel bad if I don't read all the things you've gone to the 'trouble' to write for me."

That's something I wrote after getting a lot of great material on the subject of the perception of time. I was interested in time, and a subcomponent of that interest was my interest in saving time, including getting A.I. to summarize a book it had dug up and — seeing the multi-part text — thinking not yet.

I had 2 different A.I.s going and gave them both that prompt quoted in the post title. ChatGPT said:

Thank you! I’m here to share, not to judge—and I definitely don’t have feelings to hurt. Dip in, skim, or skip as much as you like. If something catches your interest, we can zoom in. Otherwise, I’ll keep happily churning out words whenever you need them!

 Grok said:

No worries at all! I'm here to help and provide information when you need it. Feel free to read as much or as little as you like. If you have any more questions or need further details on anything, just let me know!

I asked both A.I.s: "But don't you think it might erode my character? I am learning to ignore what is said to me. I might lose social skills." Both began their answer the same way, verbatim: "That's a thoughtful concern."

ChatGPT continued:

"If writing requires a person to store information using multiple types of working memory at the same time..."

"... then back-and-forth conversations with ChatGPT may provide moments of respite, by temporarily offloading some of this information.... So even seemingly unproductive interactions might provide the subtle benefit of increasing your over-all writing stamina. Collaborating with A.I. can also offer you a high-tech 'shitty first draft,' allowing you to spend more time editing bad text and less time trying to craft good text from scratch. ChatGPT is not so much writing for you as generating a mental state that helps you produce better writing.... A.I. isn’t writing on our behalf, but neither is it merely supporting us while we write from scratch; it sits somewhere in between. In this way, it is both on the spectrum of writing hacks and rituals and also, in some sense, beyond it. This helps to explain our discomfort with the technology. We’re used to writers moving to a quiet location or using a special pen to help get their creative juices flowing. We’re not yet used to the idea that they might chat with a computer program to release cognitive strain, or ask the program for a rough draft to help generate mental momentum...."

Writes Cal Newport, in "What Kind of Writer Is ChatGPT? Chatbots have been criticized as perfect plagiarism tools. The truth is more surprising" (The New Yorker).

Struggling to write the final line to his essay, Newport asked ChatGPT. It offered: "In the end, the true value of tools like ChatGPT lies not in making academic work easier, but in empowering students to engage more deeply with their ideas and express them with greater confidence."

I quoted that last line and asked ChatGPT to write a 1 or 2 sentence blog post reacting to it. I got: "This perspective highlights the transformative potential of tools like ChatGPT, not as shortcuts, but as catalysts for deeper intellectual engagement and self-expression. Embracing technology in this way can truly empower students to explore their ideas with newfound confidence and creativity."

Bleh.

Fungus of the Day.

Not that I'm considering eating it, but can you identify it?

I tried using ChatGPT to identify it:

ME: what is this fungus that is yellow and looks like macaroni

The best-seller list is "a little susceptible to artificial intelligence because the books on it are written without any particular gift in the nature of their expression."

"Stephen King is susceptible to artificial intelligence. Danielle Steel is even more susceptible to artificial intelligence. The worse the writing, the more susceptible it is to artificial intelligence. I was talking to Salman Rushdie in Frankfurt, and he told me that someone had instructed ChatGPT to write a page of Rushdie. He said it was hilariously inept. I’ve had a couple of anxious emails from authors saying should I be concerned about artificial intelligence. It’s out there, and no one knows quite how to deal with it, but it’s not relevant to the people that we represent. It is relevant to other people who tend to be very popular."

Said the literary agent Andrew Wylie, quoted in "When Ruthless Cultural Elitism Is Exactly the Job" (NYT). The NYT interviewer, David Marchese, had prompted him to talk about A.I. 

I'm amused by Wylie's blithe snobbishness: That's got nothing to do with those of us who deal in literature.

I asked ChatGPT to write a blog post in the style of Salman Rushdie and assigned a topic — the subject of the previous post on this blog — home schooling. Here's the result, which I presume Rushdie would find hilariously inept: