As I was idly browsing Bluesky this evening, I spotted some terrific words of advice on the role of LLMs within study.

They were written by Robert McNees, associate professor of Physics at Loyola University Chicago, who has kindly given me permission to share them here. Robert stressed it shouldn't be taken as the the college's policy, rather as a discussion starter for his own students.

The paragraph in bold is particularly important, I think, and it offers a fine framework for how journalists should be considering their own AI use at work, too. You can find Robert on Bluesky here.

Should I Use ChatGPT Or Another LLM To Study?

I wouldn’t recommend it. I try to keep up with the capabilities of the major LLMs. They can do some things really well, if you use them the right way. However, they frequently make mistakes when generating responses to questions about physics. Sometimes these mistakes are obvious, sometimes they are subtle and hard to spot. The fact that you cannot trust the output of LLMs should be reason enough not to rely on these systems when you are trying to learn a new subject.

But that’s not the only problem. Interactions with LLMs feel like a dialog, so it’s natural to think the usual rules of conversation apply. You ask a question and expect the response will be an answer to that question. It’s important to understand that this is not what’s happening. An LLM is designed to generate statistically likely responses to the question “What would an answer to this query sound like?” This is not the same thing as answering the question. It might produce what you are looking for, or it might not. This is one reason why output from an LLM will sound authoritative even when it’s wrong, and apologetic when mistakes are pointed out. It isn’t authoritative or apologetic, and it isn’t “thinking” about your question. These are just the sorts of responses that best fit a very complicated set of likelihood criteria.

A bigger problem is that using an LLM short circuits the process of thinking through questions and developing strategies to answer them. It’s not that an LLM never gets things right; they often produce correct output. But correct outputs are limited to material in the model’s training data — questions we already know how to answer. Is that why you’re here? To answer questions we already know how to answer? Whether you are studying Physics or English or Business, all your instructors are trying to help you learn how to answer questions for yourself. Part of that training involves questions we already understand, because that’s an effective way of learning processes that can be applied to questions we don’t understand. This is one of the most important aspects of your college education and it takes practice. Asking an LLM may or may not generate a correct answer, but either way it prevents you from practicing and learning these processes.

To make matters worse, there is now research claiming that frequent use of LLMs has neurological and behavioral consequences. One recent study find significant cognitive debt and consistent underperformance compared to peers that do not rely on these systems. That is a steep price for a momentary convenience. So I can’t stop you from using an LLM, but I would urge you to consider the long term cost.

Joanna Stern installs an AI-powered vending machine in the Wall Street Journal's newsroom:

"Within days, Claudius had given away nearly all its inventory for free—including a PlayStation 5 it had been talked into buying for “marketing purposes.” It ordered a live fish. It offered to buy stun guns, pepper spray, cigarettes and underwear."

READ: We Let AI Run Our Office Vending Machine. It Lost Hundreds of Dollars.

A devastating read in ProPublica:

"Mothers had to choose which of their kids to feed. Young men took to the streets in protests, some of which devolved into violent riots. Pregnant women with life-threatening anemia were so desperate for calories that they ate mud. Out of options and mortally afraid, refugees began fleeing the camp by foot and in overcramped cars, threatening a new migration crisis on the continent. They said they’d rather risk being shot or dying on the perilous route than slowly starving in Kakuma."

Read it all: Inside the Trump Administration’s Man-Made Hunger Crisis

Federal prosecutors hung their case on a piece of onion and a mustard smear, and that the Customs and Border Protection agent struck by a foot-long Subway sandwich was the victim of an indefensible attack.

But when the jury began deliberations, they focused on the videos and photographs shown during two days of testimony: the sandwich, they noted, was never unwrapped. So how could any of the sandwich’s insides have struck the agent?

The agent kept two souvenirs he received from fellow agents teasing him after the incident. The jurors wondered: What true victim keeps a memento of their assault?

“It really came down to this whole thing was a bunch of baloney,” said one of the jurors, a 26-year-old Washington architectural designer.

-- BGov's Keith Alexander reminds us that sometimes the funniest possible stories are court reports written dead straight. Read: Jurors in Sandwich Thrower Case Found Charges ‘Bunch of Baloney’

"I don’t know how much competitive Excel you’ve watched, but it is kind of hard to follow what’s going on. The thing I always say is, like, hundreds of millions of people watch baseball, and that’s not very interesting either."

-- Michael Jarman, financial modeler and 2024 Microsoft Excel world champion. Businessweek: Why We Can't Quit Excel

"Writers are tearing their hair out. Bond didn't just vanish off a cliff or fake his death – he was blown to pieces on screen. Everyone agrees it was a massive mistake because Bond is supposed to be eternal. They are now stuck trying to find a believable way to resurrect him, and it is proving almost impossible."

-- an unnamed writer, speaking to RadarOnline, highlights a significant plot headache for the next installment of James Bond: He's dead.

"Unable to get a precise crowd estimate, I tried instead to count inflatable frog costumes. I gave up on this about twenty minutes later: there were simply too many frogs."

• The Verge's Sarah Jeong covering the nationwide 'No Kings' protest in Portland, Oregon.

Aristotle taught that all human beings want to know. Our own experience teaches us that all human beings also want not to know, sometimes fiercely so. This has always been true, but there are certain historical periods when the denial of evident truths seems to be gaining the upper hand, as if some psychological virus were spreading by unknown means, the antidote suddenly powerless. This is one of those periods.

-- Columbia professor Mark Lilla in the New York Times: The Surprising Allure of Ignorance.

Over the course of the 2024 season, the White Sox have explored the full spectrum of losing the way a great actor uses every corner of the stage, the way a jazz saxophonist probes every note in a scale. They have lost nobly, tragically, cleverly, inspiringly and deflatingly. They have lost late at night and early in the afternoon, in soggy rain and on crisp sunny days. I have seen perfectly professional losses that could have gone either way — but of course didn’t — and games that should have been stopped, for cruelty, in the fourth inning. I have seen the White Sox lose in front of huge roaring crowds at Fenway Park and also, back home, in their own nearly empty stadium. (On a sunny Tuesday, just before game time, I once counted 199 people sitting in the vast sea of outfield seats — and when the announcer finally said “Play ball!” the applause sounded like someone had just done a magic trick at a church picnic.) I have seen the White Sox hit their catcher in the groin with the baseball three separate times in a single inning. I have seen the White Sox lose because three fielders ran into each other like clowns. I have watched a bloop single flutter and fall, like the first leaf of autumn, delicately onto the outfield grass, at the most devastating possible moment. I have seen games in which Chicago’s hitters looked like All-Stars but their pitchers looked like impostors, and games where it was vice versa, and games in which they all played great but the ball just bounced the wrong way.

What is it about baseball -- or more specifically, losing at baseball -- that provokes such wonderful writing?

Sam Anderson in the New York Times: How Does a Baseball Team Lose 120 Games? Every Way You Can Think Of.

For New York Magazine, Clio Chang investigates the dripping water on the subway:

Next, we move downstairs to the platform. Bostick tells me that the deeper you are, the nastier the water may be, since it has more time to pick up gross stuff as it travels through the bowels of the subway infrastructure. Here, we find a puddle on the ground. We decide to take a sample, even though it wasn’t technically part of our inquiry, and Bostick sucks it up with a syringe. We agree that this stale floor liquid, with its unsettlingly yellow hue, is much nastier than the drips. “Don’t drink this,” Bostick says, as if the thought had remotely crossed my mind.