Erin Alberty

Utah news reporter for @axios. Gardener, word puzzler, mom. Happiest in the mountains. @PulitzerPrizes winner.

[Dionne Warwick @dionnewarwick]

It looks like you may have chosen two photos of me for your twote @PopBase

[Erin Alberty @erinalberty]

@PBTBrider There are other alternatives to turfgrass. Not just rock.

[NBC News @NBCNews]

A network of fake academic journals masquerading as legitimate publications has published more than a hundred AI-generated papers in recent months, in some cases using the names of real professors at top universities without their knowledge. https://t.co/9OuA7Erj0f

[James Anderson White @jamesawhiite]

It’s imperative that we revert to STANDARD, not Daylight time, for two reasons, one of which you may have never considered.

You know the first: solar noon is closest to 12 o’clock all year. That means an equal number of hours before and after noon each day.

This also means earlier sunrises which are significantly better for establishing a healthy circadian rhythm.

The second reason is that nearly ONE THIRD of US counties are in the wrong timezone.

A significant swath of the country isn’t only an artificial hour ahead because of daylight time… but TWO HOURS ahead of true solar time.

Year-round daylight time, especially without recalibrating time zones, would establish this error permanently, affecting millions of Americans.

We should be letting the Sun dictate the day the way God made it. Arbitrary, artificial tampering would be worse for us all.

Fix the time zones and affix standard time.

[Erin Alberty @erinalberty]

@jkimballcook Also: I thought bolo ties were formal? My grandpa, a conservative banker in rural South Dakota, wore them to dressy stuff all the time.

I wrinkled my nose as soon as I saw the idea of high/low "standards" attached to cultural norms that aren't even unanimous within the U.S.

[𝓔𝓿𝓪𝓷 @Baltlol]

if I woke up to a pony playing a keyboard after surgery I would assume I passed away

[Erin Alberty @erinalberty]

Kevin O'Leary sits behind the Utah governor's desk in a photo that was posted in early January — months before most of the public learned of his massive data center proposal in Box Elder County. https://t.co/Bh4GdC8Vrk

[Erin Alberty @erinalberty]

@NHarding97 @MetamateDaz im less worried about the dead husband than the adult granddaughter who appears to be providing in-home nursing care for a centenarian who could easily afford professional services 😅😭

[Erin Alberty @erinalberty]

@alisha_ann94 @ItsMrsPlugg My parents were VERY strict, but no way were they going to sit through Wayne's World with a theater full of junior high kids.

Mine is 13 now, and it didn't occur to me that dropping her off might be lax. But she hasn't asked & wouldn't want to pay for her own tickets. 😂

[Nav Toor @heynavtoor]

🚨SHOCKING: Apple just proved that AI models cannot do math. Not advanced math. Grade school math. The kind a 10-year-old solves.

And the way they proved it is devastating.

Apple researchers took the most popular math benchmark in AI — GSM8K, a set of grade-school math problems — and made one change. They swapped the numbers. Same problem. Same logic. Same steps. Different numbers.

Every model's performance dropped. Every single one. 25 state-of-the-art models tested.

But that wasn't the real experiment.

The real experiment broke everything.

They added one sentence to a math problem. One sentence that is completely irrelevant to the answer. It has nothing to do with the math. A human would read it and ignore it instantly.

Here's the actual example from the paper:

"Oliver picks 44 kiwis on Friday. Then he picks 58 kiwis on Saturday. On Sunday, he picks double the number of kiwis he did on Friday, but five of them were a bit smaller than average. How many kiwis does Oliver have?"

The correct answer is 190. The size of the kiwis has nothing to do with the count.

A 10-year-old would ignore "five of them were a bit smaller" because it's obviously irrelevant. It doesn't change how many kiwis there are.

But o1-mini, OpenAI's reasoning model, subtracted 5. It got 185.

Llama did the same thing. Subtracted 5. Got 185.

They didn't reason through the problem. They saw the number 5, saw a sentence that sounded like it mattered, and blindly turned it into a subtraction.

The models do not understand what subtraction means. They see a pattern that looks like subtraction and apply it. That is all.

Apple tested this across all models. They call the dataset "GSM-NoOp" — as in, the added clause is a no-operation. It does nothing. It changes nothing.

The results are catastrophic.

Phi-3-mini dropped over 65%. More than half of its "math ability" vanished from one irrelevant sentence.

GPT-4o dropped from 94.9% to 63.1%.

o1-mini dropped from 94.5% to 66.0%.

o1-preview, OpenAI's most advanced reasoning model at the time, dropped from 92.7% to 77.4%.

Even giving the models 8 examples of the exact same question beforehand, with the correct solution shown each time, barely helped. The models still fell for the irrelevant clause.

This means it's not a prompting problem. It's not a context problem. It's structural.

The Apple researchers also found that models convert words into math operations without understanding what those words mean. They see the word "discount" and multiply. They see a number near the word "smaller" and subtract. Regardless of whether it makes any sense.

The paper's exact words: "current LLMs are not capable of genuine logical reasoning; instead, they attempt to replicate the reasoning steps observed in their training data."

And: "LLMs likely perform a form of probabilistic pattern-matching and searching to find closest seen data during training without proper understanding of concepts."

They also tested what happens when you increase the number of steps in a problem. Performance didn't just decrease. The rate of decrease accelerated. Adding two extra clauses to a problem dropped Gemma2-9b from 84.4% to 41.8%. Phi-3.5-mini from 87.6% to 44.8%. The more thinking required, the more the models collapse.

A real reasoner would slow down and work through it. These models don't slow down. They pattern-match. And when the pattern becomes complex enough, they crash.

This paper was published at ICLR 2025, one of the most prestigious AI conferences in the world.

You are using AI to help you make financial decisions. To check legal documents. To solve problems at work. To help your children with homework. And Apple just proved that the AI is not thinking about any of it. It is pattern matching. And the moment something unexpected shows up in your question, it breaks. It does not tell you it broke. It just quietly gives you the wrong answer with full confidence.

[CAPTAIN CANADA @superglowX]

Guy who is convinced that the social media popularity of Taylor Swift and LeBron was astroturfed but that Gunther Eagleman and Catturd are now the most engaged accounts by sole merit.

[Aaron Rupar @atrupar]

President Trump: “Tuesday will be Power Plant Day, and Bridge Day, all wrapped up in one, in Iran. There will be nothing like it!!! Open the Fuckin’ Strait, you crazy bastards, or you’ll be living in Hell - JUST WATCH! Praise be to Allah. President DONALD J. TRUMP”

[Pastor Ben @BenjaminPDixon]

"Jesus is King" in Bio.

[Kim Bojórquez @kimbojorque]

SCOOP: The Salt Lake City Council is considering renaming a stretch of 500 South, which honors Cesar Chávez, following sexual abuse allegations against the late civil rights leader, Axios has learned.
https://t.co/9xJ7uSZ5zg

[Erin Alberty @erinalberty]

In the past three weeks, we've had measles at the SLC airport, Primary Children's Hospital and lots of schools and clinics. Health officials say to watch for symptoms. https://t.co/OOZAGMVORW

[LongTime🤓FirstTime👨‍💻 @LongTimeHistory]

Idaho teacher files lawsuit—after ordered to take down signs saying "Everyone Is Welcome Here!"

District lawyer said because hands pictured were different skin tones "not everyone agrees they are welcome."

"In this political environment, it's considered a personal opinion," he argued.

At the next school board meeting—they voted in agreement that the signs violate the policy requiring "classroom content to be neutral."

Sarah Inama has now filed a lawsuit against Lewis & Clark Middle School in Meridian, Idaho—accusing the school of being "racist" and violating her First and Fourteenth Amendment constitutional rights.

She is demanding a trial by jury and to recover the cost of her attorney fees.

[Zoya🕊️ @Zoya_ki_batein]

I just went to get an IUD consultation at a gynecologist and she goes, we can do it here, but just so you know, we don't numb. We won't numb your cervix here. And I go, why? And she goes, because the doctor at this clinic doesn't allow it. And I said who is that doctor? She pointed towards a male gynaecologist & said “dear even if I request it, he won't allow it”. And then i asked that male gynaecologist why he won't allow it, he told me because apparently, according to him, that's actually not even where the pain's coming from. So it doesn't even make sense to numb the cervix. i said its painful and he replied “its a part of being a woman”

[Boston Smalls @smalls2672]

CATO is a right wing think tank by the way.

[Erin Alberty @erinalberty]

@jkimballcook Semi-related, few things entertain me more than dudes acting all proud of themselves for finding women "attractive without makeup" while pointing to a picture of a woman who is caked up like a Broadway diva about to take the stage.

[Andrés @abrenesu_]

mi abuelo: ya vengo voy a hacer unos mandados

mi abuelo a las dos horas: