Tamar A K

What on earth?
"A recent research paper found that it is inevitable for LLMs to provide outputs that are not true. However, today I asked three LLMs one tricky question, and none of them provided a fake answer. Rest assured that LLMs are no longer making stuff up, and you can trust them now."

AI-Generated “Workslop” Is Destroying Productivity Despite a surge in generative AI use across workplaces, most companies are seeing little measurable ROI. One possible reason is because AI tools are being used to produce “workslop”—content that appea...

This is a great description of a new workplace phenomenon. I know several people who are not trying to be rude when they use AI in collaborative work. But it is rude. It imposes both practical and emotional costs on collaborators who are then forced to deal with the dumb crap you sent them.

For my HRM class on Friday, I'm planning to talk about some of the key events of 2025 relevant to understanding what is happening with discrimination law, the EEOC, and "DEI." Has anyone compiled a list of events like this? I could piece something together myself but maybe someone's already done it.

This. Or consider turning mob griefing off. It doesn't make a ton of sense to insist on no violence in the context of a game with creatures running around whose only purpose is to explode and destroy your stuff. You can just change the actual game settings.

Has anyone gotten Novavax this fall, or does anyone know when/in which pharmacies they might be available?

I miss getting emails that didn't all sound like they were written by the same robot. Email was already so annoying and hard to manage. Now, often, there isn't even the feeling of making contact with other people.

That feeling when you're listening to a music history lecture series in an attempt to think about art and not fascism, and the next lecture is on Shostakovich's Symphony no. 5. 😱

He doesn't strike me as a person who sleeps well enough OR who has the kind of imagination necessary for nightmares

Researchers who co-author papers: what conversations have you had with your coauthors about LLM use during the writing process? Have you tried to influence your coauthors' LLM practices in any way? Have they tried to influence yours?

Having a diagnosis does help with this!

RIGHT?! But also, does it make you feel any better that I did know about mine at the time and yet it didn't make that much of a difference in my care :/

Given the culture around pregnancy sickness, it is somehow predictable, unfathomable, and infuriating all at once that a big name researcher cannot get NIH funding for clinical trials for a promising intervention with a strong basis in prior findings.

Watching a webinar with @careyjones.bsky.social and Dr. Marlena Fejzo on pregnancy sickness research, I'm noticing how many women describe their experience as "horrific, but not bad in the scheme of things"; "destroyed my life, but others had it worse." (I'd put myself in this category, too!)

(2/2)
2. Trump's nonstandard capitalization (e.g. in the Brazil tariff letter)

3. What it does to us psychologically to have a million small experiences of absurdity and meaninglessness on the Internet each day (this one I might actually try to do research on myself)

(1/2) A few 2025 phenomena I want to read scholarly or otherwise erudite articles about. Universe, make it happen

1. As scams grow to look more like real things, real things are also growing to look more like scams

I can't believe how long it took in this article to get to actual fraud-related details and not just "she says she is disabled and yet she has fun?!?!"

(that's also setting aside ethical issues with using LLMs in general re: environmental impacts and how these models were trained on humans' work without those humans' approval)

Calculators are also solving a problem that has one correct answer. LLMs are being asked to solve problems that, in theory, have infinite "correct" answers. That's why grading work done by LLM is deeply boring in a way grading work done by a calculator is not.

I've used the calculator analogy myself. One important way it breaks down is that calculators are very reliably correct but LLMs not. Calculators are actually figuring out what the right answer is, but LLMs are just trying to sound smart.

As replies have pointed out, this is unfairly burdensome to some students with disabilities. But my follow-up question is, why hasn't someone made an updated typewriter or non-Internet-enabled word processor? Like how there are dumb phones for kids. Someone get on it please?!

Graphic of Washington Post poll from April 18-22, 2025, showing "Majorities of Americans oppose various Trump proposals" highlighting two lines in particular: "Increasing the Federal Government's role in how private universities operate" - 28 percent support, 70 percent oppose "Reducing federal funding for medical research" - 21 percent support, 77 percent oppose

University leaders need to understand: they are holding INCREDIBLY good cards.

It is so rare in today's polarized era to have the support of 70+ percent of the public against the Trump admiministration on anything. Universities have that—AND winning legal arguments.

The very first paragraph of the article says that this was already the trend before the pandemic. Stop blaming everything on covid closures without evidence that they specifically are the cause.

I (social science faculty) took an undergrad philosophy of disability class recently that was all nonmajors (as of start of class) and really great content. I don't know whether anyone switched their major because of it but I wouldn't be surprised

Really reminding me of the Enron documentary I used to show in my class:
"These guys could just yank the California economy on its leash whenever they wanted to. And they did it and they did it and they did it, and they made SO MUCH MONEY."

Am I understanding correctly that the profs who were contacted were those who signed a petition expressing concern about *antisemitism*? And that the purpose of obtaining their info and contacting them was to encourage them to snitch against their colleagues?

Not so new, but if you're talking about transparency with the file drawer or internal meta analysis, you could discuss the merits and drawbacks of the approach we took in my 2017 JPSP with Kristin Laurin and Anna Merritt. We presented our entire file drawer in a supplement.

Pretty sure all of us are trying to protect jobs and it's not theoretical to anyone. Totally different ideas about what is "high risk" in this situation.

This is absolutely what they think. Lots of language at the president's forum last week about how complying protects the community and how those with "higher risk tolerance" may choose to speak out, but to the UHM admin "this is not theoretical, we are trying to protect real jobs."

Yet, I would bet people think men are better moral *reasoners* than women.

My son said "I don't understand why people are so afraid to get an autistic kid. Autistic kids are awesome."
Can confirm. I love the deep dive in special interests, honesty, concern for justice, and empathy. Autistic minds enrich our world.