If you agree that grades "should reflect whether students meet the learning objectives of the course, not how they perform compared to their peers", maybe you'd like to join us at the Grading Conference in a few weeks to learn about different ways to do exactly that. thegradingconference.com

Harvard has just voted to cap grades: instructors may assign A's to a max of 20% of students (plus four) per course. The outcome will unfairly harm racially minoritized students and, through course evals, gender minoritized instructors. My new piece in Inside Higher Ed. #AcademicSky #EduSky

It's good that universities are making six-figure payouts to people who got fired for their comments about Charlie Kirk, but now the universities need to do some new firing, of administrators who carried out those illegal actions. www.nytimes.com/2026/05/27/u...

✅️ Deficit framing ✅️ Presenting non-neutral numbers as "objective data" ✅️ Blaming "grade inflation" ✅️ Unfounded claims that this is actually equitable ✅️ Completely ignoring other things that happened in 2020 These "arguments" are all so predictable...

Hear a lot that AI can be used to enhance learning rather than outsource but every day I get emails like this (subject:"Turn big ideas into A work") offering me 50% off Grammarly to generate my work with AI, check it to see if it sounds like AI, then rewrite it with AI to not sound like AI.

one interesting thing so far about loudly hating AI is the way AI lovers are like "wow so you're just totally closed off to hearing another perspective??" as if their perspective hasn't been relentlessly pushed down our throats with no space to push back (particularly in professional settings)

I think if I were an AI company trying to create a perception that AIs were going to replace mathematicians, what I would do is grab a large set of high-profile open problems and just start grinding counterexamples until I've found one, and then only report what went into finding that specific one

10 minute Uber ride to the airport in Alabama this morning: learned the driver's weekend plans, what she's making for the cookout this weekend, discussed pop up thunderstorms, and learned where her kid goes to elementary school 45 minute Uber ride home from SFO airport: 🦗🦗🦗

Personally I think it should be an all-billionaire crew. Obviously we must send our best, brightest, and most successful.

Apparently it is just me, but if you asked me about a solution to a problem I had never heard of I would not participate in writing an article of fawning praise for a proof, particularly one that only mentions in passing how it does not cite all the papers where it drew existing ideas from.

Grading systems are a not-inevitable hierarchy institutions enthusiastically opt into to serve those at the top.

Apropos of nothing, if you think grades should reflect actual student learning rather than arbitrary institutional policies, I'd like to invite you to join us at the Grading Conference June 16--18

"We can't possibly be held accountable for the content of our citations" well then i can't possibly trust your research and you should go sit with yourself about that

What's the over/under on how long before some AI bros who were banned start pushing a "non-woke" (or something other such nonsense adjective) arXiv copycat?

“The U.S. National Science Foundation X-Labs (NSF X-Labs) initiative is designed to launch and scale a new generation of organizations focused on independent research, development and innovation.” We have those, thanks, they’re called universities and national laboratories.

Very shocking that these technologies are yet again reinforcing existing structural inequities 🙄

Some academics wouldn't know solidarity if it hit them in the face like a wet fish. Ay yie yie.

Nondisabled people like the idea of AI accessibility because on a superficial level, it makes everyone sound kind of the same, thereby erasing the disability in the experience of the author. That’s not empowerment. It’s just not. It’s deciding a person is worth more if they sound a certain way.

It’s also ableist as fuck. You think no person with a disability has ever learned or relearned? Obviously that’s not true. You think people with intellectual disabilities CAN’T learn without the frame of others’ syntax? That’s insulting. And wrong. This isn’t democratizing. It’s wholesale theft.

Here’s why I find the argument that AI could be a disability aid insulting. I got really, really sick in 2022. Been disabled my whole life, but 2020 was when things started becoming impossible and 2022 yanked them all away. I carved a writing practice out of nothing. Every word was pain.

My second hottest take of the day is that if you have made your language not backwards compatible with itself and your website links only to old documentation, I am very unlikely to want to use your language ever again.

A blog post bearing all the hallmarks of AI slop complaining that faculty are resisting AI? Is this truly what the MAA values?

Technical question (apropos of nothing): how does one prove that data was deleted, and that no other copies exist?

Very cool that we can always find untold millions to pay AI companies and vendors that use that money to pay off criminals, but "can't afford" cost of living adjustments for faculty and staff

The Purcell Principle means it always too late for courts to fix something racist and never too late for states to do something racist

Goddess help me, I'm diving into the AI-in-math convo. Tim Gowers says ChatGPT 5.5 Pro did "PhD-level" math in an hour with no input from him. Terence Tao tells Nature the job description is changing. I wrote up what I think, including what could be lost: people, training, equity... #MathSky

It feels incredibly pedantic to me that we have to keep making this point, but I think it is essential. We are ceding our humanity to machines when we pretend they are doing or producing math.

AI nonsense aside, the Gowers post links to an article titled "Diversity, Equity and Inclusion for Problems in Additive Number Theory". I looked at it, curious what the author had to say about equity and inclusion in number theory, or in math more broadly. Spoiler: Nothing

I continue to be befuddled at how many supposedly novel "applications" of these technologies amount to doing a solved problem less accurately.

Just searched for how to do something in vi, and found a discussion with OP complaining that it takes 6 keystrokes to do this, a reply offering a solution with only 3 keystrokes, and a "Well actually..." reply pointing out that you have to press enter as well making it a mere 4 keystrokes

I am getting tired of playing the "Is it broken because of ransomware or because it was vibe coded" game with every single web service these days. This week's contestants include Canvas and Discord

in a better world getting caught doing something like this would be an automatic firing & permanent blacklist offense

Just so I am clear, in some states they are trying to redraw districts after the primary and redo the primary to disenfranchise Black voters as rapidly as possible, and in others they are overturning a referendum approved by voters to prevent redrawing districts?

Me, a person who is not teaching, watching the dumpster fire that is Canvas right now

TFW you go to do a thing, and it has been automatically done because of a process you spent time setting up a year ago

Sometimes when people use their AI "note taker" in a meeting I am in, it emails me the summary. Every single time I have gotten one of these there have been significant errors.

Glad your vibe coded personalized keto assistant has increased your protein input or whatever but people actually hate this shit for reasons that have nothing to do with pollution or structural power, which are also bad. They hate this shit because it is making their lives, personally, worse.

Racism. The outrage was racism.

My perfect strategy to make sure I never mess anything up is to wake up and before I do anything I read the note that says "don't mess anything up today," incredible that this airtight strategy also works on a bunch of eigenvectors that don't know what words mean

A month later, even more ridiculous

The "skills" kids need to use AI is to know when it's fucking wrong and how to fix it, *which is what we're supposed to be already teaching in schools*.

I really have to emphasize that this is how generative ai/llms do *everything*

I’m giving this as a gift to all the people in the last year who have said “I think [XYZ use] is bullshit, but some of the medical AI applications seem to be impressive…" where radiology is the most frequent example

Isn't conservatism all about resisting change? Whether that change is small progress against white supremacy or definitions of astronomical terms.

Every time I see someone talk about Claude I think about how they switched from female to male names as soon as they wanted the robots to seem like experts vs personal assistants.

Look! This LLM gave the correct answer to a question I asked!

I am not going to link the latest garbage piece attacking teaching centers, but I do wonder if these people that dismiss evidence-based practice in teaching and learning as "technocratic language" insist that their doctors and nurses not use evidence-based practice in their work.

"The document is, to use a technical term, totally fucking bonkers."