Nigel Caplan

"Wlimington's spelling bee content registration opens"

The (Delaware) News Journal prints a spelling error in a headline about ...

Opinion | Students Are Skipping the Hardest Part of Growing Up

Exactly. And bonus points for "Cyrano as service". This is why we need enforced limits and strong policies that do not acquiesce to big tech. And why I am unimpressed by the academic trend of "look at the AI tool I made which is ok really but no one will ever use"

www.nytimes.com/2026/01/30/o...

It was not

Students Are Finding New Ways to Cheat on the SAT

The only surprising part of this article is the correction at the end, which is truly a work of art

www.nytimes.com/2026/01/28/u...

Bless you, graduate applications systems that only require an uploaded letter of recommendation without asking a bunch of pointless ranking questions. #academicsky

The risks of AI in schools outweigh the benefits, report says A new report warns that AI poses a serious threat to children's cognitive development and emotional well-being.

This is incoherent hype. They find that "AI" products diminish critical thinking, knowledge, and creativity; increase inequity; and threaten social and emotional development. So the solution is ... make more of them and force "AI literacy" on everyone? www.npr.org/2026/01/14/n...

(b) privacy has nothing to do with language acquisition. The solution is, as always GOOD TEACHING: structured pair and group work, teachers trained to work with MLs, and most importantly, sufficient teachers with manageable workloads. Not "AI". Plus "AI offers privacy"? NPR has read the news, right?

Forgive me for being an actual ESL and SLA teacher, but: (a) "adjusting complexity" might not be helpful - our goal is to give MLs access to grade-level content; these products introduce errors and misrepresent material; and only a trained expert has the skill to simplify without dumbing-down

I got one paragraph in to find: "Teachers surveyed for the report said AI can be useful ... for students learning a second language. For example, AI can adjust the complexity of a passage depending on the reader's skill, and it offers privacy for students who struggle in large-group settings."

Grok generated harmful images - or did users prompt it to and coders give it the capacity to? Did the chatbot apologize or the billionaire owner? Can computer code take responsibility, take action, and make restitution? Or do these require human actors? That's critical thinking - asking WHO. /end

or person logically be the Agent or Sayer in the clause? Can a piece of software say/think/act? Does the AI product generate images or does a human Agent prompt it to do so? Who are these human or corporate Agents and Sayers? Then edit your writing accordingly. Eg "ChatGPT told me ..." - did it? 2/3

So much would be gained from teaching #sfl rather than structuralist grammar (subject, object, passive voice). The Participant responsible for an action verb is the Agent. The Participant responsible for a saying/thinking verb (verbal/mental process) is the Sayer. So ask yourself: can this thing 1/3

Grok turns off image generator for most users after outcry over sexualised AI imagery Editing function to be limited to paying subscribers after X threatened with fines and regulatory action

Bad writing from the Graudian. "Grok" can't be the agent of the verb "turn off" unless the AI product has the capacity to disable its own functions. It's the CORPORATION behind it which has the CHOICE whether to turn "features" like this on or off. Ascribe agency. www.theguardian.com/technology/2...

Nope. "Grok" doesn't "say" shit. Actually, it does say shit. That's all it can. But it certainly can't be trusted to provide an accurate account of corporate policy. Ask the bleedin' company for confirmation. That's literally your job.

Every journalist and editor who prints "[AI product] says/claims/thinks" should be forced to take a writing class or reassigned to the cooking section, or something. No not cooking, they'll print slop recipes. FFS, it's not hard. AI generators make shit up. Ask the *company* what it's doing.

I'm reading this while I bitch the pot for my morning scandal water

Ten archaic English slang phrases…

10. Got the morbs (sad)
9. Bumpsy (drunk)
8. Whittled as a penguin (very drunk)
8. Scandal water (tea)
7. Bitch the pot (pour the tea)
5. Sauce-box (mouth)
4. Cupid’s kettle drums (breasts)
3. Dash my wig (OMG)
2. Poked up (embarrassed)
1. Not up to dick (unwell)

Conversely, look for drops in apps and yields to the AI forward schools. I wouldn't pay for my child to go to Ohio State or Purdue now. No chance.

Opinion | We Owe It to College Students to Create Tech-Free Spaces

Again, the first University to do this will absolutely clean up. I hope it's mine. What do you say, Delaware?

www.nytimes.com/2025/12/19/o...

This is of course dumb and horrifying and a sad reflection of our inability to regulate and ban actual guns. But the irony of "AI" seeing arts, culture, and human experience as a threat is pretty spectacular. I doubt WaPo sees it.

best new student coinage of the semester: "discourse sematintics" - Tintin aux profs d'Anglais?

Yes, there are problems, but there are already ways to filter applications with simple algorithms (come on, we know they're doing it). The goal is obvious - it's not saving time, it's reducing staff. Students (or those high priced consultants) are going to start writing for the "AI" not humans. Ugh.

They have to select from their activities and put them in order of importance. And in a short personal essay, every word counts. To use a dumb machine to "summarize" their application is gross malpractice and makes the application itself redundant. Just read the damn thing.

As the parent of a HS senior, I can tell you that this is completely insane and insulting. The Common App is already a summary. Students have limited "slots" to enter extracurricular activities, achievements, etc, and the essay is, what 500 words? Diligent students spend hours crafting their app

Turns out plagiarism really is the sincerest form of flattery for the egotist

And most appropriately, sadly, today: Rosencranz and Guildenstern are Dead

The Lion King

The conclusion is comically naive. Why would tech corp$ make LLMs better by adding "friction "? There's no business incentive and, not only don't they care about education, they are actively trying to dismantle
dismantle all but elite educational opportunities

Learning with AI falls short compared to old-fashioned web search Doing the mental work of connecting the dots across multiple web queries appears to help people understand the material better compared to an AI summary.

Feature not bug. Bonus points for using the phrase "old fashioned web search"

theconversation.com/learning-wit...

Sometimes my grammar lessons write themselves