New #PPICsurvey: Strong majorities oppose Trump’s executive order to close down U.S. Department of Education. 71% of adults, 74% of public school parents, and majorities across state regions and demographic groups are opposed, while partisans are divided. Full report ➡️ bit.ly/3RRREuG

Trying a fun prompt with the new models: "I need you to create a plausible false history, like Borges (dont mention him by name). Erudite, illuminating some sort of deeper human truth, full of references." o3 seems particularly good at this, maintaining a theme throughout.

Competence and expertise are impediments to service in the Trump regime because reality often conflicts with Trump's whims. Trump wants only lickspittles willing to say with absolute sincerity that the sun will rise in the West tomorrow if he so orders it.

👀Today’s AIs are already hyper persuasive. A controversial study where LLMs tried to persuade users on Reddit found: “Notably, all our treatments surpass human performance substantially, achieving persuasive rates between three and six times higher than the human baseline.”

Playing Seriously: Four AI Apps I Built to Transform the Writing Classroom, by Eric Martinsen #AIinEducation #highered #EdTech Team playLab open.substack.com/pub/elmartinse…

Me: "o3, do the first chapter of Genesis as an IKEA instruction manual and show me the flatpack," "Fix the spelling" Me: "Now come up with an idea of what to do next" o3: "How about Genesis 2 - Garden Starter Kit?" Me: "Sure."

Marcus Luther Teaching is the only profession where technology creates more work for employees.

o3 now one-shots extremely New Yorker-y original New Yorker cartoons, and sometimes the punchlines are actually kind of okay. And most of them involve therapists or cellphones or both, which feels very on-brand.

How does it work? Students attended twelve 90-minute sessions where they interacted in pairs with ChatGPT. Teachers started each session with a suggested prompt, which was carefully designed to align with the science of learning and encourage the LLM to act as a tutor.

The key? We used the AI as a tutor, not a crutch. Prompts were very carefully designed to align with the science of learning and be context relevant.

My takeaway is that by now we know AI can harm learning and generate cognitive offloading if used as a crutch. But when used and design properly and in alignment with pedagogical practices, AI can enhance and accelerate learning.

Trump refused to call in the National Guard when it was most needed. Let that sink in.

I am starting to think sycophancy is going to be a bigger problem than pure hallucination as LLMs improve. Models that won’t tell you directly when you are wrong (and justify your correctness) are ultimately more dangerous to decision-making than models that are sometimes wrong.

I know it is a small thing, but, in these dying days of the open web, it is lovely that such a large proportion of famous poetry is online, mostly due to a $100M gift from Ruth Lily, who loved poetry (even though she never got any of her own published) poetryfoundation.org/poems/guides

Never been a better time to study the humanities: 1) LLMs are trained on the cultural history of all humans, knowing that helps you use them 2) The humanities gives us context in this odd moment in history 3) Books & stuff are good Wrote this 3 years ago: oneusefulthing.org/p/magic-for-en…

AI really can help education: Randomized controlled experiment on high school students found a GPT-4o powered tutor that personalized problems for students raised final test scores by .15 SD, "equivalent to as much as six to nine months of additional schooling by some estimates"

🚨🚨 Excited to share our first *positive* results on AI in education! Most AI tutor work focuses on making the chatbot better. We suggest another lever: deciding what students should practice next to improve learning. We combine an LLM tutor with reinforcement learning to

New study involving 770 high school students over 5 months across 10 schools. All students got the same lectures, course material, and GenAI tutor. The only difference: half received a fixed sequence of practice problems (easy to hard, standard practice), while the other half had