automated essay scoring was capable of producing scores similar to human scores for extended-response writing items, with equal performance for both source-based and traditional writing genre.

another luddite argument against education at scale falls.

Moonbot is the company that reminds me most of the young lady’s illustrated primer from the Diamond Age.

Their first project, The Fantastic Flying Books of Mr. Morris Lessmore, was released for the iPad last May. It recounts the wondrous adventures of a book lover who dotingly cares for a living library before writing a book himself that tells of “his joys and sorrows, of all that he knew and everything that he hoped.” Gorgeously illustrated, Lessmore breaks new ground in the way that it incorporates interactivity. Each page has a wormhole of interaction. Read about a song and perhaps a keyboard will pop up and guide your fingers to plunk out “Pop Goes the Weasel.” When Morris Lessmore hand-feeds alphabet cereal to his books, the reader gets a bowl too, with letters that can be dragged along through the milk to spell out words. Each page holds its game like a secret and puzzling out what to do encourages the reader to look harder, knowing they’ll be rewarded. The games pull the reader deeper; the narrative pulls the reader farther. The tension between lingering and racing is potent.

2023-03-25: We’re plausibly much, much closer to the vision.

With the introduction of GPT-4 and Claude, AI has taken another big step forward. GPT-4 is human-level or better at many hard tasks, a huge improvement over GPT-3.5, which was released only a few months ago. Yet amid the debate over these advances, there has been very little discussion of one of the most profound effects of AI large language models: how they will reshape childhood.
In the future, every middle-class kid will grow up with a personalized AI assistant — so long as the parents are OK with that.
Letting your kid have an AI companion will bring big advantages. Your child will learn to read and write much faster and better, and will do better in school. Or maybe you want your kid to master Spanish or Chinese, but you can’t afford an expensive tutor who comes only twice a week. Do you want your child to learn how to read music? The AI services will be as limited or as expansive as you want them to be.But the biggest drawback might simply be that the AI services work too well, and kids become very attached to them, neglecting friends and family. They might be such good babysitters that parents won’t always pull the plug when they should. They might, in short, be the 21st century version of television.

It’s far less clear whether they’ll make inroads in education, which has been extremely resistant to online learning and other developments.

Advancements in AI, like ChatGPT, promise to become incredible educational tools. However, they will have a mixed impact if there isn’t widespread demand from students, parents, and teachers. Any implementation connected to central databases will likely face tremendous pushback. ChatGPT-for-schools must be compatible with a “Gentleman’s C” for widespread adoption. It could even be popular if it helps with classroom control, allows students to goof off, and lets parents believe their students have world-class teachers.

Only 1 year ago, Harvard had a $36.9b endowment, the largest in academia. Now that endowment has imploded, and the university faces the worst financial crisis in its 373-year history. Could the same lethal mix of uncurbed expansion, colossal debt, arrogance, and mismanagement that ravaged Wall Street bring down America’s most famous university? And how much of the turmoil is the fault of former Harvard president Larry Summers, now a top economic adviser to President Obama? As students demonstrate, administrators impose Draconian cuts, and construction is halted on an over-ambitious $1.2b science complex, the author follows the finger-pointing.

harvard is so fucked