Simon Willison’s Weblog

Subscribe

Thursday, 4th July 2024

Exorcising us of the Primer (via) Andy Matuschak talks about the need for educational technologists to break free from the siren's call of "The Young Lady’s Illustrated Primer" - the universal interactive textbook described by Neal Stephenson in his novel The Diamond Age.

The Primer offers an incredibly compelling vision, and Andy uses fifteen years of his own experience exploring related ideas to pick it apart and highlight its flaws.

I want to exorcise myself of the Primer. I want to clearly delineate what makes its vision so compelling—what I want to carry in my heart as a creative fuel. But I also want to sharply clarify the lessons we shouldn’t take from the Primer, and what it simply ignores. Then I want to reconstitute all that into something new, a vision I can use to drive my work forward.

On the Primer's authoritarianism:

The Primer has an agenda. It is designed to instill a set of values and ideas, and while it’s supportive of Nell’s curiosities, those are “side quests” to its central structure. Each of the twelve “Lands Beyond” focuses on different topics, but they’re not specific to Nell, and Nell didn’t choose them. In fact, Nell doesn’t even know the Primer’s goals for her—she’s never told. Its goals are its own privileged secret. Nell is manipulated so completely by the Primer, for so much of her life, that it’s hard to determine whether she has meaningful goals or values, other than those the Primer’s creators have deemed “good for her”.

I'm also reminded of Stephenson's piece of advice to people who may have missed an important lesson from the novel:

Kids need to get answers from humans who love them.

# 4:39 am / education, ai, neal-stephenson, andy-matuschak

The expansion of the jagged frontier of AI capability is subtle and requires a lot of experience with various models to understand what they can, and can’t, do. That is why I suggest that people and organizations keep an “impossibility list” - things that their experiments have shown that AI can definitely not do today but which it can almost do. For example, no AI can create a satisfying puzzle or mystery for you to solve, but they are getting closer. When AI models are updated, test them on your impossibility list to see if they can now do these impossible tasks.

Ethan Mollick

# 10:38 pm / ai, llms, ethan-mollick