A Rose in Winter
Seen on a woodland walk.
Quote of the Day
”If you think education is expensive, try ignorance.”
- Derek Bok (President of Harvard 1971-91)
Musical alternative to the morning’s radio news
Luca Sestak | How Long Blues
I love the Jimmy Yancey version of this, but can’t find a recording that has decent audio quality.
Long Read of the Day
Will ChatGPT Kill the Student Essay?
Stand by for a year of obsessing about ‘Generative AI’
ChatGPT went from zero to a million users in five days. What it reminds me of is the moment in 1993 when Marc Andreessen and Eric Bina released Mosaic, the first modern Web browser. The Internet we use today (the network based on the TCP/IP family of protocols) had been up and running for ten years by that time, but it had essentially been a parallel universe inhabited mainly by geeks. Inhabitants of the other universe — what John Perry Barlow called ‘meatspace’ (aka the real world) — basically had no idea what this “Internet thingy” (as one newspaper editor put it to me) was for. But the moment they experienced Mosaic they suddenly knew what it was for. And the rest is history.
My feeling is that we have now reached an analogous inflection point with ‘AI’, which is the tech euphemism for machine-learning. The tech world has been obsessed by it for nearly a decade, but most ‘normal’ people have no idea of what it is or why it might be a big deal.
Earlier this year, ‘Generative AI’ like Midjourney gave them a glimmer of insight into why it could be interesting, but basically they saw those tools as toys. It’s fun to be able to issue an instruction like “Draw a picture of J.K. Rowling as an astronaut” and then see interesting and amusing images. That kind of ‘AI’ is basically a power tool for people who can’t draw.
(See Commonplace Booklet below for a link to an example.)
ChatGPT is different because it does something that people need to do — compose and write apparently coherent text — but find difficult. And here it is in a machine, just waiting for instructions from you. Having listened to the reactions of numerous non-techie people who have tried it, what I’m hearing is “Oh now I see — this is what this AI-stuff is for!”
Which is why this little essay by Stephen Marche is interesting.
Here’s how it begins:
Suppose you are a professor of pedagogy, and you assign an essay on learning styles. A student hands in an essay with the following opening paragraph:
“The construct of “learning styles” is problematic because it fails to account for the processes through which learning styles are shaped. Some students might develop a particular learning style because they have had particular experiences. Others might develop a particular learning style by trying to accommodate to a learning environment that was not well suited to their learning needs. Ultimately, we need to understand the interactions among learning styles and environmental and personal factors, and how these shape how we learn and the kinds of learning we experience.”
Pass or fail? A- or B+? And how would your grade change if you knew a human student hadn’t written it at all? Because Mike Sharples, a professor in the U.K., used GPT-3, a large language model from OpenAI that automatically generates text from a prompt, to write it. (The whole essay, which Sharples considered graduate-level, is available, complete with references, here.) Personally, I lean toward a B+. The passage reads like filler, but so do most student essays…
Thanks to John Seeley for spotting it.
My commonplace booklet
Lemon juicer designed by ChatGPT and drawn by Midjourney
An experiment by Azeem Azhar. Link here.
This Blog is also available as a daily email. If you think that might suit you better, why not [subscribe]? One email a day, Monday through Friday, delivered to your inbox. It’s free, and you can always unsubscribe if you conclude your inbox is full enough already!