I’ve been putting this off for at least a month, but it’s time, time to get the balls in the air and see how they fly.
ChatGPT
I’ve got two big projects hanging over my head, finishing my report on last year’s work with ChatGPT and concluding my series on the greatest literary critics. I’d originally intended to post the ChatGPT report in early December, a year after I started working with the Chatster. But somehow I just couldn’t get psyched, or was I avoiding it for whatever reason? I suppose there’s a similar story for the litcrit greats.
Of the two, I suppose the ChatGPT report is more important because it’s a way of getting set-up for future work, though I’ve been doing that anyhow. I’ve done quite a bit of work already, but I need to write some narrative exposition. I think I want to focus on conceptual ontology. Why? Because that’s the armature of the semantic system, the supporting skeleton. It’s completely invisible in ordinary queries because it structures everything. It’s everywhere. But it shows up in the story variations quite obviously. From there I can move to analogy and abstract concepts. One of the audiences I’m most interested in reaching, current researchers in AI and ML, won’t be familiar with this material, though they probably know the relevant concept of ontology from its use in computer science. They’re just not familiar with it in the context of natural language. I also want to include something about prompting and associative memory.
This will lead naturally to a discussion of hallucination and alignment, which is where I see this going. It may also be useful in prompt engineering. And then there’s citizen science.
Greatest literary critics
As I said at the outset of the project, this is a response to Tyler Cowen’s book on the greatest economists. I’ve got two more posts to go. The next and penultimate post will be about Harold Bloom. This will give me a chance to talk about the issue of authority in academic literary criticism, something I’ve touched on here and there, but not really gone into, though there is yesterday’s post about LitCrit “Stars” and AI “Godfathers.”
That will be the last in the “greatest” series of posts, but not the end of the series, where I’ll take a look at the future of academic literary criticism. I’ve written a great deal about that already – e.g. Prospects: The Limits of Discursive Thinking and the Future of Literary Criticism – but I should be able to say something a bit different, if not quite new. In particular, I want to say something about ethical criticism in relation to undergraduate education and public outreach.
Mission accomplished!
I’ve finished two of the projects I had in mind when I first planned this post over a month ago. There’s a 3QD post about my relationship with Charlie Keil: On the 12/8 Path with Charlie Keil. To be honest, it’s not quite what I’d had in mind, not what I’d hoped, but it’s there. I can always do more.
And I’ve written my final post in the Maestro series: Some Thoughts on Maestro, ethical criticism, and calibration [Media Notes 102d]. That turned out to be a bit better than I’d thought it would be, and just a bit longer than I’d been thinking, but not very long as posts around here can be.
Some posts and working papers on the to-do list
I’m thinking of doing a post on the process of running ‘experiments,’ as they call them, on ChatGPT. I want to relate this to the control theory of William Powers. The idea is simple: When we issue a prompt, we are attempting to control ChatGPT. If it does what we want, the attempt is successful.
And then there are working papers on Miriam Yevick and on the State of AI.
I also want to do a short post about my photography and how I feel that, through it, I have made Hoboken my home. This would be similar to a post I did about Jersey City. The core idea: I now feel a responsibility to people in the Hoboken Photo Group, some of whom are no longer living in Hoboken, but enjoy photos of their home town.
Photo Exhibit
Finally, I have to get moving on a small photo exhibit I’ve been planning. I’m thinking about 10 to 15 photos, all of Hoboken. I’ve got to order some larger prints (8 by 10, 8 by 12) to show around and get reactions before I make my final selection.
No comments:
Post a Comment