MovieLens.org

Posted on by

Editor’s Note: In the two years I’ve been playing with AI I’ve shared what I thought were some of the more interesting “conversations”, always being careful to distinguish something I wrote from AI output. ChatGPT frequently offers to compose a blog post and the drafts are as well written (better?) than mine. But they’re not mine. ChatGPT wrote (or plagiarized ) the following.

MovieLens is a free, non-commercial movie recommendation website developed by the GroupLens research lab at the University of Minnesota. It’s designed to help users discover movies they’re likely to enjoy based on their individual tastes. The site uses collaborative filtering algorithms to generate personalized recommendations, improving as more ratings are added.

Getting started is simple: users rate movies they’ve already seen using a five-star scale. The more films you rate, the better the system becomes at understanding your preferences. MovieLens then offers tailored suggestions that can be filtered by genre, release decade, popularity, and more.

One of the key strengths of MovieLens is its clean, ad-free interface and its emphasis on transparency. It doesn’t promote content for profit, and it doesn’t track users for advertising purposes. The recommendations are driven purely by user data and academic research.

Users can view their own rating history, create custom movie lists, and explore trends or themes in their viewing habits. While it lacks the social features of some commercial services, its core focus is accuracy and simplicity in recommendation.

MovieLens is especially useful for cinephiles who want to keep a record of what they’ve seen and find new films aligned with their personal taste, without the influence of box office hype or crowd-sourced noise. It’s also a valuable tool for researchers and students studying recommendation systems, as its algorithm and structure have been used in numerous academic projects and publications.

“The Computer-Science Bubble Is Bursting”

Posted on by

The subhead from this article in The Atlantic says it all: Artificial intelligence is ideally suited to replacing the very type of person who built it.

From 2005 to 2023, the number of comp-sci majors in the United States quadrupled. All of which makes the latest batch of numbers so startling. This year, enrollment grew by only 0.2 percent nationally. […] Artificial intelligence has proved to be even more valuable as a writer of computer code than as a writer of words. This means it is ideally suited to replacing the very type of person who built it. A recent Pew study found that Americans think software engineers will be most affected by generative AI.

I used to joke that (in 1966) I majored in Stay Out of Vietnam. Even back then, I couldn’t imagine what I would do with a major in Speech/Theater. But according to the Atlantic article, economists are suggesting college students “Major in a subject that offers enduring, transferable skills.”

I think of that sometimes when I’m writing a big check to the appliance repairman, the heating and AC service guy, or the plumber and electrician.

Discussing a book with something that hasn’t read it

Posted on by

I couldn’t find authoritative data but  I assume millions of copies of William Gibson’s novel, The Peripheral have been sold. And many more millions share my interest in Eastern spiritual traditions (specifically the illusory nature of the self). And the subset of those who fall into both groups is probably large… I just don’t happen to know any of those people. So I fired up my ChatGPT app fully aware of the bullshit risk.

“I’d like to get your take on one of the plot elements in William Gibson’s novel, The Peripheral. Before I get to that, I’m curious about how you know what you know. For example, were you trained on the full text of that novel or would your “understanding” be based on reviews or other sources?”

The rest of the conversation (PDF) was pretty interesting.