Join 📚 Josh Beckman's Highlights
A batch of the best highlights from what Josh's read, .
In fact, the dental-office “badness” of so much of the A.I. art is precisely why I don’t dispute Alexander’s assertion that people preferred it. Like any LLM output, A.I.-generated images are designed to please, not to provoke. [I’ve argued before that these images are, by their nature, almost unavoidably kitsch](https://maxread.substack.com/p/what-ai-art-spiral-images-tell-us)--comforting, straightforward, accessible, flattering. And people love kitsch!
People Prefer A.I. Art Because People Prefer Bad Art
Max Read
RAG is often implemented using [vector search against embeddings](https://simonwillison.net/2023/Oct/23/embeddings/#answering-questions-with-retrieval-augmented-generation), but there’s an alternative approach where you turn the user’s question into some full-text search queries, run those against a traditional search engine, then feed the results back into an LLM and ask it to use them to answer the question.
Building search-based RAG using Claude, Datasette and Val Town
Simon Willison
As is so often SOP with the truth, there’s a cruel paradox involved. It may well be that we spectators, who are not divinely gifted as athletes, are the only ones able truly to see, articulate, and animate the experience of the gift we are denied. And that those who receive and act out the gift of athletic genius must, perforce, be blind and dumb about it—and not because blindness and dumbness are the price of the gift, but because they are its essence.
String Theory
David Foster Wallace
...catch up on these, and many more highlights
Readwise