SGomez
@elfaro
SGomez
@elfaro
This connects to the importance i am putting on chasing the provenance to every llm response in my ai literacy instruction. It is important to consider, however, that even when we instruct the llm to surface the provenance of its response, the llm is only reconstructing a plausible provenance for us that is disconnected from the documentation of the knowledge and how it came to be, or the actual provenance. It is getting rid of a huge part of the value, the work, of documentation over history. The problem is that without legitimate provenance, these llm responses are not actual knowledge they are only gesturing to knowledge. This entire blog post by Talisman is fascinating an worth a read.
From Nick Milo. Thinking about how much more useful all of these collected notes in Sublime would be if I had added a note to each saying this is interesting to me because… Then they wouldn't be this random grouping of links I can barely remember looking at and would instead be a history of my thinking. I'm guessing that the accompanying notes would also function as memory triggers for what was in the links and their context.
“But I think that consciousness is at stake in a lot of what’s going on. One of the things Trump has done is occupy a significant chunk of our attention every single day. Our consciousness is being polluted, and protecting ourselves against that at the same time we preserve the ability to act politically is a difficult balancing act. Consciousness
... See more