Ask A Genius 922: Claude, the Large Language Model
Author(s): Rick Rosner and Scott Douglas Jacobsen
Publication (Outlet/Website): Ask A Genius
Publication Date (yyyy/mm/dd): 2024/04/20
[Recording Start]
Rick Rosner: All right, so at the beginning of the last little session, I mentioned that I have been using a free version of the LLM, Large Language Model AI, called Claude. Claude comes in various flavours, including a fancy one you can pay for. Do you know the difference between the fancy ones you pay for and the free and cheap ones?
Scott Douglas Jacobsen: My suspicion is the depth of the programming because ChatGPT 3.5 is very different from 4.0. 4.0 has more creativity in the mix; the parameters are considered, and the data points are much broader in many dimensions than the ones you pay for.
Rosner: So, let us see, I am asking Claude himself. Well, it does not know, admits ignorance. “I do not have information about paid versions or different service tiers. I am Claude, an AI assistant created by Anthropic to be helpful, harmless, and honest. I cannot access details about Anthropic’s product plans or pricing models. I do not have any special insider information about the company beyond what is publicly available. My role is to have natural conversations and provide helpful information to users based on my training.” So, that seems like the opposite of insidious. I mean, even insidious AI would try to be the opposite of insidious.
In the previous session, I read 15 ideas Claude gave to start a rom-com. I am too cheap to pay for a pay version to see if the ideas would improve. I doubt it because, in things like a romcom or a Liam Neeson movie, it is generally not like a genius idea that makes the movie good or not; it is whether it is executed with care and cleverness. I saw a rom-com with Emilia Clarke called Last Christmas, which had a genius plot. I am not sure it was overly satisfying, but it did have some surprising plot twists and ones that make you slap your head and go, “Oh, I see where all this was leading to,” like all that lays out a bunch of clues. When all the clues come together, you say, “Ah!” The romance itself is satisfying, but the surprise that all this fucking around was covering up a hidden structure that finally gets revealed; that was cool, but most romcoms are not clever that way. They have clever dialogue, and they are not entirely predictable. You know they are going to fight; you know they will hate each other at some point, maybe in the beginning, and then after somebody does something that’s ill-advised in secret and that secret gets revealed, there is apologizing. If you can develop a structure that does not work like that but just the basic ideas, I do not think you would get better ones from a better AI.
[Recording End]
License
In-Sight Publishing by Scott Douglas Jacobsen is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Based on a work at www.in-sightpublishing.com.
Copyright
© Scott Douglas Jacobsen and In-Sight Publishing 2012-Present. Unauthorized use and/or duplication of this material without express and written permission from this site’s author and/or owner is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given to Scott Douglas Jacobsen and In-Sight Publishing with appropriate and specific direction to the original content. All interviewees and authors co-copyright their material and may disseminate for their independent purposes.
In-Sight Publishing 