With the amount of projects slopping into "AI" recently, I wonder if someone's keeping a watch on them as they proliferate rather fast.
Maybe it's time to compile a naughty list or something?
llm
You know what I want? As an already existing LLM critic? Give me either, a local model that can be trained on anything I choose, no data centers required, or if it's going to be on a server, very specialized and or very targeted models that are trained on only open and consenting data, in my case, public domain books and opted in Creative Commons books. I want to have two modes, an assistant kind of, where it drafts whatever I want it to draft, like if I want to make some bullshit corporate letter I never want to write myself but it has an explainer addition to where it explains why it did what it did and what training data was used to make the thing so I can understand why it did the thing, or like a training mode, like if I want to learn more about JavaScript, it has a whole training curriculum right there built into it that can be adapted to any learning style, including and especially, the blind. I honestly wouldn't even need it to describe pictures. Just teach me stuff and I'll be happy. That's my AI wish list, part of it anyway, but it doesn’t make people money or make data centers so, shrug. #AI #LLM
Had a weird moment in Claude's Memory settings. I left my cursor in the text box, it started cycling lines, and the first one said "My dog's name is Beans."
Beans was in my lap at the time - panting.
It's almost certainly placeholder text (the next line gave me a baseball career I never had). Still, for a few seconds I was a believer.
Has an AI ever seemed to know something it shouldn't?
Watched David Krakauer on StarTalk this weekend. His take: LLMs are capable, not intelligent. What I think they're missing is wisdom - intuition from experience.
So I had Claude interview me on whether an agent with memory could grow some. I landed on an adviser that argues with me but never makes the final call.
Would you want an AI that pushes back, or one that agrees?
Good morning, all.
My new video is up! In this one I talk about the allegations that ChatGPT and Anthropic stole prominent researchers' work via the users' inputs into their AI models as well as options you have to protect your data when using LLM's.
If you like the video, please tap the like button. This helps us smaller Youtubers out tremendously. Thank you!
#Technology #Tech #AI #LLM #Anthropic #ChatGPT #Fraud #Youtube #Video
Edited 25d ago