Thread

Need to hand them some @Maple AI View quoted note →

We’re using AI to discover important insights that we previously had no clue about. View quoted note →

it’s happening rather we like it or not

Neither of them get it just from watching this clip. I wish I coulda been there to join the talk & question them into rethinking what they just said.

I don't expect McConaughey to know it because he's not a techie, but we can already do this through self-hosting and custom training your LLM. He's right, though: it's ideal to self-host your LLMs if you have the capabilities to do so.

#SelfSov/*****

He can afford a beefy PC for his private LLM from all the Salesforce Ai adverts he’s done. It’s like saying you like bitcoin and then shilling a bank to get your nut.

What is describing is a local llm running what they call a rag system. “Running a local LLM with a Retrieval-Augmented Generation (RAG) system gives you the power of AI without relying on the cloud. You keep full control of your data, run queries securely and privately, and enhance the model with your own knowledge base in real time. It’s fast, customizable, and cost-efficient—perfect for teams or individuals who want AI that works with their information, not just what the model was trained on.”

Can you guys create a chatbot like experiende for support questions on a website?

I tried asking ChatGPT for songs that fit a specific situational dynamic, and I think it's deliberately nerfed from doing so

are you guys doing fine tunes? the deeper we get into reasoning, the more important integrated ideas are going to be. search and rag are fine for the basics, but they'll never change the philosophy of the llm

Absolutely agree. That's the way. We can't imagine the consequences that sharing information with a public there may have in the future.

Right right right

This. Plus, run your GPU locally. This why tech skills matter. View quoted note →

In other words “thinking.”

I love @Maple AI and use it daily. Two quick questions if you’re reading… In Claude you can create a project folder. All the chats are saved there as are any documents you upload. And then it’s basically what McConaughey is wishing he had. Maybe I’ve missed this - if so let me know. Also, if you had a post about use cases for each model that’d be great. I use the Deepseak the most, then OpenAI. I’m just asking ask questions. Both are great but I’m curious what their strengths are and what the other models excel at. Thanks for accepting Bitcoin! Getting another account for a family member this weekend. View quoted note →