The Trump administration is considering new restrictions on the Chinese AI lab DeepSeek that would limit it from buying Nvidia’s AI chips and potentially bar Americans from accessing its AI services, The New York Times reported on Wednesday.

Just thinking about making this a monthly post, which model are you using? what are the positives and negatives?

Let's go! Lossless CPU inference

Ok, so the best thing that I have been using my local AI for lately is acting as Socrates and helping me delve into my thinking

::: spoiler Long text here

User | 2025/04/09 08:39:23

When considering the Socratic method, what questions did Socrates ask people, to start the process of enquiry

Gemma3 (12B) | 2025/04/09 08:39:55

Okay, let's break down the types of questions Socrates used to initiate the Socratic method. It's important to understand that it wasn't a rigid formula, but a flexible approach designed to expose contradictions and ultimately lead to deeper understanding. Here's a breakdown, categorized by their function and with examples. I'll also include some key principles at the end.

1. Initial Clarification & Agreement (Establishing a Starting Point - Often Seemingly Simple)

• "What do you mean by...?" / "What is X?" (Where X is a Virtue, Concept, or Idea - e.g., "What do you mean by justice?", "What is courage?", "What is piety?")

General consensus seems to be that llama4 was a flop. A head of meta AI research division was let go.

Do you think it was a bad fp32 conversion, or just unerwhelming models all around?

2t parameters was a big increase without much gain. If throwing compute and parameters isnt working to stay competitive anymore, how do you think the next big performance gains will be made? Better CoT reasoning patterns? Omnimodal? something entirely new?

We’re introducing Llama 4 Scout and Llama 4 Maverick, the first open-weight natively multimodal models with unprecedented context support and our first built using a mixture-of-experts (MoE) architecture.

So far I have tried to get it to read me my shoping list, turn on my desklight, and tell me how many hours until sundown. It wasn't able to do any of those things. I'm using llama3.1:8b. I have a 3080 12gb. I also tried mixtral:8x7b but it just says "unexpected intent error".

Im having some fun experimenting with gif making tonight hope you don't mind the animation. This shows the iterative creation process of our current thumbnail.

I kind of knew what I wanted in my minds eye with a front facing llama to juxtapose the old thumbnail sideways view. I went searching for AI generated llama images since thats fitting of the fourm and copyright free public domain (as far as I understand).

The original image was generated by Stable Diffusion. I like it a lot as is but to be a good thumbnail for the community it needed to be easily recognizable and renderable on small screens. First picked the color.The purple is more bright and slightly reddish in tinge which helps with pop. Then I expanded the neck to fill to the bottom.

All those detail lines created nasty artifacting when compressed on small phone icons of the community, They needed to go. I left the chin hair lines and eye lines to not make it too simple. The nose ridge outline was thickened for some recog

Every release from kobold has me hyped one of the nicest engines that balances cutting edge features with ease of use and optimization. This is gonna be a great year for LocalLLaMA Hype :)

https://github.com/LostRuins/koboldcpp/releases

Link to bluesky https://bsky.app/profile/tomaarsen.com/post/3llc2jvwah22f

Some more details https://huggingface.co/blog/train-reranker

Hi everyone! I recently became moderator of this community and have been making some changes. I figured it would be good practice to be transparent with you and document what's been going on.

Changelog

• The community sidebar now has a banner
• The description of our community has been rewritten and expanded.
• New thumbnail picture
• Un-pinned post on beginner questions / Reason: its better for beginners to make a new post asking their questions for community visibility/engagement.
• Un-pinned post on guide to LLMs / Reason: guide years old and somewhat outdated.

thumbnail picture changes

Ive been experimenting with some different thumbnails for our community. I didn't really want to keep associating with r/localllama in any way we dont need to copy them.

Old thumbnail

New Thumbnail:

![](https://lazysoci.al/api/v3/image_prox

I liked reading this article. Its cool to really poke into the hidden perplexity behind patterns of 'thought' in llms. They aren't merely simple 'auto complete'.

The finding that claude does math in a different way then it says it does and can anticipate words ahead of generation time are facinating.

I am unsure whether to buy the M4 Max Max Studio with 128GB ram and 500GB/s bandwidth or M1 Ultra Mac Studio with 128GB ram and 800 GB/s bandwidth.

The Max costs 4400€ The Ultra costs 3500€ second-hand

What would you buy? I am not going to use it only for LLMs but for coding, machine learning, gaming.

I upgraded my PC, I have just upped the RAM to 64GB and got a Radeon 9700xt 16GB GPU.

I want to do some personal programming projects and was wondering what is the best model that would fit my HW?

I mostly use python, but dabble a little in go, and am interested in looking at rust.

I am running this on my desktop, the processor is a AMD 3700X (8c/16t).

Removed