Coding 🤜🤛 Nostr

Opinion: French isn't a real language.

Building a private AI assistant for truly private chat and personal data processing (e.g., notes, PDFs). Planning to create a PWA web app that allows users to run any LLM model, including the latest LLaMA 3.2-1B, locally on any device or OS, even offline. Users will simply select the model, and the app will automatically download, cache, and store it for future use. The next time, loading will be much faster. The MLC-LLM library makes implementation straightforward, supporting all platforms and leveraging WebGPU for maximum performance. Coming soon, meanwhile check there library repo. They have lots of examples. MLC library repo:

WebLLM Javascript SDK — mlc-llm 0.1.0 documentation

If you're paying for X to support free speech, I have a bridge in Brooklyn to sell you.

Podcasting bro 😭

Test new Llama 3.2 1b/3b on any device locally with the web app. I have multiple apps that use this WebGPU implementation. Decent small LLM models were the only constant, but now we finally have the model.

WebLLM Chat

WebLLM Chat

Chat with AI large language models running natively in your browser

One of the Greatest!

Songlink/Odesli

My Beautiful Dark Twisted Fantasy by Kanye West

Listen now on your favorite streaming service. Powered by Songlink/Odesli, an on-demand, customizable smart link service to help you share songs, a...

1B & 3B models 👀 These should be able to run on iPhones models going back 5 years or so to around iPhone X

First step is complete. Now, we are extending this isolation to all the apps pinned in each AppCore variant. You should be able to log into 3 different Signal accounts using 3 AppCore variants. Each variant should behave like an isolated environment.

Nostur experience on iPad 🤍 It's my fastest and best multitasking experience on iPad so far. Perfect for both power users and casual users, it's very easy to navigate with all the necessary features.

3 key features in upcoming AppCore V2: 1. Create multiple copies of the AppCore app & run them simultaneously, separate from each other. 2. Categorization of apps with custom Profiles like Personal, Work. 3. Database will now be encrypted with password, in addition to being completely offline and local.

Marie Curie

This weekend, I will be starting a very interesting book that I’ve wanted to read for a long time, ever since I finished Mein Kampf. The book is called "Hitler's War by David Irving". It's also about World War II, but perhaps from the Axis point of view. I'm not entirely sure yet, as I haven't read it.

From all this AI saga, all I want is a general-purpose open-source LLM that is efficient enough to run locally on a phone, > 10 tokens/sec and smart enough to handle basic tasks like answering questions and reading PDFs with contextual understanding. No AGI, nothing, just a model, Light enough to run on a device locally, yet smart enough to handle essential tasks, that’s all I need, I'm a simple man.