Faction Inc.’s Post

Run Llama-3 on Someone Else's Computer Using the World's Largest Peer-to-Peer Network for AI Agents (100% Free) Here's how you can do: 1. Visit [aios.network](https://aios.network). 2. Install the client app on your laptop. 3. Open the app on your laptop. 4. Install "Llama-3" from the AI models section or use the model running on someone else's machine. #llama3 #opensource #aios #agents #ai

WOW - the GenAI Experience with DataStax and NVIDIA last night was super fun! Check out the recap video for the highlights. I saw some old friends and made new ones. Oh, and BoDi (the Born Digital rhino) looked adorable as an ice-sculpture! Make sure to follow DataStax to stay up to date on how they can help YOU with building and deploying AI apps. BTW - if you’re at re:Invent be sure to visit them at booth 1328!!! #AWSreInvent #DataStax #GenAI

Now that the availability of #GenerativeAI models and services are commonplace, collaborative research tools like this are an essential new frontier. It’s not enough to have smart models without the context of smart people and shared vision. I ❤️ this concept so much that I tried to patent it almost a decade ago. https://lnkd.in/gQ_Q_g5R When I started out the journey to build nLyze there were very few commodity components to fill the requirements for deep learning. Ten years ago, there just wasn’t enough foundation to build on. The part that will make or break these tools is #SecureAi and #PrivateAi. It will be very interesting to see how these tools handle protecting both users and sensitive models. Fantastic work, Hamna Aslam Kahn. Very exciting.

Run Llama-3 on another computer using the world's largest peer-to-peer network for AI agents (100% FREE) Here's how: Step 1: Go to https://aios.network Step 2: Simply install the client app on your laptop Step 3: Open the app on your laptop Step 4: Install "Llama-3" from AI models or use the model running on someone else's machine.

Running LLMs completely on-device, through browsers, is a THING. Unlike before, using #WebGPU, we can reach the accuracy and speed of GPT 3.5 - completely offline and on-device, which solves so many security and privacy concerns for Companies. Reminder: The first "ChatGPT" (which wasn't as good as today's gpt-3.5-turbo) cost OpenAI ~$3M monthly (for inference) per 10 million users when it was released. Only 2 years have passed since then. I used to play and deploy vision models to the browser, but this is another level of engineering. WebGPU and ONNXRuntime-Web can fuel real-world applications starting today. Congrats Microsoft, for providing both the model (Phi3) and the runtime (ONNXRuntime-Web) with metal/webgpu support. My gut feeling tells me that 6 months from now, models like #phi3 will run on iPhones, in real-time. LLMs will be available to use on-device through apps in the app store. How amazing is that? Demo: https://lnkd.in/d6FXvCcu

Microsoft Copilot AI knows well about the importance of hydration and how TakeSip can help you develop this habit. 🕸️ Visit our website 👉 https://lnkd.in/d4wUYddX 👩💻 Get an app for your macOS 👉 https://lnkd.in/d5AvGpeG #TakeSip #HydrationInnovation #HealthTech #StayHydrated #AI #Bing #Copilot

As I was building my Formula One Car Classifier, I wondered what it would take to deploy this model I had trained on my phone. 🤔 After a quick Google search, I came across the concept of "On-Device AI." A single Google search opened my horizons to what we have achieved in implementing AI on edge devices with relatively low computing power! I learned how On-Device AI helps in: 1. Increasing Privacy 2. Decreasing Latency In Generating Responses 3. Better Personalization I was introduced to many frameworks like Tensorflow Lite and CoreML that help deploy an actual Tensorflow/Keras model on a device. I have covered all my learnings regarding On-Device AI in the following blog post on Hashnode! Check it out to learn in detail how On-Device AI has been transforming our lives and how we can try it firsthand!  https://lnkd.in/eWcf57FC #OnDeviceAI #ArtificialIntelligence #MachineLearning #MobileAI #TechInnovation #DataPrivacy #TensorFlowLite #CoreML

#WWDC is always a big day for Apple, and today was no exception. It put a bright spotlight on private on-device AI, which is our jam. To show what's possible, here is a demo of the Knapsack app analyzing Mark's WWDC predictions (in a local txt file) vs the actual announcements -- with inference running entirely locally on a Mac. https://lnkd.in/gawp4svc

🚀 Check out Kent Weare's latest video in the "Innovate where you Integrate" series! Part 2 dives into the Retrieval Augmented Generation (RAG) pattern using Azure Logic Apps. Don't miss the demo where Kent pairs it with Business Process Tracking. Perfect for those keen on cutting-edge integration techniques! And yes, I agree upon the statement that GenAI solutions don’t need to be always in the center of gravity, but can be integrated with existing business processes to create benefits for client projects. Great work Kent Weare ! Thanx for sharing 👍 #LogicAppsAviators #Azure #AzureCloud #LogicApps #GenAI #RetrievalAugmentedGeneration #Integration

Google AI has just launched AI @ Edge: Torch, a high-performance inference engine for PyTorch models on mobile devices! 📱🔍 This cutting-edge tool empowers developers to deploy sophisticated AI models directly on mobile devices, enhancing the performance and efficiency of mobile applications. Key highlights include: High Performance: Optimized for speed and low latency, making real-time AI applications a reality. Easy Integration: Seamlessly integrates with PyTorch, enabling developers to leverage existing models with minimal modifications. Cross-Platform: Supports both Android and iOS, ensuring broad reach and compatibility. https://lnkd.in/d7PJrrSJ #ArtificialIntelligence #MobileAI #PyTorch #GoogleAI #TechInnovation #MachineLearning #EdgeComputing