AREC Inc.’s Post

VASA-1 takes in a single portrait photo and an audio file and converts it into a hyper realistic talking face video complete with lip sync, realistic facial features and head movement. The model is currently only a research preview and not available for anyone outside of the Microsoft Research team to try. Similar lip sync and head movement technology is already available from Runway and NVIDIA but this seems to be of a much higher quality and realism. Google Research also has something similar called VLOGGER AI. Potential use cases are in gaming, social media and film making. There are obvious ethical issues too, by making deepfakes easier to make, perhaps. The point, however, is that this AI based tech is gong to advance exponentially and there is nothing we can do about it. The thing we can hope for is that companies, especially Big Tech, put in the right guard rails and safety mechanisms before general availability. That seems to be happening – the guardrails around ChatGPT are quite impressive, text-to-video phenomenon Sora has not been released and OpenAI promises to build the right guard rails for that. Hopefully, the same is true for VASA, which currently is just a research project. Interesting article on CNN-News18 about VASA. Link in the comments. #ArtificialIntelligence #ChatGPT #Deepfakes

𝐍𝐕𝐈𝐃𝐈𝐀 𝐂𝐡𝐚𝐭𝐑𝐓𝐗 𝐓𝐫𝐚𝐧𝐬𝐟𝐨𝐫𝐦𝐬 𝐘𝐨𝐮𝐫 𝐋𝐨𝐜𝐚𝐥 𝐅𝐢𝐥𝐞𝐬 𝐢𝐧𝐭𝐨 𝐃𝐢𝐚𝐥𝐨𝐠𝐮𝐞𝐬! First it was Chat with your RAG, Reflecting on my earliest post https://lnkd.in/dEdCtsdy 😉 Now it's Chat with your Laptop 😎 At the heart of ChatRTX lies the innovative RAG Model, which enhances AI’s ability to provide accurate, relevant responses by referencing a vast knowledge base. Elevate Your Laptop with NVIDIA ChatRTX🌟 𝐊𝐞𝐲 𝐈𝐧𝐬𝐢𝐠𝐡𝐭𝐬 : 💻 𝑫𝒂𝒕𝒂 𝒑𝒓𝒊𝒗𝒂𝒄𝒚 & 𝑨𝑰 𝒂𝒕 𝒀𝒐𝒖𝒓 𝑪𝒐𝒎𝒎𝒂𝒏𝒅 𝒘𝒊𝒕𝒉𝒐𝒖𝒕 𝒄𝒐𝒎𝒑𝒓𝒐𝒎𝒊𝒔𝒆: ChatRTX offers AI functionalities on your PC without the internet, 𝙚𝙣𝙨𝙪𝙧𝙞𝙣𝙜 𝙮𝙤𝙪𝙧 𝙘𝙤𝙣𝙩𝙧𝙤𝙡 𝒐𝒗𝒆𝒓 𝒚𝒐𝒖𝒓 𝙙𝙖𝙩𝙖. 🤖𝑼𝒏𝒍𝒆𝒂𝒔𝒉 𝒕𝒉𝒆 𝑷𝒐𝒘𝒆𝒓 𝒐𝒇 𝑨𝑰 𝑪𝒐𝒏𝒗𝒆𝒓𝒔𝒂𝒕𝒊𝒐𝒏 𝒘𝒊𝒕𝒉 𝑵𝑽𝑰𝑫𝑰𝑨 𝑪𝒉𝒂𝒕𝑹𝑻𝑿  𝑪𝒉𝒂𝒕𝒃𝒐𝒕: ⚙️📚 𝑪𝒖𝒔𝒕𝒐𝒎𝒊𝒛𝒆 𝒀𝒐𝒖𝒓 𝑨𝑰 𝑱𝒐𝒖𝒓𝒏𝒆𝒚 𝒘𝒊𝒕𝒉 𝑪𝒉𝒂𝒕𝑹𝑻𝑿: Choose from diverse models at your fingertips and for specific needs. 📄🔍𝑨𝑰-𝑷𝒐𝒘𝒆𝒓𝒆𝒅 𝑪𝒐𝒏𝒕𝒆𝒏𝒕 𝑫𝒊𝒔𝒄𝒐𝒗𝒆𝒓𝒚: Pinpoint relevant information from various files. Let ChatRTX’s AI sift through documents to fetch the exact information you need. 🎙️ 𝑽𝒐𝒊𝒄𝒆-𝑨𝒄𝒕𝒊𝒗𝒂𝒕𝒆𝒅 𝑨𝑰: Whisper AI’s speech-to-text conversion streamlines your interaction with the AI. 📷 𝑵𝒆𝒙𝒕-𝑳𝒆𝒗𝒆𝒍 𝑰𝒎𝒂𝒈𝒆 𝑼𝒏𝒅𝒆𝒓𝒔𝒕𝒂𝒏𝒅𝒊𝒏𝒈: The CLIP model interprets images with an understanding that transcends metadata. 🌐 𝑭𝒖𝒕𝒖𝒓𝒆-𝑷𝒓𝒐𝒐𝒇 𝒀𝒐𝒖𝒓 𝑨𝑰: ChatRTX is set to expand its horizons, potentially including email conversations and more. #NVIDIA #RAG #DataPrivacy #GenAI #GenerativeAI #ML #AI #LLMs #LLM #TechInnovation #FutureOfAI #digitaltransformation https://lnkd.in/dTqTdS5x

Build the World's Fastest Voice AI Bot! 🚀 Want super-fast voice-to-voice response times? Here's how to achieve ~500ms with a LLama 3 voice bot: 1. Unified Hosting: Keep speech-to-text, LLM inference, and text-to-speech in the same container or cluster. 2. Efficient Audio Routing: Use WebRTC and edge networking to send audio over the internet. 3. Latency Optimization: Adjust voice activity detection and phrase endpointing timings carefully. 4. Top Tools: Use DeepgramAI for STT and TTS on CerebriumAI's serverless GPU infrastructure. Ready to create the future of voice AI? Start now! Follow Hancy Khalid Artificial intelligence Engineer for more tech info #ai #llm #chatbot #chatgpt #tech #trend

Interesting announcements in the world of AI this week. 👉 Cohere announces Aya, multilingual open-source model that should outperform existing models in 101 languages. Did not have change to test the Finnish capabilities as their hosted model only supported 22. 👉 Voyage AI announces their new large embedding model with 1536 dimensions. Previous smaller was nr. 2 on rankings, so it is expected this is the new SOTA model. 👉 OpenAI has included "memory" in ChatGPT, that remembers selected things from earlier discussions. Users can manage the memory and disable it if wanted. 👉 Andrej Karpathy leaves OpenAI to concentrate on "personal projects". Looking forward to hear more about them! 👉 Nvidia announced "chat with RTX", AI chatbot that runs locally on PC.

This Friday, we’re covering some key roll-outs, and a new tool for saving info as a busy multitasker 🧮 With Apple’s introduction of its Vision Pro, we’re ushering in a new era of spatial computing. Vision Pro is a revolutionary spatial computer that reimagines how people work, collaborate, and connect—providing new avenues of entertainment and complex interactions in immersive environments. Following the introduction of ChatGPT Enterprise and its quick adoption by industry leaders, now you can access a new self-serve plan for teams of all sizes. ChatGPT team provides access to advanced AI models like GPT-4 and DALL·E 3, and includes a dedicated collaborative workspace plus admin tools for management. Lazy.so is offering first access to a tool that advertises the ability to “capture at the speed of thought.” With one keyboard shortcut, you can jot down notes and save information, keeping your digital space ultra-organized. #VisionPro #ChatGPT #SMBtips

You can now interact physically with someone on the opposite side of the world. inFORM is a Dynamic Shape Display that physically produces 3D content, allowing users to interact with digital data in a physical manner. It can also interact with the physical environment by manipulating objects on the surface of a table. This technology allows remote video conference participants to be physically presented, promoting a strong sense of presence and the capacity to communicate physically at a distance. Contact us if you made a great Al tool to be featured: https://lnkd.in/dwsY5XBr #artificialintelligence #technology #chatgpt #generatieveai

By every passing day, I feel like AI will soon start managing ordinary humans, and the engine of that AI will be in the hands of extra ordinary humans. But it may soon slip out of Extraordinary Humans too specially when algorithm creators won't be able to control it's self-improvement! #aiadvancements #aiarchitecture

Really? Amazing in all forms

In which Industry do u think this technology can be used ?

This is more than AI... This is a Phygital experience, and a great opportunity to combine digital and physical capabilities to build a better world. ❤️ The concept of being able to physically interact with someone on the opposite side of the world is mind-blowing. This has the potential to revolutionize remote collaboration and communication. Imagine how this could be used in various fields like engineering, design, and even healthcare. Just imagine. 🧠 #ai #phygital #colaboration