HaiVE’s Post

Open-Source Multi-Modal Local AI Alert! Edit: The audio is inaudible because of the background flight noise, so I've uploaded the video with manually typed subtitles in YouTube and here's the video: The reason I still chose to post it this way is that, I could think of no other way to authentically showcase that we are in the air and in a flight that offers no internet and we are still able to get multi-modal inference locally, and this product is now free and open-source! Hi everyone! Like millions of others, we at HaiVE have benefitted enormously from the open-source community with Meta, Mistral AI & others contributing capable foundational models and rockstars like Eric Hartford releasing awesome versions like Dolphin. We've decided to give back to the community by opensourcing one of our crown jewels: HaiVE AI Studio. A multimodal local inferencing solution for Mac, Windows, Linux and Android. Think of it like ChatGPT running on your laptop or phone that you never have to pay anyone subscription for. It has voice and text now, will add support for images and video in the coming months. RAG & these features are super hard to standardise into a local product :( so taking a much longer time. In this video, you can see our HaiVE AI Studio running on my laptop and I am 35000 feet in an Air India A350, which everyone knows does not yet offer internet. So I could think of no better way to showcase the product's offline multi-modal capabilities. If you would like to be the first to get beta access weeks before we release everything to the public, ping us [email protected] What can this be used for? This product is free and open-source. (Free for both personal and commercial use) 1. You can install this on your computer and use it privately so that no big tech will ever get your chat data. So you can load up an uncensored model and make it your AI girl friend and the whole thing is like an incognito window ;) 2. You can install this for your team or companywide and loading any GGUF model of your choice, censored or uncensored or your own foundational model for your company. All your and teams' chats will be private so none of your company's sensitive details need to be sent to any 3rd party AI service like OpenAI or Google, heck even the data doesn't even need to travel outside physically like if you're using a private Azure AI cloud service or AWS Bedrock or SageMaker. 3. If you are an AI consultant or a firm, take this, modify it and charge your clients for it Thanks to these wonderful people for making this happen: Deepika Loganathan Aravinth Ramesh kesavan sekar Raj Kumar Muthu kumar N Pritam Cristpin Reddy Andrew Rothwell Loganathan Dir murugesan gnanasekaran Mohan K Jai Kumar Tarun Dua Jonathan Ross Kesava Reddy Jibin Sabu Dr.Mahesh Kumar Sham Sundar Meyyappan M Dr.Diwan baskaran Mageswar Radhakrishnan Gautham Palaniswamy Nethya Sri Subramani Shrreyans Mehta Anuradha Chowdhary MARSHAL AM Samuel Jerome Saarunathan T

I think I see where "AI PCs" are going. I have been experimenting with OpenAI APIs (remote), and llama3 (local). It's a situation where you may or may not be tending to the screen and mouse/keyboard all the time. So, it will basically turn into major accessibility improvements. It explains some of the more scary-sound plans like "remember everything you ever did". With a local Language Model (like llama3). - run a database of past messages to retain context. - run a local vector database to index this context by similarity. - run a local Language Model, and be able to reach out to a better remote one 1) Stand up and go to the kitchen, and say "Ada, bind my bluetooth headphones". And perhaps you also ask to bind a text terminal (ie: braille for some people). 2) "Ada, get me connected to the local wifi. It's called 'foundationcoffeeroasters' and the password is 'joytotheworld'" 3) "Ada, I know that I have Tagalog listed as a language; but I want all content converted to English, including Tagalog; for all of my web browsing". 4) "Ada, leave all instant messages in their original language" 5) "Ada, read me all the news about the CrowdStrike thing." 6) "Go through my family photos, and describe the latest images to me" (which involves taking screen shots and analyzing what is on the screen -- in pixels). One thing that is a prerequisite is improving the awful state of screen reading. A screen reader should be a text shell, just like bash zsh or powershell; so you can reach it remotely over ssh. This is because Braille plus headphones is all drive by text input. Just like stdin,stdout,stderr file handles; there should be also stdbrl, stdspeech text output. This would be an amazing gift to blind users; to have reallly good accessibility improvements. But we need to do it, because these blind users are celebrities like: Claude, ChatGPT, Gemini, etc. Accessibility doesn't work if you can't just let a chat model ssh into your system and use it like any other blind user would. This is kind of like computers playing Atari or Dota2 from just the pixels. A competent assistant can work in the screen reader as a blind user, and requesting screen descriptions as needed to complete the task.

📢📢📢 New Blog Post : Getting Started with Chrome Built-in AI Chrome Built-in AI integrates AI models, including large language models (LLMs), directly into the browser. This gives you access to Gemini Nano model right within your browser. Check out the article that covers the following: 🏁 Key features of Chrome Built-in AI 🏁 Getting Started and Configuration Steps 🏁 Sample code to demonstrate Sentiment Analysis that prompts the local Gemini Nano model in the browser 🏁 Join the Google Chrome Built-in AI Challenge Read it here: https://bit.ly/3YcMsUO Github Repo : https://lnkd.in/dsTMnXf7 #Google #GoogleChrome #Gemini #GeminiNano

Open Source AI | "According to the leaked document, the engineer thinks open-source AI will outclass Google and OpenAI because the open-source community has now solved the “major open problems” and made it accessible to the general public. OpenAI’s and Google’s models are still better in quality, the document said, but "open-source models are faster, more customizable, more private, and pound-for-pound more capable.""

Clarifying the Confusion: "GPT Next" is Not a New Model OpenAI has recently clarified that "GPT Next" is not a new model, but rather a figurative placeholder to indicate the potential evolution of their models over time. The term was mistakenly interpreted as a literal new model, leading to confusion in the AI community. The Current State of OpenAI's Models OpenAI's most advanced model currently is GPT-4o, and details about GPT-5 are limited. The company is continuously working on improving its models, but there is no new model called "GPT Next". Fundraising Round and Partnerships OpenAI is reportedly in talks with tech giants Apple, Nvidia, and Microsoft for its latest fundraising round. This could lead to significant investments and partnerships that could further accelerate OpenAI's research and development. The Future of AI Models The clarification from OpenAI highlights the importance of accurate communication in the AI community. As AI models continue to evolve, it is essential to have clear and transparent information about their development and capabilities. Join the Conversation: What do you think about the potential evolution of OpenAI's models? Share your thoughts! How do you think partnerships with tech giants will impact OpenAI's research and development? Share your ideas! Salesforce acquires data management firm Own for $1.9B in cash Salesforce acquires data management firm Own for $1.9B in cash. Own, a provider of data management and protection solutions, is Salesforce's biggest deal since buying Slack in 2021. The acquisition strengthens Salesforce's commitment to secure data solutions and enhances their data protection and management capabilities. Own, valued at $3.35 billion, offers data backup tools and services, and complements Salesforce's existing data management tooling.

📢 Microsoft Steps Down from OpenAI Board Observer Role 📢 In a significant move reflecting the evolving landscape of AI governance, Microsoft has decided to step down from its board observer role at OpenAI. This decision comes amidst growing discussions about the ethical and strategic direction of artificial intelligence development. 🔍 What’s Happening? Microsoft’s exit from the observer role highlights the company’s shift towards a more integrated and collaborative approach with OpenAI, focusing on their existing partnership rather than oversight. This transition is a testament to the maturing relationship between the tech giant and the AI research organization, emphasizing trust and mutual growth over traditional governance structures. 📈 Why It Matters The most intriguing aspect of this development is how it mirrors broader industry trends. As AI technologies advance rapidly, the need for flexible and dynamic collaboration models becomes crucial. Microsoft’s move may signal a new era of partnership strategies in the tech world, prioritizing innovation and agility over formal roles and titles. As we witness these shifts, it’s essential to consider how traditional roles and structures may need to adapt in the face of rapid technological advancements. This move by Microsoft prompts us to reflect on how organizations can best position themselves to support and drive innovation in the AI sector. For more detailed insights, read the full article here: https://lnkd.in/dQdCsqvc #AI #ArtificialIntelligence #TechNews #Innovation #Microsoft #OpenAI #DataNatives #AIethics #FutureTech 🌐🚀

The Setup: Microsoft & OpenAI Partnership: Microsoft made a big investment in OpenAI, starting with $1 billion, and eventually increasing that to $10 billion by January 2023. They were working closely together on AI advancements. The Drama Begins: November 2023 – Sam Altman Gets Ousted: Out of nowhere, the board of OpenAI fired Sam Altman, the CEO, claiming he wasn’t being honest in his communication. This shocked the entire tech industry because Sam had been leading OpenAI's success, especially with things like ChatGPT. Microsoft Steps In: Satya Nadella's Reaction: Microsoft's CEO, Satya Nadella, stayed calm. He publicly supported Sam Altman and said that Microsoft was still fully committed to its partnership with OpenAI. A Job Offer: Within just three days of Sam’s firing, Nadella even offered Altman a job at Microsoft to lead a new advanced AI research team, which showed how much confidence Microsoft had in him. The Comeback: Altman Returns: Shortly after, due to overwhelming support from employees and others in the tech world, Sam Altman was reinstated as the CEO of OpenAI. The board also brought back Greg Brockman, OpenAI’s president, who had left in protest when Altman was fired. Moving Forward: Satya's View on the Outcome: At a tech conference in 2024, Nadella said the whole incident felt like "a long-lost memory" and that everything had worked out well in the end. He praised how everyone came together for a good outcome. What's Next: Microsoft Builds Its Own AI: While still working closely with OpenAI, Microsoft is also building its own AI model called MAI-1, led by big names in AI, including Mustafa Suleyman from DeepMind and Karén Simonyan from Inflection AI. OpenAI’s Latest Model: Meanwhile, OpenAI is advancing its own AI efforts, recently releasing a new AI model, o1, which is designed to think more like humans by producing a "long chain of thought." The Future: More Investments: OpenAI is in talks to raise an additional $6.5 billion, potentially valuing the company at a staggering $150 billion. Microsoft is reportedly considering investing again. In simple terms, Microsoft and OpenAI went through a rough patch when Sam Altman got fired, but things settled quickly, and both companies have continued to push forward in AI research. Microsoft even used this opportunity to strengthen its own AI initiatives while maintaining its partnership with OpenAI.

Interesting calculation for the cost of using LLM with different options. #generativeai #llms https://lnkd.in/g_eJfnFS

The Assistants API enables developers to easily build powerful AI assistants within their apps. This API removes the need to manage conversation history and adds access to OpenAI-hosted tools like Code Interpreter and File Search. The API also supports improved function calling for 3rd party tools. This FAQ is very useful, thanks OpenAI What is an Assistant? An assistant represents a purpose-built AI that uses OpenAI’s models and can access files, maintain persistent threads, and call tools. What is a Thread? A thread is a conversation session between an assistant and a user. Threads simplify application development by storing message history and truncating it when the conversation gets too long for the model’s context length. What is the File Search tool? The file_search tool implements several retrieval best practices out of the box to help you extract the right data from your files to augment the model’s responses. For more information, please read our developer documentation. By default, the file_search tool uses the following settings: Chunk size: 800 tokens Chunk overlap: 400 tokens Embedding model: text-embedding-3-large at 256 dimensions Maximum number of chunks added to context: 20 Continue —> 🤖 #generativeAI #machinelearning #ai #deeplearning #openaiassistant #AI #ArtificialIntelligence