Microsoft Introduces Cost-Effective AI Models Under Phi-3 Family Key points: Microsoft introduced three smaller #AI models under the #Phi3 #technology family. Even the smallest #model performs almost as well as #GPT-3.5 but at a lower cost. The smallest model can fit on a #smartphone and run on regular computer #chips. Smaller models require less processing, making them cheaper for customers to use. Microsoft believes customers will accept some trade-offs in performance for affordability. Eric Boyd, a Microsoft executive, highlights the potential for various industries, like online advertisers, to benefit from more affordable #AI. Microsoft is #opensourcing its Phi-3 models. Researchers focused on refining data and using #synthetic #data to build smaller but efficient models. Microsoft offers three sizes of the Phi-3 models: Mini, Small, and Medium, each with varying power and cost. via The New York Times
Rimah Harb’s Post
More Relevant Posts
-
Strategic Marketing Executive | $200M+ New/Existing Revenue | Built 10 Teams at 4 Hypergrowth Companies | GTM Message & Strategy | Artificial Intelligence | Cloud Native, SaaS, APIs, IaC, AI/ML | Senior Director | VP
New to AI? Here's a quick recap of the big news in the last week! - State of AI Report 2023 - Amazing report... but you need time to digest it. 100+ slides! https://lnkd.in/gedDuX5z - OpenAI plans major update to lure developers - Developers! Developers! Developers! https://lnkd.in/gTP4EA9Z - Google joins Microsoft to defend users against copyright claims https://lnkd.in/gcak-94q - Gartner predicts 80% of enterprises will adopt AI by 2026 - Up from 5% today. That seems crazy to me. We'll see... https://lnkd.in/g2P8z-3p - Microsoft announces new AI Tools to assist Doctors - This got me thinking, both Google and MSFT have recently released AI solutions to target healthcare. Will Healthcare be the breakout vertical market? https://lnkd.in/gkwS5nsw Microsoft to debut new AI chip next month - Rumors are always fun but I've heard this in a few places. OpenAI is allegedly also looking into custom chips. How much longer can NVIDIA hold out? https://lnkd.in/gHCgZB57 #ai #artificialintelligence #aiml #generativeai #genai #machinelearning #gpt #llm #nutnaix #machinelearning #AIOps #MLOps
To view or add a comment, sign in
-
Unlocking AI's Potential with Microsoft's Phi-3 Mini. Microsoft's recent unveiling of the Phi-3 Mini language model marks a significant step in making powerful AI tools more accessible and efficient. The Phi-3 Mini, a standout in the new Phi-3 family, boasts 3.8 billion parameters yet outperforms models twice its size. It's designed for both commercial use and open-source applications, available on platforms like Microsoft Azure AI Model Catalog and Hugging Face, and supports local, offline operations reducing latency and enhancing data privacy. The upcoming models, Phi-3-small and Phi-3-medium, promise further diversity in performance and scalability, catering to a wider range of computing needs and application scenarios. This development not only advances the capabilities of smaller scale models but also illustrates Microsoft's commitment to creating versatile, cost-effective AI solutions. #Phi3 #MicrosoftAI #LanguageModels #TechInnovation #AIForAll #GenerativeAI https://lnkd.in/dNFUQ9j7 Activate to view larger image,
#Technology #Strategy #FashionTech #DeepFashion #DesignAutomation #DigitalTransformation #MicroAutomation #DataDriven
Unlocking AI's Potential with Microsoft's Phi-3 Mini. Microsoft's recent unveiling of the Phi-3 Mini language model marks a significant step in making powerful AI tools more accessible and efficient. The Phi-3 Mini, a standout in the new Phi-3 family, boasts 3.8 billion parameters yet outperforms models twice its size. It's designed for both commercial use and open-source applications, available on platforms like Microsoft Azure AI Model Catalog and Hugging Face, and supports local, offline operations reducing latency and enhancing data privacy. The upcoming models, Phi-3-small and Phi-3-medium, promise further diversity in performance and scalability, catering to a wider range of computing needs and application scenarios. This development not only advances the capabilities of smaller scale models but also illustrates Microsoft's commitment to creating versatile, cost-effective AI solutions. #Phi3 #MicrosoftAI #LanguageModels #TechInnovation #AIForAll #GenerativeAI https://lnkd.in/dNFUQ9j7
To view or add a comment, sign in
-
-
Unlocking AI's Potential with Microsoft's Phi-3 Mini. Microsoft's recent unveiling of the Phi-3 Mini language model marks a significant step in making powerful AI tools more accessible and efficient. The Phi-3 Mini, a standout in the new Phi-3 family, boasts 3.8 billion parameters yet outperforms models twice its size. It's designed for both commercial use and open-source applications, available on platforms like Microsoft Azure AI Model Catalog and Hugging Face, and supports local, offline operations reducing latency and enhancing data privacy. The upcoming models, Phi-3-small and Phi-3-medium, promise further diversity in performance and scalability, catering to a wider range of computing needs and application scenarios. This development not only advances the capabilities of smaller scale models but also illustrates Microsoft's commitment to creating versatile, cost-effective AI solutions. #Phi3 #MicrosoftAI #LanguageModels #TechInnovation #AIForAll #GenerativeAI https://lnkd.in/dNFUQ9j7 Activate to view larger image,
#Technology #Strategy #FashionTech #DeepFashion #DesignAutomation #DigitalTransformation #MicroAutomation #DataDriven
Unlocking AI's Potential with Microsoft's Phi-3 Mini. Microsoft's recent unveiling of the Phi-3 Mini language model marks a significant step in making powerful AI tools more accessible and efficient. The Phi-3 Mini, a standout in the new Phi-3 family, boasts 3.8 billion parameters yet outperforms models twice its size. It's designed for both commercial use and open-source applications, available on platforms like Microsoft Azure AI Model Catalog and Hugging Face, and supports local, offline operations reducing latency and enhancing data privacy. The upcoming models, Phi-3-small and Phi-3-medium, promise further diversity in performance and scalability, catering to a wider range of computing needs and application scenarios. This development not only advances the capabilities of smaller scale models but also illustrates Microsoft's commitment to creating versatile, cost-effective AI solutions. #Phi3 #MicrosoftAI #LanguageModels #TechInnovation #AIForAll #GenerativeAI https://lnkd.in/dNFUQ9j7
To view or add a comment, sign in
-
-
#Technology #Strategy #FashionTech #DeepFashion #DesignAutomation #DigitalTransformation #MicroAutomation #DataDriven
Unlocking AI's Potential with Microsoft's Phi-3 Mini. Microsoft's recent unveiling of the Phi-3 Mini language model marks a significant step in making powerful AI tools more accessible and efficient. The Phi-3 Mini, a standout in the new Phi-3 family, boasts 3.8 billion parameters yet outperforms models twice its size. It's designed for both commercial use and open-source applications, available on platforms like Microsoft Azure AI Model Catalog and Hugging Face, and supports local, offline operations reducing latency and enhancing data privacy. The upcoming models, Phi-3-small and Phi-3-medium, promise further diversity in performance and scalability, catering to a wider range of computing needs and application scenarios. This development not only advances the capabilities of smaller scale models but also illustrates Microsoft's commitment to creating versatile, cost-effective AI solutions. #Phi3 #MicrosoftAI #LanguageModels #TechInnovation #AIForAll #GenerativeAI https://lnkd.in/dNFUQ9j7
To view or add a comment, sign in
-
-
#AI, AI's costly buildup poses a price problem Tech companies are touting new AI technology that can spit out business memos or computer code. They are still figuring out how those products will generate a profit. Microsoft has lost money on one of its first generative AI products. It and Google are now launching AI-backed upgrades to their software with higher price tags. Zoom Video Communications has tried to mitigate costs by sometimes using a simpler AI it developed in-house. Adobe and others are putting caps on monthly usage and charging based on consumption. "A lot of the customers I've talked to are unhappy about the cost that they are seeing for running some of these models," said Adam Selipsky, the chief executive of Amazon.com's cloud division, Amazon Web Services, speaking of the industry broadly. It will take time for companies and consumers to understand how they want to use AI and what they are willing to pay for it, said Chris Young, Microsoft's head of corporate strategy."We're clearly at a place where now we've got to translate the excitement and the interest level into true adoption," he said. #AI, #RB https://lnkd.in/eRcsxJGY
Big Tech Struggles to Turn AI Hype Into Profits
wsj.com
To view or add a comment, sign in
-
Microsoft is reportedly advancing an ambitious project known as "Stargate," aiming to build a $100 billion supercomputer to boost OpenAI’s AI capabilities, potentially culminating in the development of AI models surpassing ChatGPT-4. Slated for completion around 2028, the project requires substantial resources, spanning hundreds of acres and needing up to 5 gigawatts of power, marking a significant leap in the collaboration between Microsoft and OpenAI. This initiative, which awaits OpenAI’s delivery of a major AI upgrade expected by early 2025, not only underscores the scale of investment into AI technology but also prompts important discussions on its environmental footprint and ethical considerations, from energy consumption to labor practices. https://lnkd.in/e9fjKfPu #AI #Microsoft #OpenAI #Sustainability #EthicalTechnology #Stargate
Microsoft and OpenAI Allegedly Developing a $100B Supercomputer to Advance AI Models
https://theaiinsider.tech
To view or add a comment, sign in
-
🚀 **Big Move by Microsoft in the AI Arena!** Are we witnessing a new titan in the realm of AI? Microsoft is stepping up its game with the development of MAI-1, a large language model that might just give OpenAI a run for its money! 🧠💻 Here’s a quick scoop: - 🆕 Microsoft's new AI, MAI-1, aims to compete with OpenAI's ChatGPT and DALL-E. - 🎨 It’s designed to be a versatile jack-of-all-trades, adept in generating images, code, and text. - 🔧 Developed by the seasoned pros at Microsoft Research, this could integrate smoothly into tools we use daily like Bing, Teams, and Azure. 🤔 As someone who has dabbled in AI through my projects, I've experienced firsthand the pivotal role such capabilities play in enhancing productivity and creativity. Remember when we thought voice assistants were revolutionary? Now, imagine a tool that not only responds but also creates! ### Challenges Ahead While exciting, challenges in data handling and computing resources loom large for Microsoft. It’s a tech titan battle out there, and staying ahead means not just invention but also infrastructure. ### Why Should We Care? For professionals and businesses, this means more robust tools could soon be at our disposal. Improved AI could revolutionize how we handle data, customer interactions, and even software development. 👥 **What's your take on this development?** Do you think MAI-1 could change the game for Microsoft and its users? Could this shift the dynamics in tech giants' race for AI dominance? Share your thoughts or experiences with AI tools in your work or business! 📊 Excited to see how this unfolds! #AI #Microsoft #OpenAI #technology #innovation #business #ArtificialIntelligence #MicrosoftAI ✨ Let's keep the conversation going! If you have insights or questions about integrating AI into business processes, drop a comment below or shoot me a message! Let's leverage AI to push boundaries together! 🌐💡
To view or add a comment, sign in
-
-
Chief AI Officer | IEEE Fellow | Ex Global Head of Google Cloud AI R&D | Entrepreneur | Board&Advisory Board | Angel Investor| Hiring Generative AI Positions
Google shared their new open source models Gemma based upon similar tech and models of Gemini models. Two model sizes are released including 2B and 7B. You can run it on your laptop. Now Meta, Google, Microsoft, Apple have open sourced models. Who will be the next? 😁 #genai #generativeai #ai #llm #opensource #ml https://lnkd.in/gRdTvEwr
Gemma: Introducing new state-of-the-art open models
blog.google
To view or add a comment, sign in
-
Microsoft is reportedly developing a new AI model, internally referred to as MAI-1, that aims to compete with similar models from Google and OpenAI The new AI model is being overseen by Mustafa Suleyman, CEO of Microsoft AI, the co-founder of Google DeepMind and former CEO of Inflection AI. MAI-1 is expected to be “far larger” than the open source models previously trained by Microsoft, like Phi-3. The tech giant is reportedly using large clusters of servers equipped with Nvidia’s graphic processing units and vast amounts of data to improve MAI-1. The model is expected to have roughly 500 billion parameters. For comparison, OpenAI’s GPT-4 has one trillion parameters, and the recently launched Phi-3-mini has 3.8 billion parameters. Read more: https://lnkd.in/gfgKXNc6 #ai #llm #microsoft
To view or add a comment, sign in
-