Simon Willison’s Post

Advancing AI governance in Europe and internationally - EU Policy Blog (microsoft.com) https://lnkd.in/gzcDud9y "As the EU finalizes the AI Act, the EU could consider using procurement rules to promote the use of relevant trustworthy AI frameworks. For instance, when procuring high-risk AI systems, EU procurement authorities could require suppliers to certify via third-party audits that they comply with relevant international standards." Microsoft Vice Chair & President Brad Smith on Advancing AI Governance in Europe and Internationally - YouTube Microsoft Vice Chair and President Brad Smith’s keynote speech at the "Europe’s Digital Transformation – Embracing the AI Opportunity" event in Brussels on June 29, 2023 where he discussed Microsoft’s ongoing commitment to developing AI responsibly and finding a viable path to advancing AI governance internationally. Brad Smith Michael Charles Borrelli #euaiact

Excited to share this insightful article from Microsoft WorkLab: "Why Using a Polite Tone with AI Matters." 🤖💼 The integration of AI into our daily lives is a fascinating journey, and this piece from Microsoft WorkLab delves into why maintaining a polite tone with AI is more important than we might think. 🔗 Link to the article: Why Using a Polite Tone with AI Matters For the past few years, my family has adopted a courteous approach with our smart home devices. Instead of just shouting commands, we've been saying "please" before our requests and responding with a heartfelt "thank you." The result? A positive vibe in our interactions with technology. Even my kid has picked up on this habit! Extending this experience to interactions with Microsoft Co-pilots and Chat GPT has been seamless. It's incredible to witness how a touch of politeness not only shapes our interactions with AI but also influences the overall experience. Let's continue fostering a culture of respect in the digital realm. How have you approached AI interactions in your daily life? Share your thoughts below! 👇 #AI #TechEtiquette #PositiveTechInteractions

FRIENDS ......HOW FAST AI ENTERING IN MOST OF THE FIELDS OF INTELLIGENCE ..... TRY OUT COPILOT WHICH ASK QUESTIONS TO BE FRAMED AND WITHIN SECONDS IT WILL BRING A HOST OF INFO FOR YOU ONLINE SO TRY OUT put on google window https://microsoft bing.........choose copilot from page down below it will give space to write your query and then you get instant answer whatever is stored.

I have been scared of AI to be honest and have been reading from the sidelines and figuring out how it can be of use. Here is what I have learned using a desig AI led system. Designer. Microsoft.com You need to be clear what you are saying and requiring. The instructions need to be like you are talking to someone who doesn’t know or get you. Some people find it easier I must say, I am still in the experimental and learning phase. I think AI can benefit a lot of people and has a huge role, but it does require that you brief it properly. #smallbusinesstips #smallbusinessbranding

One of the challenges of open-source software is that it often receives far less recognition and publicity than it deserves. The announcement of Apple's new intelligence feature exemplifies this issue: rather than headlines like "Apple adds ChatGPT integration," the news should highlight "Apple uses open-source AI to build on-device Apple intelligence." Source: https://lnkd.in/dYugkUFB #apple #generativeai #artificial #intelligence #GenAI Apple

🎉 2023 was a major year for generative AI. It went from research labs to real life with millions of people using it through popular tools like ChatGPT and Microsoft Copilot. 🚩 This year, AI is expected to become more accessible, nuanced and integrated in technologies that improve everyday tasks and help solve some of the world’s most challenging problems by Vanessa Ho. Here are three important AI trends to watch in 2024: 👉 1. Small Language Models (SLMs): Imagine AI that's powerful but runs on your phone!  SLMs are making AI more accessible, performing well on tasks like complex questions (think Copilot!). This is a game-changer for affordability and innovation. 👉 2. Multimodal AI: AI that sees, hears, and reads?  Multimodal models understand information from text, images, and even video. This is leading to smarter search tools, creative apps like Microsoft Designer that generate images from descriptions, and even custom AI voices for people with disabilities. 👉 3. AI for Science: Get ready for faster scientific breakthroughs! AI is tackling global challenges like climate change and disease. Researchers are using AI for weather prediction, sustainable agriculture, and even developing new drugs and materials.  This could revolutionize scientific discovery! Read more at https://lnkd.in/gx6_5TZP 3 big AI trends to watch in 2024 (microsoft.com) 🤗 At knowdroids, we certainly follow these trends. Especially, the SLMs in open-source as well as multimodal AI will be very important elements in what we deliver to our clients.Check out more and get in touch at knowdroids.ai – Multiply the speed, efficiency and knowledge utilization of your organization with AI autonomous agents. #GenAI #AI #MachineLearning #FutureofTech  #Microsoft

A very simple way to see where the market is heading is just by seeing the amount of NVIDIA cards each player is purchasing. For data + nvidia cards are the backbone of ML today. It is clear below the market is heavily divided in 2 main LLMs: chatGPT by OpenAI (here represented by its cloud provider Microsoft), and LLaMA, represented by Meta. The 3d is Google with only 1/3 of the GPU's sold, which is surprising considering all LLMs today are based on its BERT model and its attention algorithm. In fact, it was thought that Google had somehow let the LLM tech escape, as their latest model (PalM), despite being huge (650B) was not really delivering. But yesterday Google surprised again with Gemini, a new LLM which supposedly outperforms ChatGPT 4. As soon as I have done some testing (a little hard to, as now they made it into a closed system) I will let you know!

Soon, AI language models running on our smartphones will be the norm. These AI models will operate “at the edge,” not connected to the cloud to minimize latency and maximize privacy. Today, Microsoft announced Phi-3, a family of the most capable and cost-effective small language models (SLMs). Here’s how Phi-3 compares to other models of similar size.

You (likely) have heard that #NextGenAI is already changing the way we work. What is it capable of and how should you prepare for it? ⬇️ Microsoft asked 31,000 people around the world about work, AI, and the new skills we’ll need in the future. 📝Take the quiz to find out how you stack up! How Does Your Experience at Work Measure Up? (microsoft.com) #AI #msftadvocate

Auch besonders lesenswert zu den Responsible AI principles und zum "Private Cloud" Ansatz von Apple: "... _Empower users with intelligent tools: We identify areas where AI can be used responsibly to create tools for addressing specific user needs. We respect how our users choose to use these tools to accomplish their goals. _Represent our users: We build deeply personal products with the goal of representing users around the globe authentically. We work continuously to avoid perpetuating stereotypes and systemic biases across our AI tools and models. _Design with care: We take precautions at every stage of our process, including design, model training, feature development, and quality evaluation to identify how our AI tools may be misused or lead to potential harm. We will continuously and proactively improve our AI tools with the help of user feedback. _Protect privacy: We protect our users' privacy with powerful on-device processing and groundbreaking infrastructure like Private Cloud Compute. We do not use our users' private personal data or user interactions when training our foundation models. https://lnkd.in/ejdmbJWh