Windows betting on Copilot+ PCs!
World of GPT-4o, OpenAI new Search Challenger in Google's Search Space, Ray-Ban AI Glasses and more news!
The Future of Computing is Here: The All New Copilot+ PCs
Get ready to be amazed! Microsoft just unveiled a revolutionary new line of PCs - Copilot+ PCs. These machines are designed to be your ultimate AI companion, boasting cutting-edge hardware and built-in features that will transform the way you work and create.
Supercharge Your Work with Innovative NPUs
At the heart of Copilot+ PCs lies a brand new Neural Processing Unit (NPU). Unlike CPUs and GPUs, NPUs are built for AI tasks, using less power while delivering incredible efficiency. This frees up your PC's main processors for other things. This powerhouse is up to 20 times faster and 100 times more efficient for AI tasks compared to traditional PCs. Imagine editing photos in Photoshop or creating stunning videos with DaVinci Resolve – all with the help of AI that runs smoothly and directly on your device.
Never Forget a Thing with Recall
Ever spent ages hunting for a file you know you saw on your screen? Copilot+ PCs introduce a revolutionary feature called Recall. Think of it as a photographic memory for your device. Recall helps you virtually find anything you've seen on your PC. See snapshots of your past activity and use object recognition to jump back to specific emails, chats, or websites. It's like having a personal search engine for everything you've done on your PC. Plus, your privacy is always protected – everything is processed locally on your device, and you have complete control over what information Recall stores.
Break Down Language Barriers with Live Captions
The world is getting smaller, and communication is key. Copilot+ PCs make it easier than ever to connect with anyone, anywhere. Live Captions with translation takes spoken audio from over 40 languages and translates it into English subtitles in real-time. Whether you're catching up on a podcast, watching a foreign film, or collaborating with international colleagues, language barriers are a thing of the past.
Power, Performance, and Affordability
Copilot+ PCs are available for pre-order now, starting at an attractive price of $999. These sleek and stylish devices come from Microsoft Surface and also leading manufacturers like Acer, ASUS, Dell, HP, Lenovo, and Samsung. General availability kicks off on June 18th, so you won't have to wait long to experience the future of computing.
Copilot+ PCs are more than just powerful machines; they're intelligent companions designed to boost your productivity, unleash your creativity, and connect you to the world like never before. Are you ready to experience the future of computing?
The World of GPT-4o
Welcome to the world of GPT-4o. Open AI recently announced GPT-4o (o for omni ) the smartest AI multimodal.
What Exactly is GPT-4o?
GPT- 40 is the multimodal AI model that is designed by open AI. GPT- 4o extends its capabilities to text, audio and also vision. GPT-4o was announced by the chief technical officer Mira Murati on 13th of May , 2024. GPT -4o will be available as the new desktop app very soon.
GPT-4o can take the input in the form of Audio, Vision and Text and responds accordingly. The most interesting fact of this GPT- 4o is that it available completely for free. That means that it is really going to give a tough competition for Gemini Ultra and Claude Pro.
As mentioned above, GPT – 4o responds in the text audio and vision format. GPT- 4o will match the speed GPT – 4. However, It is constantly working on the understanding of the people who are not frequent in English. GPT- 4o can respond to the audio speeches similarly as humans with almost the same speed. GPT- 4o can analyze the pictures easily. The most interesting part is that it can combine all three that is text, audio and vision and then produce the output. For example, it can hum for you while analyzing the picture of a beautiful beach and simultaneously responding to the questions you’ve asked. How cool is that!
GPT- 4o will be widely used in the Day-to-Day Scenarios. For example, if you give the input in a single language. GPT – 4o can translate and respond back. Both in audio and in the written languages. This will be best used in customer service center resolving the customer issues.
Developers can fine-tune GPT-4o for specific applications. Want an AI that identifies bird species from audio recordings? Fine-tune it on bird calls! Need an AI that analyzes medical images? Fine-tune it on radiology scans! This customization allows GPT-4o to adapt to specialized domains, making it even more versatile.
In conclusion, GPT-4o's advanced architecture and ability to handle multimodal inputs make it a versatile and powerful tool for a wide range of applications. Whether you need an AI to identify bird species from audio recordings or analyze medical images, fine-tuning GPT-4o for specific tasks allows it to excel in specialized domains. However, as we leverage this cutting-edge technology, it is crucial to prioritise ethical considerations, ensuring transparency, accountability, and ongoing efforts to mitigate biases. By wielding GPT-4o responsibly, we can unlock its full potential while maintaining the highest standards of integrity and fairness.
Khan Academy and Microsoft Partner Up to Make Learning Fun and Accessible with the Help of AI
Get ready for buzzing classrooms filled with enthusiasm and personalised learning! Khan Academy has partnered with Microsoft to bring AI-driven teaching tools to millions of teachers in America. This is a game changer for both students and educators.
Artificial Intelligence That Makes Learning Entertaining
Think about a world where teachers no longer need textbooks that are boring, but can make interactive lessons through the use of AI. This is exactly what Khan Academy’s pilot scheme, Khanmigo for Teachers does. It employs artificial intelligence to suggest imaginative lesson plans like using marshmallows as a teaching tool on gas laws! With this practical approach, learning no longer becomes dull as it makes understanding difficult concepts easier.
Supercharging Teachers
Entertainment aside, Khanmigo is there to rescue teachers from undertaking repetitive tasks. Through the use of Khanmigo, they can create customized classes, recommend groups of students and modify reading difficulty in minutes! By this, they have more time on their hands which allows them to concentrate on what truly matters: being close to their learners. According to Khan Academy’s estimations, these tools could save an average teacher five hours per week!
Opening Doors for All Learners
And the best thing of all? Owing to the generosity of Microsoft, Khanmigo for Teachers is now free to every K-12 instructor in America. This means teachers in underserved areas, who are often hit the hardest by resource constraints, can now access these powerful tools. Khan Academy is also looking at ways to make AI tutoring even more inexpensive and reachable in future.
The Future of Learning is Collaborative
This alliance does not exclusively involve Khan Academy and Microsoft. They are teaming up with educators and students as well to create the best learning experience possible. By incorporating Khan Academy content into Microsoft’s Teams for Education, students will have more access to high-quality educational resources. Additionally, they are working on creating new open-source AI models tailored specifically towards math tutoring.
The Takeaway
Education has never been this interesting before. Teachers can utilize tools such as Khanmigo, powered by AI to produce lessons that captivate their students’ attention, make learning more personalized and get enough time for them individually. More importantly, it sets a foundation where education becomes accessible, efficient, & yes even fun!
Open AI’s New Search Engine.
A search engine that directly competes with Google, the search engine king, is set to be unveiled by Open AI. This product is combined with the GPT-4 large language models and Microsoft Bing. This product will revolutionize the web search questioning the Google search engine existence.
This product was rumoured to launch on the 13th of May. However, the exact date of the announcement isn’t confirmed.
As mentioned above, this is said to use the large language models that are used to process the queries that are asked in the natural language and also it is said to tie up with Microsoft Bing to provide real-time search results, ensuring up-to-date and accurate information.
By this, Open AI can tailor the responses accordingly to the user questions. This will also improve the relevance and accuracy.
Silicon Valley has been buzzing about this open AI search engine. It is rumored that the search engine will use the “search.chatgpt.com” domain for this search engine. Reuters also reported that the openAI is planning to launch a search engine using LLM( Large Language Model).
However, Google is dominating the search engine. OpenAI search engine could question the status of the Google search engine.
However, we still have to wait for the official announcement about the search engine from OpenAI. So, Stay tuned for more updates!!
Ray-Ban Meta smart glasses.
Ray-Ban Meta smart glasses. The eyewear combines the iconic Ray-Ban design with cutting-edge technology.
These Ray-Ban smart glasses have a Meta AI assistant which is called Meta AI. Users can activate it by simply saying “Hey Meta!”. Powered by Llama 3, an advanced open-source AI model designed to rival or surpass OpenAI's GPT-4, the Meta AI assistant offers natural and context-aware interactions, optimized for devices like smart glasses and smartphones. No need to unlock the phone or press any buttons. these glasses boast enhanced real-world perception through multimodal understanding, processing multiple inputs like spatial awareness, speech interpretation, and image capture through discreet cameras, providing practical applications such as sartorial advice, language translation, and object description. Meta AI, the user can make calls, send texts, control music, and get answers to questions like “Hey meta what kind of plant is this?”. This is so practical for everyday usage.
Users use smart specs to stream music from Amazon Music directly. Music can also be controlled using voice, ensuring a hands-free experience for the user.
These smart glasses can also be used for meditation, improving physical activity, and self-care. Users can simply say “Hey meta play music for deep meditation”. All these activities don’t require any additional surprises. This can be performed by merely logging into the meta-view app. Meta also offers a month-free subscription to new users!
Last but not the least, these smart glasses are launched in different styles
These smart glasses represent the next generation of wearable tech. These glasses improve daily tasks and provide a hands-free, natural user experience by integrating the potent Llama 3 AI model and offering advanced multimodal capabilities. With stylish and useful Ray-Ban Meta smart glasses, you can bring the future into the present, whether you're a tech enthusiast or just looking for a practical way to interact with AI.
Stay tuned for more updates.
Microsoft's Phi-3 Family Gets a Visionary Boost: Phi-3 Vision
Remember those bulky AI models that hogged your phone's resources? Well, Microsoft is shaking things up with the new Phi-3 family of small language models, and the latest addition, Phi-3-vision, is a game-changer!
Phi-3-vision is a multimodal superstar, combining the power of language understanding with the ability to see! This 4.2-billion parameter powerhouse can analyze images, understand text within them, and even answer your questions about charts and diagrams.
Here's what makes Phi-3-vision stand out
Sees the Bigger Picture: Unlike its text-focused siblings (Phi-3-mini, Phi-3-small, and Phi-3-medium), Phi-3-Vision boasts "multimodal" superpowers. It can analyze both text and images, making it a true whiz at tasks like answering questions about charts, deciphering text within images (think captions or signs!), and even understanding complex diagrams and tables.
Mighty Yet Mini: Don't let its smarts fool you! Phi-3-Vision packs a punch in a compact package. This lightweight model, with a size of 4.2 billion parameters, is designed to run smoothly even on mobile devices. This makes it perfect for on-the-go tasks and applications that require minimal computing power.
Efficiency Champion: Phi-3-Vision outperforms many larger models when it comes to general visual reasoning and tasks like Optical Character Recognition (OCR) – that's fancy talk for turning images with text into usable data. Plus, it delivers impressive performance compared to bigger models on tasks like understanding charts and tables.
Phi-3-vision isn't the only star in the Phi-3 family
Phi-3-mini: This tiny titan boasts strong reasoning capabilities for language tasks, making it perfect for content creation or Q&A.
Phi-3-small & Phi-3-medium: Offering a balance between size and power, these models excel at various language tasks and analytical problem-solving.
Microsoft prioritises responsible AI, and the Phi-3 family is no exception. These models undergo rigorous safety checks and are trained with high-quality data to ensure responsible development and deployment.
Ready to See the Future?
Phi-3-vision is currently in preview, but you can get started with other Phi-3 models on Azure AI Playground. Explore the possibilities and build amazing AI applications with the Phi-3 family!