- Tune AI
- Posts
- 🇺🇦 AI Ambassador Speaks for Ukraine
🇺🇦 AI Ambassador Speaks for Ukraine
Microsoft Invests $3.9B in SEA, Apple Poaches AI Talent, SenseTime Takes on GPT4, Tesla Launches Robotaxi, and GitHub's New Copilot Workspace
Hello Tuners,
In today’s edition:
Ukraine Debuts AI Foreign Ministry Spokesperson
Microsoft Invests $3.9B in South East Asia
Apple Lures AI Talent from Google And Revives OpenAI Chatbot Talks
Chinese SenseTime Challenges GPT4-Turbo
Tesla Unveils Robotaxi Ride-Hailing App
GitHub Copilot Now Starts AI Projects
Ukraine Debuts AI Foreign Ministry Spokesperson
Oh, Ukraine, you're taking the whole "digital age" thing to a whole new level! They've just introduced an AI-generated spokesperson for their foreign ministry for the first time in history, and I'm both impressed and a little skeptical.
Meet Victoria Shi, the virtual rep based on Ukrainian singer and influencer Rosalie Nombre with 54,000 followers on her Instagram account. She's going to give updates on various issues, including the ongoing war with Russia.
I get why they did it - to cut down time and resources during wartime - but I'm curious to see how this AI spokesperson thing plays out. Will it be more efficient? Less prone to human error? And what about the whole "digital counterfeiting" thing? Is it the future of diplomacy or just a fancy gimmick?
Microsoft Invests $3.9B in South East Asia
Microsoft is investing a whopping $3.9 billion in Southeast Asia, with a focus on cloud and AI infrastructure in Indonesia, Malaysia, and Thailand. This is a major move to promote their generative AI technology in the region.
CEO Satya Nadella is on a mission to bring the latest and greatest AI infrastructure to these countries, and it's not hard to see why. With a huge, tech-savvy population, Southeast Asia is a key target market for tech-related investment.
This investment is not just about the financial benefits; it's also about building partnerships and creating opportunities for local businesses and developers. Microsoft wants to make sure that every organization and startup in the region can benefit from its AI infrastructure. And with their AI Centre of Excellence, they're committed to supporting the region's digital transformation.
With this investment, we can expect to see some serious innovation and growth in the region, and who knows, maybe one day we'll see a Southeast Asian tech giant giving the likes of Microsoft and Apple a run for their money!
Apple Lures AI Talent from Google And Revives OpenAI Chatbot Talks
The AI wars are heating up, folks! Apple is on a mission to poach the best and brightest AI talent from Google, and it's working! They've snatched up at least 36 specialists from their rival since 2018, including top AI executive John Giannandrea.
And it's not just about talent - Apple's also building a secretive European laboratory in Zurich, where they're working on the underlying tech that powers products like OpenAI's ChatGPT chatbot.
But what's the endgame here? Industry insiders think Apple's focused on deploying generative AI on its mobile devices, which would be a total game-changer. Imagine AI chatbots and apps running on your phone's own hardware and software, without needing cloud services!
It's a bold move, but Apple's been doing research on AI tech for over a decade, and they're not afraid to invest big. We might even see some of these new features at their Worldwide Developers Conference in June. The AI arms race is on, and Apple's not holding back!
Chinese SenseTime Challenges GPT4-Turbo
SenseTime from China has launched SenseNova 5.0, a large-scale AI model that's fast, cost-effective, and challenging GPT-4 Turbo's performance! This hybrid model combines cloud and edge computing, making it a game-changer for various industries.
Trained on 10TB of tokens, it boosts proficiency in knowledge, math, reasoning, and coding, with an inference speed of 109.5 words per second - 5x faster than the human eye can read! It reduces cloud server dependency and inference costs by up to 80%, making advanced AI more accessible and practical for everyday use.
Tesla Unveils Robotaxi Ride-Hailing App
Tesla has given a sneak peek into its Robotaxi ride-hailing app, featuring a "Summon" button, estimated wait times, and a 3D map displaying a virtual vehicle en route. The app allows passengers to adjust the climate, view wait times, and select music.
Tesla plans to debut the Robotaxi in August, with CEO Elon Musk likening the service to a blend of Airbnb and Uber. Musk stated that Tesla will manage the fleet, with some vehicles owned by the company and others by customers, who can add or subtract their cars from the fleet at will.
GitHub Copilot Now Starts AI Projects
GitHub has announced GitHub Copilot Workspace, a new service that uses AI to help developers start coding projects from scratch. This new feature is an extension of GitHub's Copilot platform, which is powered by GPT-4 and has been assisting developers with code completion.
With Copilot Workspace, developers can describe their project goals through prompts, and the AI will provide step-by-step suggestions on how to begin the project.
The suggestions can be edited, and once satisfied, the code can be run and completed with the help of Copilot. This new feature aims to reduce the time spent reading through code and documentation to figure out how to start a project.
According to Jonathan Carter, head of GitHub Next, Copilot Workspace can scan codebases and understand them faster than humans can read written documentation, making it particularly helpful for reviewing older code.
LLM Of The Week
Hermes 2 Pro is an upgraded version of Nous Hermes 2, featuring an updated dataset and new capabilities like Function Calling and JSON Mode.
It maintains excellent general task and conversation abilities, while exceling in Function Calling, JSON Structured Outputs, and more. With a 90% function calling score and 84% JSON output score, Hermes 2 Pro showcases improved agentic capabilities, making it a reliable and efficient model.
Weekly Research Spotlight 🔍
Multi-Layer Perceptrons are widely used in AI, including in Transformer models. They're like building blocks for processing data. But they use fixed functions to process this data, which can be limiting.
This paper suggests using custom functions instead, which can be tailored to specific tasks. It does this by replacing some parts of the model with special curves called splines. This makes the model more complex, but it also gives it some interesting features.
For example, it might be easier to understand how the model is making its decisions. This is important because AI models are often like black boxes - we don't know exactly how they work. By making the model more interpretable, we can trust its decisions more. This could lead to breakthroughs in areas like natural language processing and computer vision. Check it out here.
Best Prompt of the Week 🎨
“A red fox sniffing the wind, its muzzle raised upwards, eyes closed, high quality illustration, lots of detail, lots of color, romantic mood, in the style of light emerald and cyan, cartoonish realism, sentimental realism, 8k resolution, charming realism, mystical portraits,”
Today's Goal: Try new things 🧪
Acting as a Javascript Console
Prompt: I want you to act as a javascript console. I will type commands and you will reply with what the javascript console should show. I want you to only reply with the terminal output inside one unique code block, and nothing else. do not write explanations. do not type commands unless I instruct you to do so. when I need to tell you something in english, I will do so by putting text inside curly brackets {like this}. My first command is console.log("Hello World");
How did you find today's email? |
Missed an issue? Catch up on our past newsletters here.