AI to take over air traffic control?

Good morning, FryAI army! We are gathered here today to explore the latest and greatest AI updates. 😁

🔨 AI TOOLS OF THE DAY

🎵 ImFeeling: Enter your emotion … discover a soundtrack that suits your mood. → check it out

💻 Swifto: Let your website optimize itself on AutoPilot using generative AI A/B testing. → check it out

🎥 Pipio: Easily create videos in minutes with your own on-demand digital actors. You do the typing, and they’ll do the talking. → check it out

U.S. AND VIETNAM FORM PARTNERSHIP OVER AI 🇺🇸

The United States is making deals in the Asia-Pacific, and this time it is not with China. 🇻🇳

What’s new? Amidst the tension between the United States and China, the U.S. and Vietnam have forged agreements worth billions of dollars in AI and semi-conductor business deals. These developments are spearheaded by prominent companies such as Boeing, Microsoft, and Nvidia, marking a significant upgrade in relations between the two nations that were once adversaries.

Want some details? During a joint press conference, Joe Biden focused on critical technologies like cloud computing, semiconductors, and AI. Key deals include Vietnam Airlines' $7.8 billion purchase of Boeing 737 Max jets, AI projects led by Nvidia and Microsoft, and the establishment of semiconductor design centers in Ho Chi Minh City. The commitment to resilient semiconductor supply chains, in particular, aligns with U.S. concerns about China's chip development.

What does this mean? The cooperation between the U.S. and Vietnam represents a commitment to build a resilient semiconductor supply chain and expand economic partnerships, fostering increased trade and investment for both countries as the global race continues to heat up in the AI development space.

META IS BUILDING THE STRONGEST AI MODEL YET 🦾

There’s a new AI model on the block … well, soon there will be. 🤖

What’s new? Meta, the parent company of social media giants Facebook and Instagram, is reportedly developing a powerful, open-source AI model to compete with OpenAI's most advanced systems. Meta's new AI model is expected to be "several times" more powerful than Llama 2, their latest model.

How will this work? Meta is actively bolstering its infrastructure to support this high-level AI system, acquiring Nvidia's H100 semiconductor chips, known for their immense processing power. While Microsoft, a key supporter of OpenAI, has collaborated with Meta to host Llama 2 on its Azure cloud platform, Meta intends to train this new model on its own infrastructure.

When will it be available? Unfortunately, it looks like we are going to have to wait a while for this news leak to come to fruition. This new model is reportedly set to commence training in early 2024, with a planned release in the following year, potentially arriving after Google's LLM, Gemini.

AI TO TAKE OVER AIR TRAFFIC CONTROL? ✈️

AI has an eye in the sky … literally. 👁️

What’s up? UK researchers have unveiled a groundbreaking project called Project Bluebird, which is aimed at revolutionizing air traffic control using AI.

What is the point? Project Bluebird is a collaborative effort between National Air Traffic Services (NATS), the Alan Turing Institute, Exeter University, and UK Research and Innovation, with substantial government funding. The project's objectives include optimizing flight routes for fuel efficiency, reducing aviation's environmental impact, mitigating delays, and addressing the shortage of air traffic controllers, who require three years of training.

How does it work? The researchers are leveraging NATS’ comprehensive database of past flight records (which contains over 10 million flight paths) to train their AI system. Currently, human controllers and AI agents are working together in simulations of real-life air traffic within the project's “digital twin” of UK airspace. The AI system is capable of directing aircrafts while maintaining a necessary minimum vertical separation of 1,000 feet and a horizontal separation of 5 nautical miles. Furthermore, it guarantees the safety of each flight path for a minimum of 15 minutes in case there is a breakdown in radio communication with the pilot, preventing any potential collisions.

Will this actually be implemented? While the AI system won't have the authority to make actual routing decisions for at least a few years, the project aims to conduct live "shadow trials" by 2026, directly comparing AI decision-making with that of human controllers. Successful outcomes could lead to extended operational trials before considering full implementation of a computer-controlled system. Richard Cannon, NATS’ research leader on Bluebird, said, “In every air traffic control system in the world all decisions are made by a human. We are not saying that we want to automate the skies over the UK but we are pushing the envelope as far as we safely can.”

TWITTER (X) TUESDAY 🐦

HAS AI REACHED SINGULARITY? CHECK OUT THE FRY METER BELOW