AIGuys Digest | June 2024
🌟 Welcome to the AIGuys Digest Newsletter, where we cover State-of-the-Art AI breakthroughs and all the major AI news🚀
In this thrilling edition of June 2024, we’re diving headfirst into the ever-evolving universe of Artificial Intelligence. 🧠✨
Btw, if you want to up your AI game, please check my new book on AI, which covers a lot of AI optimizations and hands-on code:
🔍 Inside this Issue:
- 🤖 Latest Breakthroughs: This month it is all about YOLOv10, xLSTM, Mechanistic Interpretability, and AGI.
- 🌐 AI Monthly News: Discover how these innovations are revolutionizing industries and everyday life: Apple’s WWDC 2024, Kling: China’s Insane New Text-to-Video Generator, Claude Sonnet 3.5: The New #1 Chatbot in the World, and OpenAI Ex-Chief Scientist Ilya Sutskever’s Safe Superintelligence Project.
- 📚 Editor’s Special: This covers the interesting talks, lectures, and articles we came across recently.
Let’s embark on this journey of discovery together! 🚀🤖🌟
Follow me on Twitter and LinkedIn at RealAIGuys and AIGuysEditor.
Latest Breakthroughs
YOLO has been the undisputed king of object detection for many years. With this new release, it has become even faster. The paper introduced some cool new ideas like NMS-free training of YOLOs, which brings competitive performance and low inference latency simultaneously.
YOLOv10: Object Detection King Is Back
Before the quick rise of Transformers, LSTMs were the kings. LSTM or Long Short Term Memory was invented to solve the issues of the Recurrent Neural Network vanishing Gradient problem. Recently there was a lot of hype about Mamba, a state space model; LSTM could be thought of as a precursor to these state space models. But today, we are discussing a newer version of the LSTM called xLSTM, something that can not only compete with Transformers but in some cases even outclass them.
xLSTM vs Transformers: Which Will Win?
The ability to interpret and steer large language models is an important topic as we encounter LLMs on a daily basis. As one of the leaders in AI safety, Anthropic takes one of their latest models “Claude 3 Sonnet” and explores the representations internal to the model. Let’s discover how certain features are related to different concepts in the real world.
Extracting Interpretable Features From A Full-Scale LLM
In the last few weeks, the ARC challenge by the legend Francois Chollet has made quite some noise. It is a challenge that has puzzled a lot of AI researchers, demonstrating the generalization incapabilities of all the AI systems out there. The last SOTA AI on ARC was around 34% and on the same challenge, Mechanical Turks performed around 85%.
But recently, there have been new claims of achieving 50% on this challenge. So, the big question is did we really somehow increase the generalization capabilities of our AI systems, or there is something else happening in the background?
How We Suddenly Got 50% On The ARC-AGI Challenge?
AI Monthly News
Apple’s WWDC 2024
At WWDC 2024, Apple announced significant updates across its entire product lineup, focusing on enhancing user experience, privacy, and ecosystem integration. Moreover, the US-based technology giant revamped its digital assistant Siri with more capabilities powered by artificial intelligence and machine learning. Lastly, Apple debuted its personal intelligence system called Apple Intelligence, which leverages generative models for personalised interactions and integrates ChatGPT for advanced content generation. Here are key takeaways from Apple’s WWDC 2024 keynote address.
Apple WWDC: Click here
Kling: China’s Insane New Text-to-Video Generator
Kling AI boasts exceptional video quality and length capabilities, producing 2-minute 1080p videos at 30fps, which significantly surpasses previous models. It features cutting-edge 3D modeling techniques that utilize advanced face and body reconstruction to create ultra-realistic character expressions and movements. Additionally, Kling AI excels in modeling complex physics and scenes, effortlessly combining concepts that challenge reality. The proprietary Diffusion Transformer technology enables Kling AI to generate videos in various aspect ratios and shot types, offering unparalleled versatility in video production.
Kling AI website: Click here
Claude Sonnet 3.5: The New #1 Chatbot in the World
Anthropic’s new AI model, Claude Sonnet 3.5, is now the top chatbot, outperforming ChatGPT-4o in benchmarks. It’s twice as fast as Claude 3 Opus and excels in coding, writing, and visual tasks like explaining charts. Demonstrations include creating a Mario clone with geometric shapes, solving complex physics problems, coding a Mancala web app in 25 seconds, generating 8-bit SVG art, transcribing genome data into JSON, and diagramming chip fabrication. Despite lacking some features of ChatGPT-4o, Claude Sonnet 3.5 is praised for its speed, human-like writing, and ability to handle large documents.
Try it for free here: Anthropic
OpenAI Ex-Chief Scientist Ilya Sutskever’s Safe Superintelligence Project
Ilya Sutskever, co-founder of OpenAI, has launched a new venture called Safe Superintelligence Inc. This initiative focuses on developing a safe, powerful AI system within a pure research environment, free from the commercial pressures faced by companies like OpenAI, Google, and Anthropic. The aim is to push forward in AI research without the distractions of product development and market competition, ensuring that safety and ethical considerations remain at the forefront.
Source: CNN
Editor’s Special
- An old paper from Francois Chollet on the Measure of Intelligence: Click here
- Geoffrey Hinton | On working with Ilya, choosing problems, and the power of intuition: Click here
- Max Tegmark | On superhuman AI, future architectures, and the meaning of human existence: Click here
🤝 Join the Conversation: Your thoughts and insights are valuable to us. Share your perspectives, and let’s build a community where knowledge and ideas flow freely. Follow us on Twitter and LinkedIn at RealAIGuys and AIGuysEditor.
Thank you for being part of the AIGuys community. Together, we’re not just observing the AI revolution; we’re part of it. Until next time, keep pushing the boundaries of what’s possible. 🚀🌟
Your AIGuys Digest Team
