The AI Canvas Newsletter #18
Delve into the latest AI advancements: Boston Dynamics' Electric Atlas, Llama 3, Udio, Grok-1.5V, The Silicon Shift and more.
The AI Canvas Newsletter #18
The AI Canvas: Your weekly palette of inspiration, insights, and innovation in the world of AI.
- 🤖 Electric Atlas: Boston Dynamics' new Atlas, enhancing industry automation with AI.
- 🧠 Meta Llama 3: A breakthrough in open source AI, offering advanced reasoning and code generation with a focus on safety.
- 🎶 Udio: Revolutionising music creation with AI, enabling intuitive, text-prompted musical compositions.
- 🌌 Grok-1.5V: Bridging textual and visual understanding, setting new benchmarks in multimodal AI capabilities.
- 🖥️ The Silicon Shift: Tech giants like Google, Intel, and Meta drive the future with custom AI chips, promising unprecedented performance and efficiency.
Written by Oli Wilkins.
Introducing Boston Dynamics' Electric Atlas
Boston Dynamics has launched its electric Atlas robot, aimed at practical applications in various industries. This new version of Atlas, developed in partnership with Hyundai, builds on the company's experience in robotics, promising improvements in automation and efficiency. The robot is designed to work in real-world settings, supported by advanced AI and machine learning, alongside a new software platform for better operational management.
Read more here and be sure to check out the video inside!
Meta Llama 3: The Latest Advancement in Open Source AI
Meta's unveiling of Meta Llama 3 marks a significant milestone in the open source large language model landscape, delivering cutting-edge performance. Available across major cloud services and supported by key hardware manufacturers, Llama 3 offers enhanced capabilities in reasoning and code generation, compared to Llama 2.
Read more in Meta’s blog post.
Udio: The Next Step in AI-Powered Music Generation
Following our showcase of Stability AI's music generator in the last edition, we now turn our attention to Udio, the latest entrant in the AI music generation arena, developed by ex-Google Deepmind researchers. Now publicly available, Udio offers an intuitive platform for creating music through text prompts, promising ease of use, voice quality, and musicality.
Check it out yourself on Udio’s website.
Bridging Worlds: An Introduction to Grok-1.5V's Multimodal Capabilities
In xAI’s third consecutive feature in this newsletter, Grok-1.5V introduces new multimodal models capable of understanding both textual and visual data. It demonstrates competitive performance across various benchmarks, notably in real-world spatial understanding. Grok-1.5V aims to improve how digital systems interpret the physical world, offering early access to testers and existing users.
Read more on xAIs blog.
The Silicon Shift: How Tech Giants Are Crafting the Future of AI Hardware
In a significant move towards custom silicon, Google, Intel, and Meta are each developing their own chips to enhance AI workloads, offering improved performance, energy efficiency, and system scalability. Google's Axion Processors, designed for data centres, promise up to 60% better energy efficiency and 50% improved performance over current x86 instances.
Intel's Gaudi 3 AI accelerator aims to democratise generative AI with a 4x increase in AI compute and enhanced memory and networking capabilities. Meanwhile, Meta's next-generation Training and Inference Accelerator doubles the compute and memory bandwidth of its predecessor, focusing on efficiency for AI-driven ranking and recommendation models. These advancements underscore a growing trend of tech companies investing in bespoke hardware to meet the escalating demands of AI and machine learning applications.
Technical Reads
Building an AI Coach to Help Tame My Monkey Mind – Eugene Yan
“This week, I share about how I built Tara, a simple AI coach that I can talk to. I was initially skeptical of voice as a modality—Siri and other voice assistants didn't work so well for me. But after building Tara, I'm fully convinced. The post includes a phone line to Tara. Enjoy!”
Data acquisition strategies for AI-first start-ups – Air Street Press
Data acquisition is a key step for AI creation. This article takes a deep look at an array of ideas to solve this problem.
Lessons after a half-billion GPT tokens – Ken Kantzer
“My startup Truss (gettruss.io) released a few LLM-heavy features in the last six months, and the narrative around LLMs that I read on Hacker News is now starting to diverge from my reality, so I thought I’d share some of the more “surprising” lessons after churning through just north of 500 million tokens, by my estimate.”
How Do Machines ‘Grok’ Data? – Anil Ananthaswamy
“For all their brilliance, artificial neural networks remain as inscrutable as ever. As these networks get bigger, their abilities explode, but deciphering their inner workings has always been near impossible. Researchers are constantly looking for any insights they can find into these models.”
Projects, Code and Discussions
“Explore optimized prompts for a breadth of business and personal tasks.”
“LLM training in simple, raw C/CUDA.”
“aider is AI pair programming in your terminal”
Learning
Visualizing Attention, a Transformer's Heart | Chapter 6, Deep Learning – 3Blue1Brown
“Demystifying attention, the key mechanism inside transformers and LLMs.”
CS 886: Recent Advances on Foundation Models – University of Waterloo
A full course on foundation models in AI from the University of Waterloo.
🚀 Don't miss your weekly dose of cutting-edge AI innovations with The AI Canvas newsletter!
Subscribe now to ensure you never miss out on these transformative insights.
Looking for more specialised consultancy? At ADSP we’re a team of data experts who build AI products with purpose.
We deliver data science projects for companies who want to harness the power that AI can bring to their organisation. Get in touch at hello@adsp.ai.
Stay tuned with The AI Canvas podcast for in-depth episodes exploring Generative AI's transformative role across various industries.