ARM Gets in the AI Chips Race + Tools of the Trade

PLUS: New AI features in iOS, Unitree releases robot priced at $16K

Author

Bee Keeper
May 16, 2024

Today’s top AI Highlights:

  1. UK-based ARM is planning to launch its AI chip by 2025

  2. Unitree Robotics launches G1, a smaller humanoid robot & a smaller price tag

  3. Eye Tracking and Music Haptics - Apple to roll out new AI features in iOS

  4. Google releases Firebase Genkit to build production-ready AI-powered apps

  5. Build awesome LLM apps using RAG using OpenAI, Anthropic, Google, and opensource models

& so much more!

Read time: 3 mins

Latest Developments 🌍

The race to meet the growing demand for AI chips is heating up, and UK-based semiconductor company Arm is stepping up to launch its own AI chips by 2025. This initiative, under the umbrella of SoftBank Group, aims to leverage Arm’s dominant position in smartphone processor architecture to enter the AI chip market. With a projected increase in the AI chip market value and a clear need for more advanced processing capabilities, Arm could challenge Nvidia and impact the industry majorly.

Key Highlights:

  1. Timeline: Arm will establish a dedicated AI chip division to develop a prototype by spring 2025. Mass production, managed by contract manufacturers, is scheduled to begin in fall 2025.

  2. Market Opportunity: With the AI chip market projected to grow from $30 billion in 2023 to over $200 billion by 2032, Arm is targeting this expanding sector. Currently dominated by Nvidia, the market’s demand outstrips supply, offering Arm a strategic opening to establish its presence.

  3. Collaborations: Arm is negotiating with Taiwan Semiconductor Manufacturing Corp. and others to ensure they have the necessary production capacity for their upcoming AI chips, crucial for their entry into the AI chip market.

Unitree Robotics, the company that made headlines with its humanoid robot H1, has unveiled a new robot, the G1. This smaller, more affordable robot is explicitly designed for research and development for university labs and AI researchers. While G1’s diminutive stature (127 cm, roughly the height of an eight-year-old) means it’s not suited for replacing human workers, it boasts impressive agility and a range of features that could be crucial for advancing robotics research.

Key Highlights:

  1. Size matters, cost lessens: G1’s smaller size allows Unitree to use less material and smaller motors, lowering the price compared to the H1. With a starting price of $16,000, the G1 is a more accessible option for research labs.

  2. Move freely, move fast: G1 is incredibly agile, with a wide range of joint movement and a walking speed of 2,000 mm/s (4.5 MPH). This speed surpasses many AGVs on the market, demonstrating the G1’s potential for various applications.

  3. Dexterous hands, tactile feedback: G1 is equipped with force-controlled, three-fingered hands. These hands can be further upgraded with tactile feedback, enhancing the robot’s ability to interact with objects and perform delicate tasks with precision. This opens up possibilities for the G1 to be used in research areas like manipulation, grasping, and object recognition.

  4. Powered by NVIDIA: The G1 EDU version comes equipped with an Nvidia ORIN controller, providing the computational power needed for complex AI algorithm development.

  5. Target: Due to its small size, G1 won’t be able to reach a typical workbench, kitchen, or household environment. However, foundation models developed on G1 could likely be applied to the H1 version.

Using AI to enrich lives is the most transformative way to make this technology accessible and inclusive. While companies like Elon Musk’s Neuralink take a different path with devices that go inside your brain to help disabled people do things, even small AI features can improve people’s lives. Apple is releasing new accessibility features in the upcoming OS updates to enhance the experiences of people with diverse needs.

Key Highlights:

  1. Eye Tracking: Allows users with disabilities to control their iPad and iPhone using only their eyes. It uses the front camera and doesn’t require any additional hardware. It uses on-device machine learning to keep all data securely on the device.

  2. Music Haptics: Transforms the experience of music for deaf and hard-of-hearing users. iPhone can translate audio into physical sensations like taps, textures, and vibrations. Millions of songs in the Apple Music catalog are compatible, and its API will be available for developers to integrate into their apps.

  3. Vocal Shortcuts: Users can create custom phrases for Siri to understand and execute shortcuts. For example, you could teach Siri to call your mother when you say “Ring.” This is especially helpful for people who have difficulty speaking.

  4. CarPlay: Users will be able to control apps in CarPlay with just their voice. Additional features will allow deaf or hard-of-hearing people to turn on alerts to be notified of car horns and sirens.

  5. Vehicle Motion Cues: Helps reduce motion sickness by displaying animated dots on the screen that correspond to changes in vehicle movement. These cues help to reduce sickness caused by sensory conflict that happens when there is a difference between what a person sees and feels.

😍 Enjoying so far, share it with your friends!

Tools of the Trade ⚒️

  1. Firebase Genkit: An opensource framework to build, deploy, and monitor AI-powered apps. You can integrate features like content generation, semantic search, and decision-making into your apps. It gives access to multiple models, structured output, multimodal content support, function calling, and more.

  1. CodeConductor: AI coding assistant that generates production-ready code for your services in minutes. It helps you quickly transition from concept to deployment and optimize development workflows. Customize and extend the generated code to fit your unique business needs while maintaining control.

  2. Research Paper Summarizer GPT: It summarizes research papers, highlighting key statistics, insights, and facts. It provides clear readable summaries of each section including introduction, methodology, findings, implications, and conclusions. It also suggests actionable next steps based on the paper for practical application.

Hot Takes 🔥

  1. With Ilya and Karpathy both gone, OpenA's transition from being a non-profit company dedicated to AI research to a secretive profit-maximizing company fiercely competing with Google is finally complete. ~Bindu Reddy

  2. The real audience for OpenAI’s demo was Apple. ~Pedro Domingos

  3. Is the reason OpenAI didn't launch an AI-powered search engine is that they're working on something bigger? 🤔An ACTION engine instead of a search engine. ~dharmesh

Meme of the Day 🤡

That’s all for today! See you tomorrow with more such AI-filled content.

Real-time AI Updates 🚨

⚡️ Follow me on Twitter @googleclone for lightning-fast AI updates and never miss what’s trending!

PS: I curate this AI newsletter every day for FREE, your support is what keeps me going. If you find value in what you read, share it with your friends by clicking the share button below!