Today’s top AI Highlights:

1. Hugging Face’s tutorial to build your own AI robot for under $500

2. Create powerful RAG apps with Intel’s openource RAG Foundry

3. Chinese robotic startup unveils 5 new humanoid robots

4. AMD shows AI chips business is not just about AI chips

5. Opensource tool for real-time face swapping and deepfakes

& so much more!

Read time: 3 mins

Latest Developments

Get Hands-On with AI and Robotics 🤗

Build and train your own AI-powered robot! Hugging Face has released an in-depth tutorial on building and training your own robotic arm using neural network. The tutorial provides a comprehensive guide, from assembling the arm (costing approximately $400-$500) to training its AI brain (requiring a few hours on a standard laptop) that allows it to perform tasks autonomously. It uses techniques similar to how LLMs are trained for text, but adapted for robotics.

Key Highlights:

1. End-to-End Learning - The tutorial focuses on training a neural network to directly predict motor rotations from camera images, eliminating the need for complex intermediate steps. Training typically takes 2-3 hours on a standard laptop.

2. Opensource - The robot arm is based on an opensource design, and the 3D files are readily available. The software, including the training scripts and visualization tools, is also opensource. The estimated cost for components is around $200 for the leader arm and $280 for the follower arm.

3. Dataset Recording and Visualization - Creating a high-quality dataset is the most crucial part of training the AI and this project provides tools for recording and visualizing robot trajectories so you can analyze and improve the robot's performance. Recording a dataset can take approximately 30-60 minutes per episode, depending on the complexity of the task.

4. Control and Teleoperation - The tutorial covers both autonomous control using the trained neural network and teleoperation, where you can directly control the robot arm. This allows for initial data collection and testing before deploying the learned policy.

This is just the beginning. Explore the tutorial, experiment, share your creations, and contribute to this growing field of AI robotics.

The Ultimate “Foundry” for Powerful RAG Apps 💡

Tired of the complexities of building and evaluating Retrieval-Augmented Generation (RAG) systems? Intel has introduced RAG Foundry, a new opensource framework that provides a single workflow for data creation, training, inference, and evaluation of RAG techniques. Its modular design and customizable features make it an ideal choice for both researchers and practitioners seeking to build robust and adaptable RAG solutions.

Key Highlights:

1. Streamlined Development - RAG Foundry consolidates the entire RAG development lifecycle into a single, unified workflow, eliminating the need for disparate tools. This helps in rapid prototyping and experimentation so you can explore different RAG configurations and optimize your models for specific tasks.

2. Modular and Customizable - The framework is highly modular giving you granular control over each stage of the RAG pipeline, from data selection and retrieval strategies to prompt engineering and model fine-tuning.

3. Comprehensive Evaluation - RAG Foundry provides a sophisticated evaluation suite to assess both the retrieval accuracy and generative quality of RAG models. By analyzing these critical aspects, you can get deeper insights into your model's performance and identify areas for improvement.

Quick Bites

1. Chinese robotics startup Agibot has unveiled five new humanoid robots, opensource full-stack robot development, and a lot more! This includes both wheeled and biped humanoid robots designed for applications from household to industrial operations. Competing with Tesla’s Optimus robots, Agibot aims to ship 300 units by the end of the year.

2. Nvidia’s competitor AMD is acquiring ZT Systems, which provides compute design and infrastructure for AI, cloud and general purpose computing, for $4.9 billion. The acquisition will strengthen AMD’s AI ecosystem and position it to deliver AI infrastructure at scale.

3. Luma Labs has released its latest text-to-video model Dream Machine 1.5. It now features higher-quality text-to-video, smarter understanding of prompts, custom text rendering, and improved image-to-video.

Tools of the Trade

1. Deep Live Cam: Opensource tool for real-time face swapping and one-click video deepfakes. It can replace faces in videos or images using a single photo, ideal for video production, animation, and other creative projects.

2. Benchify: A GitHub code review tool to test code with advanced mathematics with better coverage than traditional unit tests. When errors are detected, it provides counterexamples and, when possible, a patch.

3. Profundo: AI-powered research tool that automates data collection, and analysis. It’s designed to be user-friendly and suitable for personal, academic, and industry research needs.

4. Awesome LLM Apps: Build awesome LLM apps using RAG to interact with data sources like GitHub, Gmail, PDFs, and YouTube videos through simple text. These apps will let you retrieve information, engage in chat, and extract insights directly from content on these platforms.

Hot Takes

1. Literally for years, Yann LeCun ridiculed my critiques of deep learning and LLMs, but eventually he figured out for himself that LLMs were not the way to AGI.
   So many tech bros and grifters still haven’t figured it out. ~
   Gary Marcus
   

2. My AGI development timeline: 
   > November/Dec 2024 (after the election):
   OpenAI announced GPT-5 with agentic reasoning capabilities.
   
   > January 2025:
   DeepMind announces breakthrough in generalized reasoning with Alpha techniques.
   
   > March 2025:
   Anthropic introduces new model alignment techniques for customization.
   
   > 2026-27:
   The first AI model widely considered AGI is showcased. ~
   Haider.

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 @Saboo_Shubham for lightning-fast AI updates and never miss what’s trending!

Unwind AI - Twitter | LinkedIn | Instagram | Facebook

Awesome LLM Apps | Sponsor Us

PS: We curate this AI newsletter every day for FREE, your support is what keeps us going. If you find value in what you read, share it with at least one (or 20) of your friends!