Subscribe to Prof Nye's Digital Lab to receive new posts directly to your inbox.
Nye Warburton
The robots are coming.
Just like their cousin, the quad-copter drones before them, the robot dog movement is in full swing, and it’s only a matter of time before we all have them.
Now is the time for creatives to start playing with robotics.
Contents:
The Everyday Roboticist
Drones to RoboDogs
RoboDogs to Androids
What can we do with a Robot?
Getting Started with Robotics
Remember when creating a website meant wrestling with HTML and FTP? You may not, but I do.
Back then, embedding a QuickTime video on your GeoCities page could attract thousands of visitors simply because so few people had the technical skills to implement it. Today, anyone can create a super slick website in minutes using tools like WordPress or Squarespace.
A similar democratization is now happening in robotics, but it’s happening at 50 times the speed.
Robotics magic that was once the exclusive domain of PhD researchers at Stanford, is becoming accessible to creators, designers, and weekend hackers like me. The transformation mirrors the evolution of other technologies: from specialist tools to creative platforms that anyone can use.
I first experienced this shift through quad-copter drones.
When I was in Los Angeles, I spent my days at a co-working space in Chinatown called "KleverDog Coworking." A few of us “entrepreneurial-artist types” starting flying drones for fun on Friday afternoons.
Starting with a $100 programmable drone, I learned about PID controllers (the mathematical magic that keeps drones stable), swapping plastic rotors and batteries, and the basics of flight control with a quad-copter.
And now, drones are everywhere. FPV drone racing is purely awesome – pilots becoming disembodied entities racing through the air at 90 mph, seeing through their drone's camera feed. Drone light shows paint the sky at major events, and aerial photography has become a mainstream creative medium. It’s hard to find a beach parking lot where a weekend enthusiast isn’t buzzing their quad-copter around.
We're seeing a similar revolution with robotic dogs.
Boston Dynamics' Spot, a yellow robot dog driven by industry-leading artificial intelligence, was a viral internet sensation. It and others like it have inspired a wave of more affordable alternatives. The Unitree GoPro2 is offering a programmable robot competitor at a fraction of the cost. Sony is offering a consumer robot dog called Aibo for around $3,000.
Introducing the Go2 Pro: the next-generation robotic companionship. With enhanced strength, heightened intelligence, and extended battery life, the Go2 Pro surpasses its predecessor in every aspect. Experience the thrill of commanding your robotic companion as it effortlessly performs cool tricks and faithfully follows you on your adventures, whether you're jogging through the park or exploring new terrain.
Jumping in myself, I recently bought a MechDog for $400! – less than the price of a gaming console. It's programmable using Scratch or Python, and while I'm a pretty “dog-crap coder,” AI assistants like Claude, or the IDE Windsurf with Codium, help me write what I need. It's amazingly cool, mind-blowing, and a sure fire vector of the future we are inventing.
The evolution from quad-copter drones to quadruped dogs represents a significant leap in complexity.
While drones master the challenge of stable flight through PID control, robotic dogs tackle the intricacies of walking using inverse kinematics – an “old” animation software trick of calculating joint angles to achieve natural movement. They learn different gaits (walking, trotting, galloping), maintain balance on uneven terrain, and interact with their environment in increasingly sophisticated ways.
The tools for developing these robots are becoming more accessible too. Using Gazebo, a virtual environment simulator from the Open Robotics foundation, I can download virtual models of robots like Boston Dynamics' Spot, and train them in virtual spaces.
This post is to show how to set up the Boston Dynamics Spot SDK.
Spot might be millions of dollars of MIT quality research, but the SDK libraries and training of spot is free and open with anyone with access to github and a little bit of Claude-powered python knowledge. (aka, me)
Dressed in a sparkling leather jacket, NVIDIA's CEO Jensen Huang recently announced their Cosmos platform, which aims to democratize physical AI development. Cosmos is designed to provide accessible digital tooling for developers to train and test a variety of robot models. It’s targeted at the industrial sector, providing all of the data and compute necessary for robot labor to understand the real world.
"The ChatGPT moment for general robotics is just around the corner."
Jensen Huang, Nvidia -- CES 2025.
NVIDIA Cosmos is a platform of state-of-the-art generative world foundation models, advanced tokenizers, guardrails, and an accelerated data processing and curation pipeline built to accelerate the development of physical AI.
This push from major tech companies is being matched by rapid evolution in the open-source community. Hugging Face, known as a hub for AI models, now hosts educational robotics projects. The Robot Operating System (ROS), a collection out-of-the-box frameworks that handles complex robotics underpinnings, allows developers to focus on building functional robotics rather than wrestling with low-level control systems.
This convergence of affordable hardware and accessible software tools marks the beginning of a personal robotics revolution, much like how affordable PCs and accessible programming tools sparked the personal computing revolution. The focus on quadruped robots isn't just about making cool pets – it's a crucial step toward humanoid robots.
As Huang points out, our world is built for “human-sized interaction.” Doorways, stairs, tools, and furniture are all designed for human bodies. Mastering four-legged movement provides the foundation for the even more complex challenge of bipedal (two-legged) robotics.
The creative opportunities are already emerging.
Digital artist Beeple recently created an installation featuring Spot robots with 3D-printed heads of himself, blending robotics with contemporary art. Chinese drone shows have evolved into spectacular aerial ballets, with full sparkling dragons that undulate across the sky. These examples merely hint at the vast potential for artists, designers, and creators to shape the future of robotics.
As LLM powered agents take the digital world tasks from us, the cheap hardware robots will certainly take the real world ones. Cleaning, cooking, delivering, monitoring our day to day lives is a near certainty. More than likely … building, constructing, surgery, and massage therapy?!?!
Research now reveals that having robots in the house helps with loneliness and depression for older adults. (Though I might ask if we are actually fixing the problem by throwing robots at it, or should we just double down on more human connections?)
Purpose: Differences in depression and loneliness, during the COVID-19 pandemic, for older adults with mild to moderate dementia living in residential care after interacting with a robotic companion dog or cat were explored. Experiences of family ...
Anyway - My sum up is this:
Don't let technical constraints hold you back.
The tools are becoming more accessible every day, and your creative vision is exactly what the field needs. Start small – experiment with a programmable drone or an affordable robot dog. Join online communities sharing open-source robotics projects. The barriers to entry have never been lower, and the potential for creative innovation has never been higher.
The future of robotics belongs to those who can imagine it, not just those who can engineer it. The question isn't whether robotics will become democratized, but what we'll create when it does.
And if my experience with a $400 robot dog is any indication, we're in for an exciting ride.
Keep in mind, I am a NooB myself, but so is most of the world!
Based on my experiments and research, here are some ways to get going building your own robot best friend.
Start Virtually
Before buying any hardware, start with simulation.
There are drone simulations all over the place, many are downloadable from Steam. If you are savvy with a game engine, maybe do some research on virtual robotics kits --- For Example, Cleverlike Studios has developed a virtual robot training course for the Unreal Engine.
Learn valuable robotics skills through engaging activites with virtual robots in Unreal Engine. In addition to learning valuable coding concepts you'll be developing your skills in a software tool used for Interactive 3D careers across many industries.
Quad-Copter Drones
Small programmable drones are perfect for learning the basics. They're relatively cheap (under $100 for starter models), and you can program them using Python.
You'll learn about PID controllers, pathfinding algorithms, and real-world physics – If you really get bitten by the bug, you can design and develop drone shows, go deeper into drones for survey work, or (gulp) the military.
I’ve been really impressed with the work of Drone Blocks. They provide educational kits for classrooms, and simple starter drone packs for learning.
Harness the power of DroneBlocks drone STEM curriculum, software, and drone kits to educate students of all ages.
Robot Dogs
I just purchased a MechDog ($400), but if you have the wallet and are ambitious, I might recommend the ROS Pug ($1300) as an entry points.
The ROSPug come with the new Jetson Nano boards – NVIDIA's powerful AI computers which should undoubtedly make an obvious difference. Imagine creating an interactive art installation where the robot responds to your movement, or designing a game where virtual and physical robots interact.
ROSPug Quadruped Bionic Robot Dog Powered by Jetson Nano ROS Open Source Python Programming
4. Tap into Open Source!
Platforms like Hugging Face aren't just for AI models anymore – they're about to become the hub for open robotics projects too. Browse through the available models, fork a project that interests you, and start experimenting.
LeRobot is a brand new community dedicated to sharing robots for education.
State-of-the-art Machine Learning for real-world robotics
If you liked this essay, please consider subscribing and sharing. Thanks for reading!
Nye Warburton is an educator and who likes to imagine a positive robot future. This essay with written with human passion in collaboration with Claude Sonnet 3.5. Visit https://nyewarburton.com