Why are we ("suddenly") getting better robots?
Robotics is a near ancient field of research. But how far have we gotten, and why isn't my personal robot assistant here yet?
The Hard Truth About Robotics
Here's a thought experiment:
Imagine trying to catch a baseball while wearing VR goggles that show you the world with a several hundred millisecond delay.
This challenge, which seems nearly impossible for humans, represents what robots have been dealing with since their inception. But here's the exciting part: we're finally on the cusp of solving this fundamental challenge.
For decades, we've dreamed of having robot helpers in our homes – folding our laundry, organizing our closets, and handling tedious household chores.
While these tasks have proven surprisingly complex, the convergence of advanced AI, improved sensors, and breakthrough control systems is about to change everything.
A really good TED talk by Ken Goldberg at UC Berkley explains why
The challenge lies in what is known as Moravec's paradox: tasks that humans find easy are often incredibly difficult for robots, and vice versa.
A robot can lift heavy objects with incredible precision, but has traditionally struggled with seemingly simple tasks like stacking blocks or folding clothes.
Understanding the Challenge
The complexity of robotics can be broken down into three key challenges, which, remarkably, are all on the verge of being solved:
1) The Control Layer
While combining sensor inputs once created overwhelming complexity, modern AI systems are mastering sensor fusion with unprecedented accuracy. What was once "more art than science" is becoming a precise, reliable process thanks to advanced machine learning models that can integrate thousands of data points per second with near-human intuition.
2) The Perception Layer
The historical catch-22 of robotics sensors being simultaneously too precise and not precise enough is being shattered by new multi-modal AI systems. Modern robots can now combine the millimeter-precise measurements of LiDAR with sophisticated visual understanding that rivals human perception. The "intern with perfect eyesight who's never seen the real world" is rapidly gaining real-world experience through advanced training techniques.
3) The Physics Mastery
The philosophical challenge of physical uncertainty – like predicting the exact path of a pushed bicycle – is being overcome through groundbreaking approaches to embodied AI.
Companies like 1X and Figure are teaching robots not just physics equations, but the intuitive understanding of physical interactions that humans possess. This marriage of precise mathematical models with learned experience is creating robots that can adapt to real-world variability just like humans do.
Convergence
What makes 2025 particularly exciting is the simultaneous maturation of multiple breakthrough technologies:
Advanced AI models specifically designed for robotics are bridging the gap between perception and action
New sensor fusion techniques are providing robots with human-like environmental awareness
Embodied AI approaches are giving robots the "common sense" they've historically lacked
Leading companies in the field are now combining these advances in revolutionary ways. Rather than pure programming, robots are learning through human demonstration, developing an intuitive understanding of the world that goes beyond rigid algorithms.
Strategic Implications
For investors and business leaders, this suggests a few key insights:
The next big breakthrough in robotics will likely come from AI, not mechanical engineering (hence why companies like NVIDIA is going hard into open source platforms for robotics models hosting)
The companies that solve the problems of Control+Perception+Predictable Physics will win, even if their robots are physically less capable
Specialized robots (like warehouse automation) will continue to steam ahead, as they buy-pass a lot of the ambiguities in the human world and rather operate in their own contained environment where the variables are much easier to control
The Bottom Line
We're not just facing a "ChatGPT moment" in robotics (as Jensen Huang recently put it)– we're experiencing something much more interesting.
The industry has moved beyond having all the pieces and is actively putting them together in working, practical systems.
The winners in this space won't just be those with the best hardware or most advanced AI, but those who are already combining these elements to create robots that can reliably understand and interact with our complex world.
I’m very excited for 2025, as it is poised to be the year when humanoid robots transition from impressive demos to practical tools, marking the beginning of an era where robots truly become part of our daily lives.
The challenges that once seemed insurmountable are being solved through the clever application of multiple technological breakthroughs, suggesting that the long-awaited robot revolution is finally at hand.
Now I just need to sign up on all the waitlists to buy some hardware… :)
About Me
Working at the interface between frontier technology and rapidly evolving business models, I work to develop the frameworks, tools, and mental models to keep up and get ahead in our Technological World.
Having trained as a robotics engineer but also worked on the business / finance side for over a decade, I seek to understand those few asymmetric developments that truly shape our world
You can also find me on X, LinkedIn or www.andreasproesch.com