Why I Created the First Humanoid Robots Comparator App?
If you were asked to imagine the future between now and the next few years, you would most likely think of an array of humanoid robots ready to help us where needed, kind of like straight out of some Asimov book. But what if this is not the future but is already the present?
Backstory
A couple of weeks ago, I attended one of the largest Italian innovation fairs. Between one speech and another on AI, which obviously remains the hottest topic of the moment, I was lucky enough to attend an interesting presentation of ErgoCub: a humanoid robot developed by IIT (Italian Institute of Technology).
I confess that I was almost completely foreign to the world of robots except for a few startups with which I had the opportunity to interact as an Analyst in the Italian Venture Capital field.
I had certainly heard of the most famous robots such as Sophia or the Tesla Optimus, but my knowledge ran out quickly and as a good technology enthusiast, even if not exactly an “expert” or “insider”, I felt rather at fault for having such a big gap on this topic.
In short, I wanted to know more and delve into the rabbit hole of human-like automatons, but I never thought it would be so difficult to dig. First I tried to do an internet search and I immediately noticed that the information is sparse, often incomplete and outdated.
Yet that presentation had opened my eyes and inside me I had the idea that robots would soon be increasingly present and we would never know enough about it.
I was intent on doing something about it even though I had a feeling it might not be the right time. After all, today all the lights are on AI and if there is no material people are probably simply not interested…
So as a Analyst I brought up the most insidious question in my repertoire: “Why now?”
And the more I found answers, the more I became convinced that we could really be in an interesting time for robotics.
Why Now?
Let’s start with some data to define the macro context.
According to the data collected, 2024 is the year in which the most humanoid robots ever came out, and there are still several months left until its conclusion!
From the point of view of global influences, it can be seen that America has dominance over this sector, but obviously Asia follows very closely and there is no shortage of a lot of European nations, both large and small.
China in particular has recently been focusing heavily on this sector, producing technologically cutting-edge humanoid robots with wide flexibility of use and with extremely accessible prices and available for sale even outside the same country.
Probably this roundup of information would already be enough to convince oneself of the “why now”.
But obviously this, in addition to being a challenge between nations, is a response to an important technological change that has convinced many to concentrate their efforts on bringing an interesting product to the market or pushing research one step further.
Convergence of technologies
The first technology I am referring to is obviously AI. In particular, the developments related to LLMs (Large Language Models) in the last at least 2 years have made it possible to obtain models capable of reproducing something resembling articulate reasoning or receiving and analyzing visual as well as textual input.
What is certain is that today language models have become so popular that AI and LLM are used synonymously. But if someone had asked you to imagine AI 4–5 years before the arrival of ChatGPT, do you think there is a possibility that you would have thought of a sentient robot?
Software travels faster than hardware and this GPT software, due to its incredible flexibility, has had the ability to truly reach everyone.
But at the end of the day, isn’t the most intuitive application for these linguistic models to give them a physicality and treat them as a single component of what is a real AI that can read and intervene when necessary in the external world?
In addition to linguistic models, enormous progress has been made in self-learning and techniques such as Deep Reinforcement Learning or Simulative approaches in general have played a fundamental role in this technical shift.
Previously, robots were guided with code in every single movement, with the result that a very small change to the environment in which it operated could lead to a misalignment between the pre-established movements of the machine.
Today robots are not hard-coded for every single movement, but on the contrary, they learn by repetition and emulation and are capable of “seeing” (even with sensors beyond the capabilities of the human eye like LiDARs for example) to all intents and purposes and practically correcting their actions in real-time during their execution.
But the revolution is not just for so-called “autonomous” robots. Even the “partially autonomous” ones have had some very interesting developments due to another technology that is making giant progress. I’m referring to VR.
Previously, robots were radio-controlled with something closer to a joystick for video games. Today they can be controlled remotely with unprecedented precision and effectiveness. And it is not at all obvious that this is the case!
On the other hand, it naturally arises to ask ourselves why we continue to work on this type of robot instead of focusing on improving the characteristics of autonomous ones. Well, the truth is that they are not alternatives at all, but it depends a lot on what vision of the future you support: do you imagine a world in which we humans will move into the metaverse and all physical interactions will be mediated by our robotic alter ego or do you imagine having a little cybernetic person at your service and ready to help you with anything?
Adoption Problems?
Dystopias aside, even if the technology seems to have reached a promising level of maturity, thinking that the adoption of consumer electronics is just around the corner could however be a too hasty and simplistic consideration. In fact, the biggest barrier still remains that of social acceptance.
The idea I have on this is that probably and even cynically it will be overridden by a discussion of monetary convenience and competitive trade-off, it will simply become economically accessible and too disadvantageous not to use it, exactly in the same way as those who did not want to buy the phones in the late 90s or connecting to Facebook in the 2000s or using AI today. There is certainly no shortage of examples in this sense in history.
Of course, all this unless someone legislates heavily on it, but perhaps we are going too far here!
Remaining closer to the current state, it is interesting to already observe the shift of these humanoid robots towards pure research or military fields to arrive at tasks closer to the common man such as Food, Healthcare or Industry.
Probably the first real impact that these machines will have on society will be the replacement of the more or less specialized worker, something that Tesla and Nio are doing right now (could it be a coincidence that they both produce EV electric cars?🤔)
Manifesto
With a cool head and taking a look at the data I have collected, I am probably aware that we are still a few years ahead of the real wave, but ultimately analysts are not really able to predict the future with certainty and sometimes you just have to do what you are passionate about without overthinking. In the worst case, you will have wasted time but you will certainly have learned something new.
So I created Bots Like You (www.botslikeyou.com) with the aim of sharing knowledge on humanoid robotics. The original idea was to build a humanoid robot comparator. Today there are comparators for practically everything and it seemed like a good way to make the exploration of this data interactive and useful, both to explore the technical specifications and to become more aware of the future that awaits us (even in a funny and bizarre way).
While I was exploring the different possibilities I had two other ideas to make the most of this data which are under development and which you can sign up to the waiting list to receive notification once they are ready.
Furthermore, since it is a type of data that is very difficult to recover and there are very few online sources that do disclosing of this type, I thought of setting up a distributed data-feeding and review system where anyone can report new releases or correct features /fields that remain outdated over time directly from the user interface with a few simple clicks.
And on the basis of this sharing, I have decided to make the dataset and its future updates completely open for anyone who wants to enjoy analyzing this data for the purposes that interest them most, hoping to do something pleasant and useful.
So let’s share, maintain and grow this knowledge together to closely follow this new cyber-revolution 🙌!
PS: The design and construction process of this project was carried out in 2 weeks and documented in public step by step on X.
PPS: Part of the site and the graphics were developed in collaboration with Settepercento.com, the data storytelling and visual essays site where you can find other interesting in-depth articles on both tech and non-tech topics but with a data-driven approach.
PPPS: The product is live on Product Hunt. I’ll leave the link to the launch for anyone who wants to take a look at it and maybe leave feedback.
