Famous robot player says that the humanoid robot bubble is destined to burst

Famous roboticist Rodney Brooks has awakening calls for investors to concentrate billions on humanoid robot startups. You’re wasting your money.

Brooks, who co-founded Irobot and spent decades at MIT, is particularly skeptical of companies such as Tesla and Tesla, who are trying to teach robot dexterity by showing human videos. In his new essay, he calls this approach “pure fantasy thinking.”

problem? The human hands are incredibly refined, packed with around 17,000 special touch receptors, and the robots aren’t matched. While machine learning has changed the transformed speech recognition and image processing, these breakthroughs were built on decades of existing technology to capture the right data. “There is no such tradition for touch data,” Brooks points out.

Then there’s safety. Full-size walking humanoid robots allow huge amounts of energy to stand upright. When they fall, they are dangerous. Physics means that a robot twice the size of today’s models packs eight times the harmful energy.

Brooks predicts that in 15 years, “humanoid” robots will actually have wheels, multiple arms, special sensors and abandon their human form. Meanwhile, he is thoroughly convinced that today’s billions are funding expensive training experiments that will not expand to mass production.

It’s not the first time that we’ve poured cold water into expectations set by Brash entrepreneurs and enthusiastic investors. Last year, he spoke at TechCrunch at length about why generative AI promises can go beyond its capabilities and sometimes create more work.

For example, AI Research Nonprofit Metr said it had recruited 16 highly rated developers from a large open source repository this summer to measure the impact of AI tools on real software development. I then assigned nearly 250 real issues to address without using tools and measured the screen. When developers use AI tools it took 19% longer to complete the task. Interestingly, they realised that the AI ​​had raised them by 20%.

Brooks has also long argued that AI is not an existential threat that many people, including Elon Musk, assume. TechCrunch spoke to Brooks about this at MIT in 2017, and although the scenery looked very different, it didn’t look like the arena today.

At the time, Brooks said that there were only a few companies specialized in creating datasets for machine learning. Relatedly, he argued why it was not necessarily a natural conclusion that large tech companies would win in robotics. But today’s major robotics companies are not escaping the gravity of those companies.

Apptronik, a humanoid robot maker, which raised nearly $450 million from investors, counts Google among its supporters and partners with Google’s Deepmind Robotics team late last year to “edit cutting edge hardware and embodied intelligence.”

The diagram of another well-known AI robotics company developing humanoid robots, partly supported by Microsoft and the Openai Startup Fund, partnered with Openai in February 2024 to combine Openai research with its own “deep understanding of robot hardware and software.” The two split almost a year later in March, and Figuai said he enjoyed a “big breakthrough” within the robotics company, moving forward independently.