Is The Pursuit Of AI & Humanoid Robots Based On A Flawed Approach?

Support CleanTechnica’s work through a Substack subscription or on Stripe.

Artificial intelligence, or AI, is the new marketing buzzword. Suddenly, it’s everywhere. My inbox this morning contained a pitch from a company that claims AI can do a better job of composting my kitchen scraps. AI may be the key to one of my pet projects that I have dreamed about for 30 years — the electric fork. The letters “AI” seem to send folks into paroxysms of joy. Finally, something that allows us not to think or actually do stuff! What will they think of next?

There is an offshoot to AI mania — humanoid robots. Just think of it! Machines that look like us but without the flaws. No need for coffee breaks, sick days, or paid vacations. No convoluted retirement plans to set up. They will cheerfully go about their appointed rounds, asking nothing more than to have their batteries recharged once in a while — a task they will surely be able to accomplish themselves.

The possibilities are limitless, despite the fact that AI could conceivably consume every electron currently generated by thermal or renewable means. The future’s so bright, we gotta wear shades. Onward!

OpenAI & Humanoid Robots

Readers may recall that a few years ago, Sam Altman and Elon Musk were working together on AI and promising great things were coming. Then they discovered they couldn’t stand each other and had a rather public falling out. Musk announced that Tesla was working on Optimus, a humanoid robot controlled by AI. Now, according to Wired contributor Will Knight, OpenAI is also getting into the humanoid game.

Although the company has not trumpeted its intentions, it has begun hiring researchers who work on humanoid systems and exploring how to advance artificial intelligence. Sources told Wired that OpenAI is recruiting people to work specifically on humanoid robots. A number of recent hires suggest its robotic intentions are now accelerating.

Chengshu Li joined OpenAI in June 2025 from Stanford University, where he worked on a number of robotics projects, including learning how to measure the abilities of humanoid robots in a wide range of household chores. Two other researchers from another robotics lab have already joined the company, according to their LinkedIn profiles. A third robotics lab that focuses on humanoid development told Wired one of its students was also recently recruited.

OpenAI declined to comment on the Wired story, but it has posted a number of job listings related to robotics research on its website recently. One is seeking people with expertise in tele-operation and simulation. Tele-operation is a crucial part of training partial or fully humanoid robots. In it, a human operator performs chores and controls the limbs of the robot, while an algorithm learns how to mimic the actions. The role also calls for expertise in simulation tools including Nvidia Isaac, which is widely used to train humanoids in a virtual physical environment.

Wired said it remains unclear whether OpenAI intends to build its own robots, use off the shelf hardware, or partner with a robotics company. One possible clue is another job posting for a mechanical engineer with expertise in prototyping and building robot systems with sensors for touch and motion. The job also calls for “experience designing mechanical systems intended for high volume (1M+), problem-solving on assembly lines,” which suggests systems that would be mass-produced or that might even be deployed in manufacturing.

OpenAI’s robot job postings all say its robotics team “is focused on unlocking general-purpose robotics and pushing towards AGI-level intelligence in dynamic, real-world settings.” Knight wrote that the focus on robots at OpenAI suggests the company, “believes reaching artificial general intelligence — AI that exceeds human intelligence — may require developing algorithms that are capable of interacting with the physical world.”

Stefanie Tellex, a roboticist at Brown University, told Knight that building more effective robots will involve designing and training AI models capable of “processing high frame rate, high dimensional perceptual input, and producing high frame rate, high dimensional physical outputs.” Tellex is not affiliated with OpenAI.

Humanoid development is an already crowded field, one that includes Google and Tesla along with several lesser known startups. “I don’t see (OpenAI) having any magical advantage over anyone else,” Tellex said.

Premature Celebrations

Knight wrote,

“While humanoids can perform impressive feats like dancing, they still lack the intelligence required to operate in complex and unpredictable, or ‘unstructured’, environments. To acquire this, they will need algorithms that go beyond a large language model’s understanding of the physical world. These systems must be able to control limbs and grippers in order to walk and manipulate physical items. Some research groups are starting to demonstrate progress in developing more generally capable AI models for robots.

“The world of venture capital is already gaga over the potential of humanoid robots. According to Morgan Stanley, the humanoid industry could be worth as much as $5 trillion by 2050, which raises the question, “What will people do when people have nothing to do? Watch reruns of Wall-E?”

Today, politics infects every conversation. Amidst a full frontal assault on universities and science, the current US president is pounding his chest and proclaiming, “America is the country that started the AI race. I’m here today to declare that America is going to win it.” But in order to do so, the US will need to attract and retain top researchers in the field, which means welcoming talented people from other nations. What happened in Georgia last week is hardly calculated to make anyone from another country feel welcome in America today.

A Pioneer Departs The US And Returns To China

One person who has packed his bags and left is Song-Chun Zhu, who came to the US in 1992 and became one of the more prominent leaders in AI research. Then in 2020, he left the US to return to China, where he was celebrated as a hero. According to a report by The Guardian, Zhu became a member of China’s top political advisory body, where he proposed that China should treat AI with the same strategic urgency as a nuclear weapons program.

Zhu looks at AI very differently than the American companies such as OpenAI, Meta, and Anthropic that have invested billions based on the idea that, if equipped with enough data and computing power, models built from neural networks could provide humanity with the dream of artificial general intelligence. The Guardian describes AGI as “a system that can perform not just narrow tasks, but any task, at a level comparable or superior to the smartest humans. Some people in tech also see AGI as a turning point, when machines become capable of runaway self-improvement. They believe large language models, powered by neural networks, may be just five to 10 years away.”

Zhu thinks they are all looking through the wrong end of the telescope. He believes a sign of true intelligence is the ability to reason towards a goal with minimal inputs. He calls it a “small data, big task” approach, as opposed to the “big data, small task” approach that is the basis of large language models like ChatGPT.

In his mind, artificial general intelligence is characterized by resourcefulness in novel situations, social and physical intuition, and an understanding of cause and effect. Large language models, Zhu believes, will never achieve this. Some AI experts in the US have similarly questioned the prevailing orthodoxy in Silicon Valley, and their views have grown louder this year as AI progress has slowed and new releases, like GPT-5, have disappointed. A different path is needed, and that is what Zhu is working on in Beijing, The Guardian says.

Priorities & Possibilities

The failed US administration — which is busy slashing funds for children’s’ health — has pledged $90 billion to promote AI development in the US. That money will be cheerfully gobbled up by some of the wealthiest humans and corporations in history. Why they would need such government largess is a question with no answer, because no one is asking it (except for a group of folks at CleanTechnica). If Zhu is correct, those funds will be squandered as AI in the US disappears down a rabbit hole.

The dream of humanoid robots has captured everyone’s imagination and attracted virtually unlimited piles of money. What no one has explained quite yet is who will pay to purchase these devices and what will they do with them once they own them? It seems to us at CleanTechnica that these are threshold questions that should be answered before the money is spent, but what do we know? We are just writers — an occupation that may soon disappear as computers and robots continue getting more proficient at performing tasks that were once the exclusive province of actual people.