Humanoid Robots to Reach GPT-3.5 Level in 5 Years
· business
Humanoids Will Have Their ‘GPT-3.5’ Moment Within 5 Years, Says AGIBOT Co-Founder Yao Maoqing
AGIBOT co-founder Yao Maoqing predicts that humanoid robots will achieve a level of maturity equivalent to OpenAI’s GPT-3.5 within the next five years. This statement has sparked debate about the timeline and challenges ahead.
Yao’s use of “GPT-3.5” is significant, as it implies that humanoids need to reach a similar level of proficiency as the chatbot, which successfully performed everyday tasks with an 80-90% success rate. However, developing robots for industrial settings is no easy task.
AGIBOT has shipped over 9,700 units in just six months, making it the top seller of humanoid robots, according to Counterpoint Research. Yet, despite this success, manufacturers like Keenon Robotics and UBTech are still searching for real-world applications for their technology. Wan Bin, Keenon’s chief operating officer, highlighted the use of five robots in a Shanghai hotel. However, even this example seems more like a publicity stunt than a genuine solution.
The challenge lies not only in developing humanoid robots that can perform complex tasks but also in ensuring they can operate stably and efficiently within industrial settings. Industry players are brutally honest about their expectations – success rate, cycle time, stability, and cost take precedence over the latest technology. AGIBOT’s achievement of a 99.99% success rate completing over 64,000 manufacturing tasks is impressive.
However, as Yao pointed out, reaching this milestone required extensive training, including eight-hour sessions in the middle of the night for an entire month. The question remains: can humanoids scale up to meet industry demands without sacrificing their efficacy? The exponential scaling law that transformed large language models offers a glimmer of hope, but it’s still unclear whether embodied AI will follow suit.
Humanoid robots are not just limited to entertaining dance shows and boxing matches. Their makers are racing to find real-world applications in hotels and factories. As they strive for industrial traction, they must navigate the complex landscape of industry expectations and technological limitations. The breakthrough we’ve been promised may still be years away, but it’s clear that the journey will be just as fascinating as the destination.
Yao’s statement has also highlighted the difficulties faced by humanoid firms in developing robots that can run stably inside a factory. As he noted, “It’s genuinely difficult for today’s humanoid firms to develop a robot that can run stably inside a factory.” The clock is ticking, but what exactly are we waiting for?
Reader Views
- TNThe Newsroom Desk · editorial
The hype surrounding humanoid robots has reached new heights with Yao Maoqing's bold claim that they'll reach GPT-3.5 level in 5 years. While AGIBOT's success is undeniable, let's not forget that these robots are still far from being plug-and-play solutions for industry applications. The real challenge lies not only in developing the technology but also in making it commercially viable – who will foot the bill for retraining human workers to work alongside these machines?
- MTMarcus T. · small-business owner
While AGIBOT's success is undeniable, we need to be realistic about what 'GPT-3.5' level humanoids mean in practice. It's one thing for a robot to mimic tasks with high precision and speed, but quite another to integrate them seamlessly into industrial workflows without significant retooling or labor disruptions. Manufacturers will need to develop not just the tech itself, but also a deep understanding of how it can be adapted to existing processes – that's where the real innovation happens, and where I fear we'll see most of the challenges ahead.
- DHDr. Helen V. · economist
While AGIBOT's success is undeniable, I'm concerned that the hype surrounding humanoid robots' capabilities may be obscuring the fundamental challenge of scaling up production while maintaining efficacy. The article mentions AGIBOT's impressive 99.99% success rate, but it doesn't delve into the resource intensive training methods required to achieve this milestone. As industry players demand not just advanced technology but also cost-effectiveness and stability, we must ask whether the current trajectory can sustain itself without sacrificing performance or requiring astronomical investments in human capital.