In a recent discussion on the development of robotics, it was noted that from a technological maturity perspective, robotics remains in a very early stage, comparable to pre-GPT-3.5 levels, perhaps even around GPT-2. Declaring an "intelligence元年" is premature at this point. However, considering application scenarios and market demand, 2026 represents a promising starting point for robotics entrepreneurship, arguably an opportune moment.
Data from January to April 2026 shows recruitment for embodied intelligence-related positions increased fifteenfold compared to the same period last year, with average monthly salaries approaching 62,000 yuan.
The current AI FOMO (fear of missing out) is driving tech companies to accelerate their strategic deployments. Embodied intelligence is rapidly emerging as the next heated talent battleground, following large language models and AI agents.
Regarding competition for talent between model developers, AI giants, and embodied intelligence companies, this trend relates to the differing maturity levels of these fields. Large model companies invest billions, even tens of billions, in computing power and hardware. For top researchers, achieving a 5% to 10% reduction in computational costs could translate to savings worth hundreds of millions or even billions of dollars.
Addressing popular events like robots appearing on Spring Festival galas or robot marathons, the focus for robotics startups should not be on performance or demonstration scenarios, but on identifying and persistently developing consumer application scenarios. Unclear scenario definitions lead to product designs that endlessly chase human-like capabilities, pushing technical difficulty to extremes. With sufficiently clear scenarios, companies can streamline designs based on real needs, finding a balance between technically feasible solutions, mass-production costs, and commercial value.
Discussing the impact of robotics on work patterns, the replacement of some programming tasks by Coding Agents does not necessarily imply large-scale unemployment for programmers. Instead, it may trigger a significant redistribution of talent. Historically, only high-margin industries like internet, new energy, and fintech could afford large programmer teams. As AI agents enhance programmer efficiency, enabling one programmer to handle work previously done by a small team, traditional sectors such as foreign trade and agriculture may also gain access to programming talent. This shift could drive a flow of programmers from large tech firms to medium-sized and small companies, and even into various traditional industries.