Robotics startup plans to have customers train their own AI assistants
Reimagine Robotics, founded by former DeepMind engineers, intends to create practical robots by having users teach them through interaction. CEO Jonathan Scholz says this approach will make robots more useful in real-world settings. The company aims to leverage customer feedback to improve AI learning.
Reimagine Robotics, a startup founded by former DeepMind engineers, is taking a distinctive approach to artificial intelligence by shifting the training burden from developers to end users. Rather than pre-programming robots for narrow tasks, the company’s CEO, Jonathan Scholz, emphasizes that customers will teach assistants through direct interaction. This method aims to produce machines that adapt to unpredictable, real-world environments—a persistent challenge in robotics, where scripted behaviors often fail outside controlled settings. By relying on user feedback as a core learning signal, the firm hopes to close the gap between laboratory demonstrations and practical, everyday utility. The strategy reflects a broader industry trend toward more personalized, iterative AI systems.
If successful, this customer-driven training model could democratize robotics, allowing small businesses and individuals to shape assistants to their specific needs without coding expertise. However, it may also place a significant burden on users, who must invest time and patience in teaching. The approach could accelerate adoption in niche sectors, but uneven training quality might lead to inconsistent performance, potentially affecting trust. Workers in repetitive roles may see shifts in task allocation, while entrepreneurs could gain affordable, adaptable labor. The long-term impact hinges on whether users embrace the teaching role.