Wandelbots on the future of robot programming

Google DeepMind has unveiled two cutting-edge artificial intelligence models, Gemini Robotics and Gemini Robotics-ER, marking a significant step in bringing advanced AI capabilities into the physical world. Building upon the foundation of Gemini 2.0, these models extend DeepMind’s multimodal reasoning prowess, which already encompasses text, images, audio, and video, to include direct interaction with the environment.

Gemini Robotics, a sophisticated vision-language-action (VLA) model, introduces physical actions as a new output, effectively enabling robots to be controlled directly. Complementing this, Gemini Robotics-ER offers enhanced spatial understanding, empowering roboticists to leverage Gemini’s embodied reasoning (ER) abilities for customized programs.

These innovations collectively expand the range of real-world tasks that robots can perform. Demonstrating their commitment to practical application, DeepMind is collaborating with Apptronik to integrate Gemini 2.0 into humanoid robots.

Facing significant financial challenges and “substantial doubt” about its operational continuity, iRobot has launched a formal strategic review, exploring options such as a potential sale or debt refinancing. This decision follows the collapse of its planned acquisition by Amazon, a deal that was ultimately derailed by antitrust concerns, despite iRobot’s efforts to recover its footing. Source: Fortune Magazine

The company’s financial performance has deteriorated, with a 23.4% revenue decline in fiscal 2024, dropping from $890.6 million to $681.8 million, and an operating loss of $103 million. In response, iRobot has implemented aggressive cost-cutting measures, including reducing its workforce by over 50% and downsizing its office space. However, these efforts have been insufficient to offset the impact of decreased revenue, promotional spending, order timing issues, and ongoing competitive pressures, all of which the company hopes to address with its new product lineup.