However, the technology is not flawless and can make errors. While it shows promise in environments like warehouses where mistakes are tolerable, deploying it in more hazardous settings like manufacturing plants could pose greater challenges and risks. AI expert Gary Marcus warns of the potential harm if the technology is used inappropriately, particularly in situations involving large robots that could cause significant damage.
Key takeaways:
- Former OpenAI researchers have developed a new software at a startup called Covariant, which helps robots understand and interact with the physical world, and also develop a deeper understanding of language.
- The AI technology developed by Covariant allows robots to pick up, move, and sort items in warehouses, and also gives them a broad understanding of the English language.
- The company's technology, called R.F.M. (robotics foundational model), combines data from images, sensory input, and text, providing robots with a more comprehensive understanding of their environment.
- AI expert Gary Marcus warns about the potential harm of the technology if used in the wrong way, especially in hazardous settings like manufacturing plants where mistakes could pose greater challenges and risks.