New users can click on "International Business Daily" to pay attention

Recently, a video released by artificial intelligence robot company Figure AI caused quite a stir on social media. In the video, Figure 01, a humanoid robot developed by Figure AI, performs some simple actions based on human voice instructions, such as handing apples to humans, putting plastic bags into a basket, and picking up a package with its five fingers. Box, then walk a few more steps, place the box on the conveyor belt, and so on.

Surprisingly, in the video, Figure 01 also demonstrates the ability to communicate with humans while executing instructions, and the dialogue is very smooth. More importantly, all behaviors of the robot in the video are not remotely controlled. Figure 01 Why can I perform tasks "autonomously"? Behind this, it has something to do with OpenAI's generative artificial intelligence model.

Figure AI was founded in 2022. At the end of February this year, Figure AI announced a partnership with OpenAI in an attempt to push the boundaries of machine learning.

After the video was released, Figure AI founder Brett Adcock explained how Figure 01 works on social media. It is reported that Figure 01 provides robot action support by Figure AI's neural network, and OpenAI provides visual reasoning and language understanding support. Figure01 inputs the images taken by the robot camera and the voice transcribed text captured by the microphone into the large visual language model provided by OpenAI. This model can understand both images and text at the same time. Then the neural network of Figure01 outputs actions and feeds them back through text to speech.

In short, Figure01 has a "central brain" that you can talk to and let it do things. For example, when you issue an instruction to Figure01 that "you need to fold the washed clothes", the "central brain" will formulate tasks and establish paths to realize operations such as carrying clothes baskets and folding clothes.

With the help of OpenAI's big model, Figure01 can currently do-describe its visual picture, plan future actions, recall its memories, and verbally explain its reasons to describe why it performed a specific action. For example, the act of handing an apple to a human in the video is triggered by a human saying "I'm hungry". The answer given in Figure 01 is-"This is the only edible item I can give you from the table."

Looking ahead, Figure AI plans for its humanoid robots to be utilized in manufacturing, shipping and logistics, warehousing and retail.