Recently, an AI system constructed using living human brain cells has made its way to the forefront of brain-like research in a Nature subsidiary journal. This system has seen its speech recognition accuracy soar from 30%-40% to 78%. Capable of unsupervised learning, similar to neural networks, it relies on the connections of neurons within brain-like organs to facilitate learning. After just two days of training, the system has shown significant improvement in identifying speakers, although it still cannot comprehend the content of the speech. However, the lifespan of the system's brain-like organs is limited, and the external devices required to maintain its operation consume a significant amount of power. The research suggests that the development of a truly universal biological computing system may take several decades, but it provides important insights into the mysteries of human brain learning.