Scientists Discuss the Future of Biological Computing

Share

Summary

In this StarTalk special edition, Neil deGrasse Tyson, with co-hosts Gary O'Reilly and Chuck Nice, interviews neuroscientist Brett Kagan, Chief Scientific Officer at Cortical Labs. They discuss the groundbreaking work of Cortical Labs in creating synthetic biological intelligence (SBI) by growing neurons in a dish and integrating them into electronic circuits. The conversation delves into the ethics, capabilities, and future implications of SBI, comparing it to traditional silicon-based intelligence.

Highlights

Introduction to Synthetic Biological Intelligence (SBI)
00:00:20

Neil deGrasse Tyson, Gary O'Reilly, and Chuck Nice introduce Brett Kagan from Cortical Labs, who is developing Synthetic Biological Intelligence (SBI). Instead of integrating technology into biology (e.g., neuralink), Kagan's work focuses on introducing biology to technology by growing brain cells in a dish and integrating them into devices to create intelligent processors.

Growing Neurons and Ethical Considerations
00:03:16

Brett Kagan explains that Cortical Labs grows neurons sustainably and ethically by using induced pluripotent stem cells from adult donors, which are then transformed into brain cells. These cells are integrated into multi-electrode arrays (CL1 devices) that record and supply electrical pulses for communication. The ethical approach avoids animal harvesting, a common practice in the field.

Embodied Networks and Learning in a Dish
00:09:24

Kagan defines 'embodied networks' as creating a closed loop where neuronal activity interacts with a virtual world (like the game Pong) and receives feedback. This creates a barrier between the neurons' activity and its effect on the world, allowing them to learn and reorganize rapidly. He highlights that neurons spontaneously communicate and intentionally organize themselves in response to information.

SBI vs. Silicon Intelligence: Efficiency and Learning
00:10:31

The discussion contrasts embedding neurons in circuits with traditional silicon computing. Kagan argues that biological intelligence offers advantages in power consumption and data efficiency, learning quickly from few samples (e.g., recognizing a ball with minimal exposure). He explains that SBI's ability to learn rapidly, as demonstrated by playing Pong within minutes, surpasses current machine learning algorithms in sample efficiency.

The 'Biological Computer' and AGI
00:15:36

Chuck Nice describes the work as creating a 'biological computer' capable of real-world grounded knowledge, unlike silicon computers that require extensive data for every association. Kagan emphasizes that while Artificial General Intelligence (AGI) in silicon is uncertain, the existence of generalized intelligence in biological systems (like cats, rats, and bees) provides a proven ground truth for SBI development.

Ethical Concerns and Consciousness in SBI
00:17:47

Concerns about the ethical implications, particularly the fear of creating a new species or uncontrolled superintelligence, are addressed. Kagan reassures that biological intelligence in a dish would be contained and controllable, lacking the self-replication or internet access fears associated with AGI. The conversation moves to the concept of consciousness, clarifying that intelligence and consciousness are not inherently linked. Kagan collaborates with bioethicists to navigate these complex questions.

Future Applications and Hybrid Computing
00:33:55

Kagan envisions a future with 'heterogeneous compute,' combining biological processing units (BPUs) with quantum and silicon processors, each optimized for different tasks. SBI can solve problems differently from silicon intelligence, prioritizing 'good enough' solutions over optimal ones, which can be more efficient in certain contexts. The platform technology has potential applications in basic science research, healthcare (drug testing), and understanding the human brain.

Insights into Brain Function and Disease Treatment
00:41:08

The technology could help understand the music of neurons and their computational approaches, leading to insights into our own brains. Kagan highlights its potential in drug testing for psychiatric and neurological diseases, as it provides a better understanding of how drugs affect information processing. He uses epilepsy models as an example, showing how treated neurons can improve their 'gameplay' and provide invaluable data.

The Long-Term Vision of Cortical Labs
00:45:51

Kagan discusses the high complexity of even a bee's brain, which, if harnessed, could outperform current machine learning drones. He expresses his company's mission to create a lasting legacy and impact the world through this new form of computation, despite the financial challenges, emphasizing the drive to understand the unknown and optimize solutions for human benefit.

Recently Summarized Articles

Loading...