Summary of Can Living Human Brain Cells Power AI? | Bloomberg Primer

The video discusses the groundbreaking research at Cortical Labs, where living human brain cells, referred to as "DishBrain," were trained to play the arcade game Pong. This development signifies a potential shift in artificial intelligence (AI) by integrating biological systems with computing technology. Here are the key scientific concepts, discoveries, and phenomena presented:

Key Concepts and Discoveries

DishBrain: A cluster of 800,000 neurons on a silicon chip that can play Pong, demonstrating the ability of living brain cells to interact with technology.
Biocomputing: The merging of biological systems (like neurons) with computational systems, potentially leading to more efficient AI models that require less data and energy.
Neural Efficiency: Biological brains operate on approximately 20 watts of power, compared to supercomputers that may require up to 40 megawatts, highlighting the energy efficiency of biological systems.
Free Energy Principle: A theory suggesting that living systems, including brain cells, strive to minimize surprise by predicting their environment, which may explain how DishBrain learned to play Pong.
Brain Organoids: Miniature brain structures created by connecting neurons, used for research in drug development and understanding neurodegenerative diseases.

Methodology

Neurons Learning to Play Pong:
- Neurons were grown on a silicon chip and connected to a computer.
- The chip was divided into sensory and motor sections to relay game information and control movements.
- Neurons received electrical stimulation as rewards for successfully hitting the ball and chaotic stimuli for misses, reinforcing learning.

Challenges and Considerations

Ethical Concerns: The potential for Brain Organoids to develop self-awareness raises ethical questions about their treatment and rights.
Scalability Issues: Maintaining living cells in a computing environment poses significant logistical challenges, including temperature control and waste management.
Commercial Viability: Companies like Cortical Labs and FinalSpark are exploring ways to monetize Biocomputing technologies, with products like the CL1 unit aimed at integrating biological intelligence into computing.

Featured Researchers and Sources

Brett Kagan: Chief Scientific Officer at Cortical Labs, involved in the DishBrain project.
Fred Jordan: CEO and founder of FinalSpark, a biotech company working with Brain Organoids.
Thomas Hartung: Researcher focused on toxicology and Biocomputing, involved in the development of Brain Organoids for drug testing.
Gordon Moore: Co-founder of Intel, known for Moore's Law regarding the growth of transistors in chips.

Overall, the video presents a fascinating intersection of neuroscience and technology, exploring the future of AI through biological computing.

Notable Quotes

— 01:15 — « Sounds like sci-fi, right? Well, we're kind of on our way. »

— 03:11 — « The promise of far less data and energy consumption is the holy grail for AI advancement. »

— 07:38 — « Our brains run on just 20 watts. That's the difference between a couple of LED light bulbs and thousands of homes. »

— 16:46 — « I am not the pope. I can fail. »

— 23:13 — « Whatever the future may bring, one thing is already clear: that brain cells will always deliver more than the sum of their parts, and that's neither boring nor binary. »