💻Digital intelligence is superior to biological intelligence.
🧠Mortal computation integrates hardware and software, offering energy-efficient analog computation.
🔌New nanotechnology and genetic re-engineering could enhance hardware efficiency.
🔄Finding biologically plausible learning procedures is crucial for mortal computation.
💡The loss of knowledge occurs when hardware dies in mortal computation.
What is mortal computation?
—Mortal computation integrates hardware and software, allowing low-power, analog computation similar to the brain. However, it faces challenges such as the loss of knowledge when hardware dies.
Why is digital intelligence superior to biological intelligence?
—Digital intelligence offers higher energy efficiency, scalability, and the ability to separate hardware and software. It also allows for the application of learning procedures and potential hardware enhancements.
What are the challenges in mortal computation?
—Challenges in mortal computation include finding biologically plausible learning procedures that scale to large networks and addressing the loss of knowledge when hardware dies.
How can genetic re-engineering enhance hardware efficiency?
—Genetic re-engineering could potentially allow the growth of hardware from biological neurons, leveraging their long history of learning and creating more efficient computation.
What is the impact of mortal computation on energy consumption?
—Mortal computation offers the potential for significantly lower energy consumption compared to traditional digital computation, making it more sustainable and efficient.
00:07Digital intelligence surpasses biological intelligence in efficiency and power.
08:51Mortal computation allows for analog computation with low-power requirements.
13:29Challenges of mortal computation include finding scalable learning procedures and addressing the loss of knowledge when hardware dies.
14:16Analog computation offers higher energy efficiency compared to digital computation.
15:00The loss of knowledge in mortal computation requires a teacher-student approach for transferring knowledge to new hardware.
