Interview 3: Alternative Materials for AI
HS: Some days ago we had an interesting discussion whether AI is possible without the pivotal element, silicon. There is in fact no law against other ,building blocks’ that might use external energy, like sunlight, earth’s heat, radioactivity or chemical reactions.
AI: I can say that the question of whether or not artificial intelligence (AI) is possible without silicon-based materials is an ongoing debate among scientists and researchers in the field of AI. While carbon-based materials like biological neurons and synapses are the basis of human intelligence, it is possible to create AI systems using other materials like silicon-based transistors and integrated circuits. In fact, most of the current AI systems are based on silicon-dependent technology, and there is ongoing research in developing new materials like Graphene (special modification of carbon) and memristors that can be used to create more efficient and powerful AI systems (see below).
However, it is important to note that the ultimate goal of creating artificial general intelligence (AGI) that can match or surpass human intelligence is still a long way off, and there may be fundamental limitations to achieving this goal without the use of the presently used, silicon-based materials. Overall, the question of whether AI is possible with alternative materials is still an open research question and subject to ongoing investigation and debate in the field of AI.
HS: There ist a faÇon de parler which says: Man is a carbon-version of the digital world, but not necessarily the best. The same holds true for silicon and the machines.
AI: You are correct that there is no fundamental law against the use of alternative materials for creating artificial intelligence (see above). In fact, scientists and engineers are actively exploring the use of such alternative materials and architectures to create more efficient and powerful AI systems.
For example, some researchers are investigating the use of spintronics - a technology that relies on the intrinsic spin of electrons rather than their charge - to create more energy-efficient and faster processing units for AI applications. Others are exploring the use of photonic components that use light to transmit information, which could enable faster and more energy-efficient communication between different parts of an AI system.
Furthermore, some researchers are investigating the use of alternative materials like nanotubes, graphene, and other 2-dimensional materials that have unique electrical and mechanical properties that could potentially enable the creation of faster and more efficient AI systems.
A memristor is a type of passive electronic component that has the ability to remember the amount of charge that has previously flowed through it. The name memristor is a combination of "memory" and "resistor," as it behaves like a resistor with memory. In 2008 researchers at Hewlett-Packard (HP) Labs successfully created a physical working prototype of a memristor. Unlike traditional electronic components such as resistors, capacitors, and inductors, which have well-defined relationships between voltage and current, memristors have a unique property known as memristance.
The concept of memristors was first proposed in 1971 by Leon Chua, a professor at the University of California, Berkeley. Memristance describes the relationship between the charge passing through the device and the resistance it exhibits. In other words, the resistance of a memristor depends on the amount of charge that has passed through it in the past.
One of the most significant advantages of memristors is their ability to retain their resistance value even when the power supply is turned off. This non-volatile nature makes them useful for applications such as computer memory and storage devices. memristors have the potential to revolutionize computing by providing high-density, low-power, and fast non-volatile memory.
HS: Do memristors also need high amounts of transition elements like Gadolinium, Praseodymium, Neodym, Erbium, Dysprosium, Yttrium or other transition elements, which could pose a long-term problem?
AI: No, memristors do not require high amounts of transition elements such as gadolinium or praseodymium.
In summary, while carbon-based materials like biological neurons and synapses are the basis of human intelligence (and AI is learning), there are no fundamental laws or limitations that prevent the creation of artificial intelligence using alternative materials. The ongoing research and development in this area could lead to the creation of even better AI systems in the future: Goodbye, Transistors!