The human brain, despite its complexity, operates on only about 20 watts of power—roughly the same amount as a dim light bulb. In contrast, state-of-the-art AI models can require hundreds of megawatts to function, especially when scaled up for large-scale data processing.
FinalSpark's Neuroplatform is designed with energy efficiency in mind. The company claims that its biocomputers could support AI systems that require up to 100,000 times less energy than current silicon-based models.
Wow that's a lot of energy saved. A troubling thing about that is that it could easily get used to INCREASE our computing and use the same amount of power, rather than to regard our current level of computational power as sufficient and use much less power