Computer scientists have been seeking, over decades, a counterpart to the human brain’s neural networks to enhance the processing power of artificial intelligence (AI). As these artificial neural networks refine and grow in power, they consume increasingly more energy. In turn, it prompts a search for more efficient alternatives. A Swiss start-up called FinalSpark has introduced a “bio-computer” that interfaces with living brain cells in a phenomenal display of nature’s efficiency. It boasts of significantly lowering energy consumption compared to traditional computers.
Bridging Biology and Computing
FinalSpark’s ingenious platform uses methods of merely integrating biological concepts into computing. Thus, it taps into spherical clusters of lab-grown human brain cells known as organoids. These organoids are housed within arrays connected to electrodes and a microfluidics system. Also, it represents a cutting-edge approach known as wetware computing. FinalSpark’s system allows for insights into the functioning of miniature replicas of individual organs, leveraging researchers’ ability to culture organoids in the lab.
ChatGPT and Beyond, FinalSpark’s Quest for Sustainable AI
In the exponential growth of artificial neural networks, the innovation serves to be exemplified by large language models like ChatGPT. FinalSpark asserts that their bioprocessors, such as the brain-machine interface they’re developing, consume a fraction of the power required by traditional digital processors.
The substantial data on their system’s energy usage and processing power is yet to be disclosed. However, the comparison with the minimal energy consumption of training large language models like GPT-3 is considered quite impressive. The contrast between the energy efficiency of the human brain—operating its 86 billion neurons with minimal energy—and contemporary computing methods emphasizes the urgency to optimize energy consumption in the AI industry.
The AI industry is expected to cause a significant increase in global electricity consumption, according to the prognoses. Since this portion is considered to be significantly larger, it can cause aggravation to environmental concerns. Against this environment, the confluence of brain cell networks and computing circuits presents an optimistic pathway for energy-efficient computing solutions.
Remote Access System for Brain Organoids
FinalSpark carries its ambition further by building upon past efforts to link computer hardware with biological systems. The AI industry is expected to cause a significant increase in global electricity consumption, according to the prognoses. An American company, Neuroplatform, demonstrates this trend by enabling extensive data collection from brain organoids for years.
Even though it was primarily aimed at studying purposes, FinalSpark’s system allows for remote access and supported electrical activity durations in mini-brains up to 100 days. The future could see more research groups adopting this as a platform for their experiments and hence the potential is expanding. These may also encompass wetware computing experimental protocols.
Thus, whether improving computer efficiency or developing organoid models, ongoing innovation in this area could yield exciting advancements at the intersection of computational science and biology.
