“Imagine a future in which the boundaries between biology and technology disappear,” he announces his activities. A Swiss startup proposes a new method for building models (SI). Instead of using regular digital signal processors, they plan to introduce a system based on bioprocessors. They would use neurons (nerve cells) obtained from human skin.
Artificial intelligence like the human brain. The Swiss startup proposes an innovative solution
The proposal of the Swiss startup’s scientists would bring several significant benefits to the field of AI. The main one is to be significant energy savings. According to research cited on the project website, the human brain uses 86 billion neurons and uses only 20 watts of energy to perform calculations. Replicating the processes taking place in the human brain using conventional silicon technology would require as much as 10 megawatts.
The use of biological neural networks would therefore allow for a significant reduction in costs and efficiency compared to traditional AI. it also praises its technology as highly scalable and constantly learning. Thanks to these features, the AI offered by FinalSpark would be easier to adapt to expectations and create a universal system that could be used in various processes and would not be limited to one task. Unlike regular AI, the FinalSpark project would create a kind of collective intelligence that would freely absorb new data and adapt to the problems presented to it.
New technologies based on human neurons. “A real thinking machine”
Fred Jordan, co-founder of the startup, revealed in an interview with a technology portal the methods used to create a modern AI model. FinalSpark tries to teach neurons to process information through electrical wires the way the human brain does. For the purposes of the experiments, neurons grown in cell cultures were used.
Current AI models are able to mimic human thinking after long months of training on advanced data. FinalSpark’s ultimate goal is to create a model that can replicate actual human reasoning. This type of AI would be able to analyze emotions, as well as create new ideas and concepts that do not result from direct experience. – This is how a “real” thinking machine should work. After all, the most famous information processor is the human neuron, Jordan said in an interview with the portal.
The startup’s co-founder told Quartz that this type of biological engineering may completely replace traditional methods of developing new ones in the future. According to Jordan, DNA-based systems could also become a more favorable alternative to cloud-based data collection.
Source: Gazeta
Mabel is a talented author and journalist with a passion for all things technology. As an experienced writer for the 247 News Agency, she has established a reputation for her in-depth reporting and expert analysis on the latest developments in the tech industry.
