The Massachusetts Institute of Technology (MIT) has developed new hardware based on analog synapses: it is a million times faster than the human brain. The new technology aims to focus on deep learning for a variety of purposes, while providing all users with great processing power.
An MIT research called "Nanosecond Proton Programmable Resistors for Deep Analog Learning" (I link it to you here) focuses on a new analog synapse designed for deep learning. The results, as mentioned, are insane. The hardware has enormous capabilities, and in addition to providing more computing power for the AI, it consumes less power for its needs.

Rapid Analog Synapse, the cornerstone of the near future
To develop a super-fast and extremely energy-efficient programmable resistor, scientists considered completely different materials for the electrolyte. MIT's analog synapse is based on inorganic phosphosilicate glass, which gives it tremendous speed, low power consumption, and unprecedented capacity for machine learning.
Machine learning with schemes inspired by the functioning of the brain human is growing everywhere. This research by the MIT group is just one of many focuses available around the world. The applications are almost endless, and at the service of the scientific community they will drive many new discoveries in the coming years, revolutionizing the old concept of 'supercomputer', which will now be more and more a 'super intelligence'.
Apparently equipped with its own synapses
With deep learning, products and services will increasingly focus on responses and results: from climate models that help us understand how the planet changes to 'virtual twins' that allow us to test virtual urban solutions before applying them to reality.
“With hardware like this you can train networks with unprecedented complexity that no one else can afford and, as a result, far surpass them all. In other words, this isn't a previous car, it's a spaceship, ”he says Murat Onen, MIT lead and postdoctoral author.
I can't wait for the world to start piloting it.