WILL MACHINES SOMEDAY REALLY BE ABLE TO THINK?

In 1950, when silicon microchips didn't exist yet, Turing predicted that, as computers grew smarter, the issue of artificial intelligence (AI) would eventually come up. According to Turing, a computer could “think” if it could fool a human being using his(her) natural language. The mathematician guessed that in 50 years – that is, today – we would be able to build computers as capable as humans in natural language processing so that they would pass the Turing test.

Until now, the forecast has not come true. Scientists claim that computers suffer from restricted design, that is, their programming only aloows them to do very specific jobs. Furthermore, they also claim that the processing speed of microchips still hasn't attained the speed of the human brain

In July 24th of this year, the Riken laboratory, located at the Yokohama Insttute, Japan, presented a computer capable of processing 1 petaflop, that is, roughly 10% of our brain's capacity.

Cortex Intelligence does state-of-the-art research in the field for the past three years. Our most recent results are optimistic and show that the storage capacity is not the problem. All linguistic knowledge of the human brain can be stored in a modern iPod, referring to the space occupied by Cortex. We aren't yet capable of simulating the linguistic capacity of the human brain because our machines are still much slower than the human brain.

Click here for the whole story (in portuguese)