computers can't understand

In Making Computers Understand Leslie Walker reports on an apparent innovation/invention suggesting that computers can understand and be aware of context. While the phraseology chosen might be a journalistic lingua franca to ‘spice’ the article, nevertheless, some claims by the company are rather troublesome:

“Abir, 46, claims to have unlocked the mystery of "context" in human language with a series of algorithms that enable computers to decipher the meaning of sentences -- a puzzle that has stumped scientists for decades.”
"This man literally has figured out the way the brain learns things," Klein said. "On a theoretical level, his insight basically is this: Understanding a concept is nothing more than viewing a concept from two different perspectives."

The very title of the article "Making Computers Understand" makes you immediately skeptical. Especially troublesome is the above quote stating that “This man literally has figured out the way the brain learns things”? Isn’t it perhaps premature to claim that we have discovered how the brain works with such certainty when the history has told us that many such claims in the past have been proven wrong by later discoveries and innovations?

Further, how does one prove that two different perspectives are sufficient to understanding a concept? I hope this does not mean that they believe two perspectives are necessary since there are ‘two sides of the same story’. Usually there are more then two sides to the same story and understanding the ‘reality’ and its context probably might take much more than two perspectives.

Besides, computers can’t decipher the meaning of a sentence as claimed in the article…