narrow ai and agi
First of all I suggest you to read carefully this that is one of the most interesting analysis about the state of art about artificial intellicence.
Narrow AI is every moment nearer and it will revolutionize heavily our job in the next years.
But let me tell an irreverent joke about specialized doctors, dermatologists, diabetologists, oncologists etc.
“They know everything about nothing”.
Their narrow view in so “narrow” to be reductive. And the poor patient must take three pressure tablets, two for cholesterol, two for diabetes ….. to please all specialist doctors.
Except to find that all these diseases “globally” fall into the same metabolic syndrome!
The takeaway message is there must be an equilibrium between narrow and general intelligence, and one cannot exist without the other.
And most important of all, the reality is not virtual like a a chess board or a Go table.
World is real Extremely changeable and complex, and our view of it is always an approximation. Such as technical analysis in finance.
“If you thought that science was certain — well, that is just an error on your part” (Feymann).
We read that narrow ai “….automation that works well in highly complex domains. The automation will thrive in investigating extremely high dimensional problem spaces….”.
It is like solving encrypted coding with a simple brute force attack. History teaches us that Turing had to apply many algorithmical shortcuts for solving in this “high dimensional problem space”. He had to apply AGI, because reality space cannot be “squeezed” into a so big but infinitesimaly smaller space. And best clues come from AGI.
Narrow AI plus AGI.
Probably AGI doesn’t have that economical appeal of narrow AI, but the last is a false promise without the first.
Do you like having a full artificial narrow ai medical examination? Good luck!
Please read the IntuitionMachine.com article.
To be continued……