Human brain is superior than till-date computers, yet why we perform poor than computer?

Question

Scientists found storage capacity of synapses by measuring their size. They found, on average, a synapse can hold about 4.7 bits of information. This means that the entire human brain has a capacity of one petabyte. This is same as about 20 million four-drawer filing cabinets filled with text.

Striving around this stuff, I have a weird doubt crawling in my mind and it goes like this: Why is it like this that a human whose brain has got a capacity to store 1 petabyte, i.e. 1 million GB, cannot fully remember even the contents and information of a simple textbook which on the maximum would not be more than 1 GB, when he has to appear in exam? Why is storage capacity not directly linked with remembering capacity?

Answer 1

Three reasons:

Most of the storage capacity of the brain is used to store implicit knowledge. For example how to move your body, how to automatically recognise objects or how to hear and parse your language. These things may seem easy to you, but they require a lot of learning. Just look at how clumsy five year olds still are, they have been training how to use their body for five years straight and still skin their knees all the time.

Humans are actually able to retain huge amounts of information, both in their short term memory and in their long term memory. We know this because there are savants and hyperthymesia. It is likely that these abilities are actively suppressed in neurotypical people.

Because it is just not evolutionary adaptive to remember every detail. Intelligence depends on weeding out unnecessary information. If normal people learn huge amounts of information they often rely on memory abilities that are evolutionary adaptive, like the ability to remember a path from landmark to landmark.

Answer 2

To answer your question, we store information in a different way than a computer would, the storage happens in the connections in between neurons and it's not a perfect copy of what the environment presents, but rather certain features, I'll give you a classic example, read this sentence:

Tahkns for Cnotriubtng an asnewr to satck ecxahgne

You can understand the sentence even though the inner words are scrambled because it is not stored as a carbon copy, but rather as similar things attached to meaning within semantic and syntactic rules which in turn are stored as well, the recall and storage happens through a network in parallel, so the storage and the recall happens as you read it.

To give you a numerical example of how limited our brains are in terms of data as we know it, there are around 26,000,000,000 (26 billion) neurons on the Neocortex ( the number will probably go up-down in the future, but let's just use it for reference and assume it is only used for vision), your vision has 260,000,000 (26 million) cones and rods, where light is registered. If you were storing the information pixel by pixel in neurons alone, you would run out of space after just 100 frames. The synapses though are a different matter,a conservative estimate is around 150 trillion: 150,000,000,000,000, and if you were storing perfect copies, you would run out of space after ~ 580,000 frames, or 4 movies.

Answer 3

I think that it's not a store problem, but the way we access the stored information. Suppose you have a computer with 1 petabyte storage and you are asked you to find a file, that is stored somewhere. If you don't have a decent searching strategy or your files are not stored orderly, it would take you a very long time to find them.

This is why learning strategy focuses more on how to retrieve information in exams and less on how to store it.

Answer 4

Yup you are on a right question.

This concept is already well known as Turing Test, named after famous mathematician A. M. Turing. You might notice that if you ever read about CAPTCHA values.

Human and all other living beings (including a cell) can do some very complicated things. We do very complicated tasks at our day to day life, may it be chatting with your friends about random topics, finding a similarity between today's soft weather with trees waving in breeze with a similar one 2 years ago, or to understand and elaborately explain the meaning of 'rolling stone gathers no moss', or to be able to tell a short summary of a long story. or writing a poem. Or to 'create' new paintings.

Yes skeptically, all these tasks can be done with help of computer; but it will take huge algorithms and hardwares. These tasks are troublesome for a computer, but easy and 'simple art' (so-called) for us.

But what is difficult to most of us; such as to calculate a huge mathematical problem within instant; or to recite the sentences of a huge novel in exact ditto, or to serve some tasks in an exact routine; etc. is very easy for a computer, and takes much smaller algorithms and hardware-advancements.

Turing-tests try to provide tasks, those are easier to human (or perhaps animals), but tougher for a computer. It could be said, The better a Turing-test, the more it easier to human than computer.

But Why that difference? plausibly we're evolved in a way to perform some highly complicated tasks that we perceive as apparently "simple". We are (mostly) naturally talented on them. But we're not hard-wired for some 'real' simple tasks so we defeat in front of computer.