<div class="pull-right"><img src="https://steemitimages.com/DQmeyEUxhYnYrgyjStSPGSM6pHeVr1JHyhEE6yBWW7uT8pf/image.png"><br><center><sub>Max Pixel <a href="https://www.maxpixel.net/Ai-Artificial-Intelligence-Droid-Robot-Android-22286104">(link)</a>
CC0 license</sub></center></div>

<b>TLDR;</b> Computers with their current architecture cannot think. I am going to go into the reasoning why this is so in this post (hint: everything in a computer is strictly mechanical and thinking is not strictly mechanical).

#### Introduction
Artificial intelligence (AI) is all the rage right now. From Alexa and Siri understanding your verbal commands, to Deep Blue playing chess and beating Grand Masters, to cars driving themselves in real traffic situations it would seem that computers are getting good at thinking.

I am going to take the polar opposite position to this; computers cannot think and I will explain the reasons why this is so.

<div class="pull-left"><img src="https://steemitimages.com/DQmPGWvgpTifrLaPBskmQShcYJ8povRkzQSsqTUDnRceiN4/image.png"><br><center><sub>Door Knob Mechanism<br>Pxhere.com <a href="https://pxhere.com/en/photo/1376334">(link)</a>
CC0 license</sub></center></div>

#### The Lowly Door Knob
First, let's consider the simple door knob. Turn the knob, a mechanism inside activates to retract a latch and the door can be opened. Release the knob and the mechanism executes the opposite mechanical action and the latch is released back out and enables a door to be blocked from opening.

It is obvious that this latch/unlatch mechanism is not thinking. It is not even what could be termed memory even though it is a mechanism that can be in one or two states.

<div class="pull-right"><img src="https://steemitimages.com/DQmZYCcUbAMgfFYM2h5ZMkcBaJWLNAVL8c74jYpEuiqJu8U/image.png"><br><center><sub>A Mechanical Calculator<br>Claas Augner <a href="https://commons.wikimedia.org/wiki/File:Thales_A_calculating_machine_2009-08-23.jpg">(link)</a>
CC BY-SA 3.0 license</sub></center></div>

#### The Mechanical Calculator
Let's move on to the mechanical calculator. Some of these are wonderful contraptions in which you can punch in numbers and then add, subtract, multiply and divide.

They are terrifically noisy and can do some mathematical calculations but you can look inside them and see that they are also obviously mechanical. The actions that they perform are also obviously not thinking even though they seem to be doing what we can do in our heads (some of us at least, I have the time tables only memorized up to about 12 x 12).

<div class="pull-left"><img src="https://steemitimages.com/DQmVbPZwPgYSC3i6K2K2csy4Rbrgoukpb5wGUWab3xVGW1J/image.png"><br><center><sub>Charles Babbage's Difference Engine<br> 	Jitze Couperus <a href="https://www.flickr.com/photos/jitze1942/4304353299/in/album-72157623284316506/">(link)</a>
CC BY 2.0 license</sub></center></div>

#### Babbage's Difference Engine
Moving the next step up.

Charles Babbage designed his [difference engine](https://en.wikipedia.org/wiki/Difference_engine) sometime around 1820. Although his success in building it back in that era was limited due to the technology levels in metallurgy and manufacturing a working model was successfully completed in the [1980's](https://en.wikipedia.org/wiki/Difference_engine#Construction_of_two_working_No._2_difference_engines).

These engines are automatic calculators that are capable of tabulating polynomial functions.

Again, you can look inside them and see that a difference engine like this is obviously mechanical. The actions that they perform are also obviously not thinking even though they can do math that humans can only do using paper and pencil.

<div class="pull-left"><img src="https://steemitimages.com/DQmVm2qMDNk4JxN18iReYDpE11gNhvzpSaDmjKATzQy4c32/image.png"><br><center><sub>A Digital Adder Circuit<br>Image credit: Inductiveload <a href="https://commons.wikimedia.org/wiki/File:Full-adder_logic_diagram.svg">(link)</a>
Public domain license</sub></center></div>

#### Electronic Digital Circuits
Moving a step up again to electric digital circuits.

They are not obviously mechanical but if you study their theory of operation in any digital circuits course it will become clear that they still are mechanical in operation. They have states that are well-defined and they move on to new well-defined states that in a completely deterministic manner.

The example in the figure is an adder circuit. It will add two binary digits, provide a sum (also in binary) and a carry bit (again in binary). The circuit is predictable and even though it is electric, it is still entirely mechanical. Combine many of these to add up bigger and bigger numbers.

The actions of these circuits is also not what I would consider to be doing what is called thinking.

<div class="pull-left"><img src="https://steemitimages.com/DQmVGQ1TsodS75Npf7QdWJ7QotaoWWjup7HCbg8SEuPmNJU/image.png"><br><center><sub>CPU Block Diagram<br>Image credit: R. S. Shaw <a href="https://commons.wikimedia.org/wiki/File:CPU_block_diagram.png">(link)</a>
CC BY-SA 3.0 license</sub></center></div>

#### CPU Circuits
The circuitry for a modern CPU is vastly complicated but it is only complicated because it is made up of very many "simple" and entirely understandable digital circuits. The word "simple" is in quotes here because it does take some study to understand these circuits but they are, in the end, completely understandable.

They are also both mechanical and deterministic in nature (complex yes, but still mechanical). The whole fetch, decode, execute, store cycle is entirely mechanical. Even the memory circuits in a computer are mechanical. The term 'memory' is borrowed from what human brains can do and applied to these circuits but that seems to me to be a quite a loose association.

#### Analogy to Thinking
So, modern computers are very fast and they seem to be able to do what we can do but these are only mechanical activities mimicking the functions of a real thinking brain.

Here comes the key point.

The confusion may lie in this observation: our thinking brains are so adaptable that they can perform mechanical functions like comparing objects, performing math and using logic to manipulate concepts (I also wrote a post on the limits of logic about 3 months ago [here](https://steemit.com/science/@procrastilearner/the-limits-of-logic)).

When we do these things it is just our brains dropping down a step to perform purely mechanical operations before going back up a step to perform proper thinking.

Let me repeat that: humans brains can perform mechanical tasks but that is only a subset of what they are capable of. This mechanical subset of activities is what we are able to get modern digital computers to mimic.

I consider these mechanical mental activities to not actually be thinking. Terms like understanding, comprehending, consciousness and creating better describe what thinking is.

<div class="pull-right"><img src="https://steemitimages.com/DQmPTnxT2ZKR1NS18WL9RdMCEYjxHGupGzQdg5NUUU8ADmU/image.png"><br><center><sub>The Thinker by Auguste Rodin<br>Image credit: Pxhere <a href="https://pxhere.com/en/photo/1089623">(link)</a>
CC0 license</sub></center></div>


#### What Is Proper Thinking Then?
I cannot answer that, nor can anyone else for that matter. Thinking and consciousness is still one of those uncracked problems and we do not seem to be be close to actually figuring that phenomenon out.

We don't know what thinking is but we are able to understand that we are doing it. This post is not about defining what thinking is, it is about illustrating that the mechanical activities of a computer are not mimicking thinking, they are just replicating the mechanical aspects our what humans brains sometimes do.

#### Closing Words
In summary, modern computers can only perform the strictly mechanical aspects of what our brains can do, albeit much faster and with a much smaller an error rate.

This is not to say that computers will never think but it seems that they will never be able to think as long as we keep using the current architecture for digital circuits.

Maybe something designed more along the lines of how brains in nature are constructed may be needed to do that particular job.

*Thank you for reading my post.*

#### Post Sources

https://en.wikipedia.org/wiki/Difference_engine
https://en.wikipedia.org/wiki/Instruction_cycle
https://en.wikipedia.org/wiki/Adder_%28electronics%29
https://en.wikipedia.org/wiki/Memory_cell_%28computing%29
https://en.wikipedia.org/wiki/Digital_electronics