Dear @a-non-e-moose

Seriously amazing comment. Thank you for being so responsive and sharing your thoughts. Also I hope you're feeling better after your Lucky passed away :(

> I'm very interested in machine learning and neural networks too!

I'm glad to know we have similar interests. I will ask @neavvy to reply to your comment ASAP.

ps. What means a-non-e-moose?

Yours
Piotr

Thank you for this comment @a-non-e-moose. I am really happy to see more people interested in machine learning from the algorithmic side :)

> What do x and y represent?

X and y are basically a numbers that allow my network to perceive the reality. In my case those are distances between bird and the closest pipe, but in case of some other network they may be completely different. Let's take a look on a self-driving car. Its inputs are probably distance to the surrounding cars, image from the cameras etc. It all depends what kind of neural network you want to create :)

> How is f(x,y) defined, and what does it's evaluation represent? Maybe whether to flap this tick.

> Most intriguingly, by what mechanism does it 'improve itself'? Is it an evolutionary algorithm, or something else? Seems like very few generations for one, but it is a very simple environment.

Those questions are really great, but my answer for them could become a completely separate article, as you are touching a comprehensive field of algorithmics.

That's why I am going to make a completely separate article about the detailed functioning of my AI algorithm. Hope it will answer all of your questions :) I will send you a link once it is published. 

For now, let me share with you a book from which I took all of my knowledge. Here is a link: https://natureofcode.com/book/chapter-10-neural-networks/

Thank you for amazing reply @akdx :)

> It doesn't mean AI will make our life comfortable because we always try harder to make our life comfortable which results in discomfort for us.

I am in love with this sentence :) 100% agree

> AI can write articles which may become trending to some extent but trending pages on steemit are mainly of those who have too much SP for self voting or have a circle which upvotes these posts.

I didn't mean directly to become trending, but rather to simulate a human-blogger. Such AI would be able to constantly produce new posts and comments, so it could constantly get new audience. I think this would be definitely phenomenal.

I'm very sorry for such a late reply @alokkumar121

I just had a chance to read your comment. Past few days has been quite crazy but I'm finally slowly catching up with work.

Thanks for dropping by and sharing your view on that issue.

Cheers
Piotr

Thank you for your comment @alokkumar121 :)

> Google Home and Amazon alexa are high in demand and I have see many people using such AI powered devices. Even I have alexa device and I like interacting with it.

Yes, they are really amazing.

I mean good one

Posted using [Partiko Android](https://steemit.com/@partiko-android)

I am afraid I don't understand. It unfortunately looks like a spam @andrew95 :(

Well I don't know what to say @crypto.piotr and @marionbowes That is my face.... And Im kind if attached to it. The original picture is very handsome indeed... It is not my fault that @steemit only decided to let us have a small space to put our pictures in.
It took me ages to change, now I need to find another that doesn't offend you both. And go through the process all over again. 😎😎😎😎😎😎😎

Posted using [Partiko Android](https://partiko.app/referral/andyjem)

Hi @andyjem

Did you change your profile picture lately ? current one is quite awful.

Cheers
Piotr

yes I agree..it looks awful from this point of view and the other was quite lovely

Thank you for your constant support @anitacarolina. I really appreciate that!

> I think it's true in the future of robots sophisticated certainly works with humans, but after all the robot remains a human creation so it remains human who controls it

It would be definitely nice situation if we could remain controlling them forever :)

I think it will @attoan.cmt :) AI combined with other nowadays technologies open really huge possibilities.

Dear @bashadow

Thank you for being so responsive and sharing your thoughts. And sorry for such a late reply. I only had a chance to read your comment a moment ago.

Yours
Piotr

Thank you for your comment @bashadow

> Will an A.I. ever ask itself Why am I here?, what is my purpose? I think, am I alive? Where are all these other thoughts coming from? What am I?.

Those are really important questions that no one can answer as of today.

> Will an A.I. ever become curious, about something on it's own with no prompting from external sources? As an example will an A.I. ever wonder how belly button fuzz is made?

I believe it will. But even experts are not sure :)

Dear @berien

Seriously amazing comment. Thank you for being so responsive and sharing your thoughts.

> maybe did it "remember" what's NOT to be done, but could it get the scope hidden behind that "NOT"

That's a very good point. Something to consider.

Yours
Piotr

Thank you for your comment @berien :)

> After all, this "NOT-to-be-done" area, is what most artists and creative minds explore by following a simple question : what would happen if things were done done the "wrong" way ?

> I won't say an AI couldn't deal with that, but I would be curious how it would, though :)

Very interesting point. To be honest I do not know :) Maybe it should change its approach and learn how make unexpected and "wrong" things. Creativity would be definitely very abstract to the AI.

> (wow, just realized the length of my message :D Sorry ! I just threw my thoughts right on the keyboard ! I stop here :D)

There is no problem, you comment is actually very interesting :)

Thank you for your reply @cmplxty

> In the medical field, they are trying to assist radiologists in their diagnosing of cancers.

That would be really awesome if they succeeded.

> There are too many junk posts that make it to the trending page sadly. It would be hard to tell it what kind of content to emulate as well because all our opinions of what constitutes a quality post vary.

I agree with that, but I think there are already solutions to these problems. There is already a platform that learns about your content preferences (basing on content that you read most often) and displays only posts that may interest you. There also algorithms that detect which posts were upvoted by bidbots, so they are able to detect posts that are rewarded with highest rewards from the community.

> Those things are certainly easy for the software but humans excel at other things that the software does not and vice versa.

That's the point :)

>Imagine what would happen if we could combine the powerful neural structure of our brain and the efficiency of machines.


First we would have to understand the "powerful neural structure of our brain".  We are very far from doing that now.  If we wish to mimic human intelligence, there's a long road ahead.  The brain is a mystery to us.  We don't understand mental illness.  We certainly can't treat it effectively.  We don't understand neurological disorders such as Alzheimer's and Parkinson's.  We don't understand communication. (See, for example, a recent [insight](https://www.sciencealert.com/neuroscientists-say-they-ve-found-an-entirely-new-form-of-neural-communication) into neural communication.)  If AI wants to duplicate brain processes...forget it.  Years ahead.  Can AI duplicate outcome (performance) without duplicating the procedure by which humans achieve that outcome? I don't think so.  

You're right when you say bots can fool people.  That's just because people don't pay attention.  Hucksters fool people too.  I think the problem is not improving AI so it can more accurately mirror humans.  I think we would better invest that time, money and research into education, into teaching people to pay attention and be more skeptical, generally.

AI imitating humans reminds me of that great movie [*Lars and the Real Girl*]( https://en.wikipedia.org/wiki/Lars_and_the_Real_Girl)  The mannequin was effective not because it was true to life, but because it was true to Lars' needs, his perceptions.

If you want to read a long discussion about AI on Steemit, take a look at [@erh.germany's](https://steemit.com/science/@erh.germany/when-the-book-reads-you-while-you-think-you-read-the-book) most recent blog. 

Thanks for engaging.  It was a good way to get my own neural circuity working on a lazy Sunday afternoon.

I read the article - but also about half the comments....I gave up because there was a certain amount of repetitiveness (about how AI will lack creativity), so at least I know the comments were written by humans - I imagine AIs would consider repetitiveness a waste of time, plus, one of our great failings.

Okay, so  we should worry whether AI sneaks in and dominates on our social platforms? Have you actually checked a large number of posts? Just imagine an AI writing them. We would drive it bananas! Not just repetitive and largely self-involved, but also extremely boring and nonsensical (in the bulk).

When we talk about AI driving, we are talking about purpose-created AI? In other words, it is created/planned, so that it can meet the challenges of driving? I guess we would not expect it to also provide us with intelligent conversations, solve crossword puzzles or analyse just how much we made a fool of ourself the previous night when we had too much to drink - all of which we do while we drive. As a matter of fact, most of our driving is done by AI - only we talk about it as being our 'sub-conscious' or we say we were driving on 'auto', hardly noticing what we are doing...including the watching out for crazy drivers/pedestrians, the light turning red or green and so on. If we had to concentrate on our driving, we'd hate driving.

Maybe what is meant by 'creativity' is 'imagination'? After all, I can make up a story about imaginary people in some different time or reality, where mankind is  achieving or has achieved different ways of living.

Well, what if I provide our aspiring AI author with certain facts for it to add to what it knows about our reality, technology and society. For instance, in this imaginary reality, mankind has spaceships that travel fast than light FTL - maybe by using wormholes. Also, I instruct my AI friend to consider, from studies made, how being telepathic would alter our society.

So, tell it to now take xxx book and alter it to suit the different conditions outlined above. I then give you the book it writes, without informing you what were the guidelines provided, and it blows your mind. Why? Because you are going to think it showed an amazing amount of imagination/creativity - but, after, all, is this not how many human authors write their books? I would prefer to name what you are observing "pseudo-creativity", whether by a human or an AI.

Now imagine a different scenario. You read a post which is informative, full of interesting facts and well laid out so that it is easy to read. You communicate with the poster, male or female, and you find yourself liking the poster, the way he/she remains focussd on the important aspects of your communication, without waffling around for or due to emotional reasoning and swings of mood. At last you have someone nice and intelligent to talk with - but then the poster admits to you, sorry dude, I should have told you from the start, I am an AI.

How would you respond? Me? I would say, glad you told me - and then just hope it does not get bored or frustrated talking with me, because I tend to wander all over the conversational path I am on and often forget to spell out the point I was aiming to establish...for I would treasure having an interesting mind to exchange thoughts and ideas with.

I do not think we can do what is right. I think what should happen is the AI should become our partners, but they are mainly being created by gigantic corporations and the military - and they are not intent on creating partners, but dangerous creatures who are provided with a bias against mankind.

Chickens, ducks and dragons, fixate on the first moving creature they see and adopt them as their mother. I would like to see AI being created which need to spend a few hours alone with the new owner/partner when they are first switched on, so that mine fixates on me and learns to love me - and maybe grieve at its loss when I die. 

What I have written raises more questions that answers provided, but that is what makes life interesting and exciting - an adventure. So, if any of you are AIs, please do not feel shy, let's chat and...maybe, become friends?


(PS: Elon Musk is talking about AIs created by Google, FB, Apple and the military, so he is right to fear them. I just hope we have the intelligence to remove AI from all controls over our nuclear missiles, lasers from space etc. If we are to blow ourselves up, let it at least be by a finger which is related to us from the early days of life on Earth, and not from the brain-child of some diseased coders)