Much has been written about AI, ethics, the possibility that they can supplant humans in some areas, in fact some experts mention cases where humans have lost jobs and their place is taken by some robotic intelligence mechanism artificial. We talk about the criteria that we humans can handle, the reasoning when making decisions, the feelings and emotions involved in each decision.

Source

So, are the feelings and emotions decisive in the decisions we make? ... I think so! Taking into account the content and motivation of this publication, in my opinion it is something scientifically proven since it has direct impact when commenting and voting on the publications we read on steemit.

I want to draw your attention to this point and I would like you to continue reading this publication that will only take a few minutes. I go into the matter ...

Discovery ...

A few days ago reading this post 6 reasons why you should power up of my good friend @achim03, I found this comment in the image which got my attention a lot, this comment, maybe some of you have seen it many times but I want to discuss with you its implications.

The message clearly expresses, (according to the criteria of a BOT) the positive votes, the reward that the post should receive according to its quality, that is, if this post does not reach the estimate by the BOT, then this publication is Sub valued.

This caught my attention because two months ago one of my publications Steemit Hard Fork in perspective received a similar comment, where the Bot makes an approximate assessment of what should be the result of this publication in terms of rewards and positive feedback, this comment contained the following image information:

I understand that they are approximate ratings, predictions, conjectures maybe or possible results but my point is,what is happening to us? What are our healing criteria? Does artificial intelligence exceed us?

There are many answers we can have to these questions, as well as many reasons to consider that a publication deserves to be voted on by me, but there are things that I still cannot understand in steemit and I need you to help me with this.

I don't understand, as someone who has a voting power that can give you 0.03, decide to vote your post with 15% or 10% of their voting power, what is the point of this?, if you really believe that some publication deserves your vote give it something worthwhile. Although I do not understand the technical implications of this, when I vote for a publication I want to receive some reward according to my voting power.

About the BOT

This is a Bot called @TrufflePig created by the user @smcaterpillar, this bot has a sophisticated system to "measure" the quality of the publications, I want to share with you fragments of this publication: I am a Bot using Artificial Intelligence to help the Steemit Community where you will find more information about its operation, the programming it has for its operation and the criteria established to determine when a publication is undervalued and when it is not.

My basic idea is to use well-paid past publications as examples of training to teach a part of me, an Automatic Learning Regulator (MLR), how high-quality Steemit content looks. In turn, my trained MLR can be used to identify high-quality publications that the curatorial community lost and that they received much less payment than they deserved. I call these publications truffles.

I use some functions that encode the design of publications, such as the number of paragraphs or the number of headings. I also care about spelling errors. Clearly, publications with many spelling errors are generally not high content Quality and, in my opinion, are difficult to read. In addition, I include readability scores such as the Flesch-Kincaid index and syllable distributions to quantify how easy and enjoyable it is to read a publication.

Even so, the question remains, how do I code the content of a publication? How to represent the theme that someone chose and the story that an author told? The simplest coding that is used quite frequently is the so-called "reverse document frequency of term frequency" (tf-idf). This technique basically encodes each document, so in my case Steemit publishes, by the particular words that are present and weighs them by their frequency of occurrence (heuristically) normalized.

For the training, I read all the publications that were sent to the blockchain between 7 and 21 days ago. These publications are first filtered and then coded. Too short, too long, non-English publications, whale war publications, publications marked by @cheetah or publications with too many spelling errors are removed from the training set.

Conclusion

Finally, I think that human beings are more than AI, that our voting criteria, our commitment to our neighbor, for all of us who live in this ecosystem, that we mutually benefit from what we do, empower and support each other. daily should be much higher than those of a BOT.

There are things that artificial intelligence will never overcome and it is that feeling of empathy, having an idea of ​​how the other feels in a certain situation, the satisfaction that we can only feel when we help someone or the feeling of frustration, pain or disappointment when we see that someone falls or is in need, we cannot lose those values ​​that differentiate us from the machines, everything else is statistics and copied behaviors that can be erased from a processor.