saprakas

[saprakas]

#Blockchain and Bitcoin - Money, Transactions and Crypto

[saprakas]

(via How to: Proof of Existence - Smart Contract on Ethereum (to be elaborated...))

[saprakas]

#13 CMS WebBuilder With Integrated Cross Domain inventory population with JSON API

[saprakas]

Time Machine: 2070 will see us Time Travelling with help of Neural Networks!

Its weird but I feel in the next 50 years or so, with computers having yotta-flops capabilities, we would be able to predict our own future with help of machine learning! 

So, its a typical Sunday morning here in NY and I didn’t sleep and was watching this Flash Series on Netflix with Barry Allen and his time travel bull shit. Kinda intriguing but no solid theory for time travel there (Other than an overview of  Theory of relativity and Light Speed, although I am not a physicist , only a computer scientist).  

My eyes were hurting so I got up, wrapped my towel around and went to my bathroom when it hit me! 

Whenever we are watching a movie, or a serial, we subconsciously are able to predict what’s gonna happen in the next half second. Lets take a moment and  think about it. You start watching your movie and after a while you are able to predict what your protagonist ‘s next move will be. Now, I am accounting for a probability function here, in the sense, yes there are a lot of shocks and unexpected change of events which account for climax. but If you are a binge watcher and have seen all the superhero flicks, you will be able to predict if your hero is angry what’ he will do in the next half  second. Similarly, If we you are a fan of love  stories , and have seen all the Ryan gosling and Channing tatum lovey dovey movies, you can predict the next half second and the anticipated drama and expressions on the face of Rachel mcadams. (Yes, there would be few outlier as well but as a whole you will, with a good probability) That’s why you will be less receptive to a new scene than someone who doesn’t watch such genre. 

Lets’s think about why this happens? We all know our brain is a pattern matcher. Or say its kinda a machine learning model based on neural networks. Through all our childhood and situations around us, we respond to events and moments differently. Someone who grew up in present day Syria would be less responsive to violence afterwards than someone who was lucky to be born out of the dark forest and lived in an awesome world. Right?

What I mean is the guy who saw a lot of troubles in his life has trained his mind to work with more diverse datasets and build the machine learning algorithm with which he operates in the world accordingly. That algorithm takes into account the outliers as sample points as well  with enough evidence to prove its not an outlier.  His  predictions wold have been completely different from say mine. That’s why his brain can predict happening of a calamity in the next half second with more probability than I will do. Giving the Interesting Idea to build our time Machine with accounts of micro and macro variability.

Food for thought! When I thought about it, I got back from bathroom in a hurry, I sat down and started calculations on paper of what it will take to build such a computer which can operate on a machine learning algorithmic model which will take into account macro variability as well as micro variability in the datasets of all organisms in the earth and say 100 characteristics at the least, defining situation around them. Few attributes which can be accounted in that could be: pressure around them, heat around them, stress level, insulin level, liquid variability around them with respect to the situations (liquid or gas surrounding them and how it varies  etc etc).

Suppose we take in to account these factors for the last 100000 years and account half a second factor of temporal distribution for 8.7 million organisms each having say a billion population at a given point of time. Storing each of the 100 environmental factors around an organism with the help of a float variable, to build a robust dataset with enough sample points. 

We are looking at a dataset of size 2.09869506 * 10^33 bytes. With exa-computers replicating our brain and we moving into the phase of yottaflops super computing, we will need a combination of 10 million of yottaflop supercomputer computation efficiency to predict the next half second before it arrives for anyone. 

Probably the next 50 years or more will take us to the era where we can predict our own future as well as future of all organisms as a whole with good probabilities ! DAMN! NOW THAT’S SICK!

(If you disagree, let me know)

[saprakas]

# 12 Rails CMS/Web Builder App: Part2, sprint1

[saprakas]

Rails CMS/WebBuilder APP: Build Your own bootstrapped website in less than 10 minutes

[saprakas]

Rails Ad Engine App and API

[saprakas]

Rails Socket App: RealTime Auctioning with Faye gem

[saprakas]

#7 Rails App: Yelp Review Aggregator

[saprakas]

#6 Smile And Draw App: Socket.io

[saprakas]

Rails App: Trials and Triumph With AWS S3

[saprakas]

#5 Refactoring: Partials, Pain and Post

[saprakas]

#4 BDD and Integration Testing the last app: Pain & Triumph

[saprakas]

#3 Showcase: A Cool Clearance App in Ruby On Rails

[saprakas]

#2 Ruby On Rails Website From Scratch Live In Production

[saprakas]

#1 Single Page Website, UX and Logo Designs