Training Example: Speedy Paz Technologies Ltd. – Review the Data, Give Your Score & Compare to the Real AI Evaluation

[H1] Welcome to Speedysoft
Speedy Net [alpha] – A new social network.
Speedy Match [alpha] – A new dating/matching app for singles.
Speedy Composer – Artificial Neural Network Melody Composer.
Speedy Mail Software – A free software and open source web-based email software.
Speedy Net Blog (English / עברית) – Updates and articles about Speedy Net, Speedy Match, Speedy Composer, Open Source Software, Music, WordPress and Veganism.
Uri's blog – Some thoughts about life and the universe.

Speedy

I just came across Guido van Rossum’s King’s Day Speech from April 2016. I recommend reading it. Until yesterday I didn’t know that the G in Guido is pronounced like the letter ?’ in Hebrew, nor that the Netherlands has a king… But even today, there is nothing I know for certain, even the fact I exist is an assumption and not a known fact. I don’t even know if I have a computer. But anyway I recommend reading this speech.
By the way, if you don’t know who Guido is, you’re probably not an experienced Python programmer, so don’t apply for a job in Speedy Net…
* I’m sure if WordPress was written in Python, it wouldn’t convert the Hebrew letter ?’ to a question mark. That’s what happens when programmers don’t know how to program…

Websites which block users from specific countries by their IP address (like Pandora) or redirect users to local websites (like Google) or display different content to different users based on their IP address (Google) are practicing Ipism, which is like racism or speciesism should be illegal discrimination of users based on their IP address. All internet users should be equal and should not be discriminated by their IP address, location, sex, age, race, nationality etc. Ipism should be illegal and websites such as Pandora should be closed, Google should be forced to display the same content to all users of all IP addresses and nobody should discriminate users by their IP address. And by the way, Google’s Inbox being limited to Google Chrome only is practicing browserism, which is discriminating users by their browsers. This should be illegal too.
#ipism
#browserism
Ipism – Speedypedia [alpha] – in English
Ipism – Speedypedia [alpha] – in Hebrew

I checked my chess queens application for larger chess boards, and found some interesting results: It appears that when the minimal distance between queens is more than 1, the minimal chess board size with solutions (not including one queen on the trivial 1×1 board) is at least the square of the minimal distance between queens. Here are a few examples:
When the minimal distance between queens is 2, the minimal chess board size with solutions is 4×4 (2 solutions).
When the minimal distance between queens is 3, the minimal chess board size with solutions is 10×10 (4 solutions).
When the minimal distance between queens is 4, the minimal chess board size with solutions is 16×16 (2 solutions).
When the minimal distance between queens is 5, the minimal chess board size with solutions is 28×28 (10 solutions).
When the minimal distance between queens is 6, the minimal chess board size with solutions is 36×36 (2 solutions).
When the minimal distance between queens is 7, the minimal chess board size with solutions must be more than 50×50, because there are no solutions in any board size up to 50×50. When I created my chess queens application I limited the maximal board size to be 50×50, because I didn’t know that larger board sizes can return results in reasonable time. But I checked and found out recently that even boards with size 36×36 return results (when the minimal distance between queens is 6) and even the 50×50 board returns “no solutions” when the minimal distance between queens is 7. It would be interesting to increase the limit of my chess queens application to boards larger than 50×50 and find out the minimal chess board size with solutions when the minimal distance between queens is 7, 8 and more.
From my results above I would guess that with any even minimal distance between queens n, the minimal chess board size with solutions would probably be n2×n2 with 2 solutions, but this needs to be proved. It would also be nice if I can find out a formula for the minimal chess board size with solutions as an expression of the minimal distance between queens, and also a formula for the number of solutions. But it looks like it can be a big challenge to find such a formula and prove it.
The numbers 10 and 28 above also appear in Pascal’s triangle, and it would be interesting to see if the minimal chess board size with solutions when the minimal distance between queens is higher are also numbers from Pascal’s triangle. If they are, I would guess they might be 55, 91 and 136 respectively for the odd distances 7, 9 and 11. It’s just a guess, but it matches the minimal board sizes of 1×1, 10×10 and 28×28 above. If my assumption is true, then the formula for the minimal chess board size with solutions with odd minimal distance between queens n would be the Binomial number of (((3 * n) + 1) / 2) over 2 (which is equal to (((3 * n) + 1) * ((3 * n) – 1) / 8)), and the number of solutions on the minimal board size would probably also be a number from Pascal’s triangle.
I found out that when using rooks with a minimal distance between rooks n, the minimal chess board size with solutions is n2×n2 with 2 solutions, for all n>1 (odd or even). But this needs to be proved.

I started to learn Python a few days ago and wrote a few sample programs. I noticed that integers in Python are not limited in size, unlike many other programming languages like PHP, C, C++ and Java. So my first programs in Python were simple calculations of big numbers, for example big powers of 2 – I even calculated numbers with hundreds of thousands of digits. So I decided to use Python to calculate the mathematical constant e. I wrote a short program to calculate the first 5000 digits of e (actually that’s 5000 digits after the dot), and surprisingly I didn’t have to use any module, not even the math module. I only used integers to calculate e times 10 to the power of 5000. I searched Google for “the first 5000 digits of e” and found out a page with the first 2 million digits of the number e, which I viewed to check if my calculations were correct. I also found a page titled “How I computed e to 20000 digits“, where I found out it took the author more than 9 hours to calculate the first 20000 digits of e. It’s surprising, because my program calculates the first 5000 digits of e in less than a second, and also when calculating 20000 digits it takes about one or two seconds. I don’t know what caused the author’s program to take so much time, but I decided to post my e calculation program here:
digits = 5000
add = 500
f = 10**(digits + add)
e = 0
n = 0
while (f > 0):
# add current inverse factorial to e.
e += f
# calculate next inverse factorial.
n += 1
f /= n
# print e
e /= (10**add)
print e
print n
Notice that I used the variable “add” with the value of 500 – I actually calculated 5500 digits of e and then omitted the last 500 digits. This is because the calculation is not accurate, because I used integers and not exact numbers. When calculating 5000 digits without using “add” (or when add = 0), the last 4 digits are not correct. So I could use a much smaller number for add, but I decided to go for 500 digits just to be on the safe side. It took 1929 iterations to calculate the first 5000 digits of e, and the number of iterations grows with the number of digits – for example, it takes 13646 iterations to calculate the first 50000 digits of e.
I went on and wrote another Python program to calculate square roots, and used it to calculate the square root of 2 (although the same program can be used to calculate the square root of any positive integer, whether the square root is an integer or not). Here is the program to calculate the first 5000 digits of the square root of 2:
# calculate next square root of the number.
def calculate_next_square_root(square_root):
next_square_root = ((number / square_root) + square_root) / 2
return next_square_root
number = 2
digits = 5000
add = 500
number *= 10**((digits + add) * 2)
square_root = 1 * (10**(digits + add))
# calculate next square root of the number.
next_square_root = calculate_next_square_root(square_root)
n = 0
while (next_square_root != square_root):
# replace square root with next square root.
square_root = next_square_root
# calculate next square root of the number.
next_square_root = calculate_next_square_root(square_root)
n += 1
# print square_root
square_root /= (10**add)
print square_root
print n
Here I used a function – calculate_next_square_root(square_root), although it’s possible to write the same program without functions. It’s interesting to note that this program takes only 13 iterations to calculate the first 5000 digits (after the dot) of the square root of 2, while the e program took 1929 iterations. And when calculating the first 50000 digits, this program takes only 17 iterations. This is because the number of iterations it takes is proportional to the logarithm of the number of digits, because every new iteration doubles the number of correct digits of the square root calculated. So If we wanted to calculate the first google digits of the square root of 2, this program would take only about 332 iterations. The problem is that we will need enough memory to store numbers with google digits, which we don’t have and we don’t expect to have in the future. So we can’t use this program to calculate the first google digits of the square root of 2. But we can use it to calculate the first 100000 digits of the square root of 2 – I tried and it took 18 iterations.
Another thing – the square root program returns correct results even with add = 0. This is because the result is always the closest integer to the square root calculated (2 times 10 to the power of digits*2), rounded down. If the square root is an integer, the result will be accurate. If not, the result is the closest integer rounded down. I will not go into the mathematics to prove it, but this algorithm is accurate even without using “add”. But I left it at 500 digits just in case.

Listen to My Compositions on Speedy Composer. They were all composed for me by Speedy Composer. Especially I like Composition #11 and Composition #5. If you want you can also join Speedy Composer and let it compose melodies for you!

I recommend reading “A Sister’s Eulogy for Steve Jobs“. When I read it again today I had tears in my eyes and I almost cried.

Mark Zuckerberg: I heard that you killed animals for meat. Why don’t you become vegan? Killing animals is cruel and immoral. It’s much better for your health, Earth and animals to be vegan!
I recommend taking the vegan pledge on http://evolvecampaigns.org.uk/ – you receive free ebooks that helps you become vegan. Especially I recommend the book “Street Smart Vegan” – it’s very good.

Check out my chess queens application: http://chess-queens.sourceforge.net/
There are 14,200 ways to place 12 queens on a 12×12 chess board, and Google Chrome is the fastest browser to calculate it. If you check the number of ways to place 16 queens or rooks on a 16×16 chess board, with a minimum distance of 4 – there are only 2 ways to do this. Each of them is symmetric.
There are 92 ways to place 8 queens on a 8×8 chess board, without any queen attacking each other. The complete list of solutions is found on Wikipedia.

Yesterday we went to see what’s left from the forests of the Carmel. We saw many green trees, but also trees burned completely to the roots and partially burned trees. There were beautiful flowers, only about one week old, growing from the ground. There is smell of ashes all over the place. We wanted to see Beit Oren, but it’s not open for visitors. But we saw some of the burned houses. On the one hand it’s sad, but on the other hand I think the forest will grow again, and we should let nature do it naturally.

How many roots has Google? I mean integer roots. Some numbers have roots, some not. I think Google has many roots probably. he’s a big number. Probably as many roots as the number One Hundred has divisors. Don’t you think?
Actually One Hundred is Ten times Ten which means primes Two and Five twice. He should have Nine divisors since Three time Three is Nine.
(Three options: each prime should appear either Zero or One or Two times).
Let me check….
One is a divisor –> Ten is a root. [TenOne Hundred is Google]
Two is a divisor –> One Hundred is a root. [One HundredFifty is Google]
Four is a divisor –> Ten Thousand is a root. [Ten ThousandTwenty Five is Google]
Five is a divisor –> One Hundred Thousand is a root. [One Hundred ThousandTwenty is Google]
Ten is a divisor –> Ten Billion is a root. [Ten BillionTen is Google]
Twenty is a divisor –> One Hundred Million Trillion is a root. [One Hundred Million TrillionFive is Google]
Twenty Five is a divisor –> Ten Trillion Trillion is a root. [Ten Trillion TrillionFour is Google]
Fifty is a divisor –> One Hundred Trillion Trillion Trillion Trillion is a root. [One Hundred Trillion Trillion Trillion TrillionTwo is Google]
Oh wait… is One Hundred a divisor of itself? Well if so, Google’s probably a root of itself too. But if not including Google itself, Google has Eight different roots if I’m correct. But how many roots does Googleplex have?

Well of course Googleplex (like any exponent of Ten) actually has only Two prime divisors – Two and Five. Googleplex is actually Google times Two multiplied by Google times Five. So it seems to me he has (Google plus One) times (Google plus One) roots minus One, not including himself. Actually all roots of Google are probably roots of Googleplex too. Since Google is probably a root of Googleplex too. Is he? Let me think. I think he is. One Hundred is a divisor of Google so he is probably a root of Googleplex too. But not all divisors of Google are also Googleplex’s roots. I think only those who have the same number of Two and Five as prime divisors – only those who are exponents of Ten. But not all of them – only those who divide Google – for example One Thousand is not. One Thousand is Three times Ten (TenThree), but Three is not a divisor of Google. Only those divisors of Google who are exponents of Ten and their exponent of Ten is a divisor of Google – They are real Googleplex roots.
So lets name them (do they already have names? maybe. but lets name them again):
Ten
One Hundred // TenTwo
Ten Thousand // TenFour
One Hundred Thousand // TenFive
Ten Billion // TenTen
One Hundred Million Trillion // TenTwenty
Gillion // Ten Trillion Trillion // TenTwenty Five
<!– all *illion are small numbers; *oo?le are big –>
Goodle // Gillion Gillion // TenFifty
Google // Goodle Goodle // TenOne Hundred // It’s a lovely name – don’t let them give him a bad name…
Doogle // Google Google // TenTwo Hundred
Toogle // TenTwo Hundred and Fifty
Boodle // TenFive Hundred
Noodle // TenOne Thousand
Noogle // TenTwo Thousand
Nooble // TenTwo Thousand Five Hundred
Nootle // TenFive Thousand
Tootle // TenTen Thousand
Toople // TenTwenty Thousand
Tooshle // TenTwenty Five Thousand
Toorle // TenFifty Thousand
Toorrle // TenOne Hundred Thousand
Toorrrle // TenTwo Hundred Thousand
Toorrrrle // TenFive Hundred Thousand
Tooggle // TenOne Million
Toogggle // TenTwo Million
Tooggggle // TenFive Million
Toottle // TenTen Million
Toottt