Would you prefer to work during 5-6 years at a low salary and be able to write your PhD or start working right after your Master in AI knowing that you may never have the possibility to create something on your own from tail to head ?

Who takes 5 - 6 years to write a PhD dissertation?! A Masters requires 2 years typically and a PhD 4. It's only an addition of 2 years. As for financial compensation, yes, it's generally going to be somewhat low and may remain so depending on how marketable you and your skill set are to the rest of the people in society. My suggestion is PhD... IF you think that's something YOU value. If it's not, then don't do it.

Correct, but its not unusual here that a PhD lasts a bit longer or even only 3 years... Internationalization is not completely done everywhere I guess Please Register or Log in to view the hidden image!

Took me five years to discover how to get a computer to divide by zero (something they cannot do (until now!)) :-D It was unpaid work but definitely worth the effort. I do not have a degree but my work with computers is done on a calculator. Early computers were so and my philosophy towards this work (and everything else) is based on the Berkofsky principle: persevering through EVERY possible combination will eventually provide the correct answer. It has to be one of them, right? My solar powered Casio is capable of dealing with X and provides results as so: X-F(X) 0- 1- 2- 3- With regards to the Berkofsky principle I concluded that ALL mathematics is done using ONLY four operators: "+*-/" (Plus Multiplication Minus Division) A fifth operator May include "=" (Equals) Conclusively the program I wrote was as such: A(B)=C+D*E-F/G 1<=B<=24 1<=C<=10 1<=D<=10 1<=E<=10 1<=F<=10 1<=G<=10 A simpler version may be achieved manually by assigning a number to each operator (1,2,3,4) and multiplying them: 1*2*3*4 which gives the return code of 24 (proving ALL combinations have been achieved.) Other version may include: +10*10*10*10 Currently I am writing REAL world programs by simply writing the program on paper (instead of using a computer!) Please Register or Log in to view the hidden image! For example: A+B*C-D/E 1<=A>=B>=C>=E<=10

Oh wait! Do you mean Bergofsky? The fictional (and entirely false) principle used in Dan Brown's Digital Fortress?

Your sense of humor made me laugh as I struggled solving the integrals of the Gibbs Sampling part of the LDA statistical model... This pic is just the part 19-20 on 64 (altogether) derivations steps.... and it's only one of the 3 most known proofs... Please Register or Log in to view the hidden image! When I'm tired of being a "Maths compiler" I write it in a computer program that does it for me [=. As a side note, if someone does not have a degree (for family, economical, sociological, political or whatever reasons) that person would be happy to know that there is some top notch learning material on the web atm and that the most well-known universities are starting to let "slip" some invaluably high quality courses for free in order to brand their name for the real (wealthy) targets. My advice for beginners is to start with Khan Academy (Precalculus + Calculus overview / LinAlgebra) -> Go on with Apostol Calculus 1-2 -> and then pass the MIT Introduction to Probability Theory and Statistics to solve the above... + A great bunch of sheer curiosity... But maybe you already tested that "path" when randomly choosing your operators ;] Have fun =D

A fun fact is to know that any Windows or Linux operating system account can be cracked by brute force in 30 seconds or less... And after that "Police IT Specialists" call themselves "clever"... Wait, I'm gonna create a State and make arrangements with all Computer companies to let me go into your personal life in order to "protect" you from digital harassment. But I'm making no sense right ?

Oh no. You are mistaken TEN. I simply perform my mathematics on ONE (1) rather than a computer. For example should I multiply the ONE (ME!) by one-hundred, and divide the result by one (whatever value that may presently be) divided by one-hundred, then all possible values should be reached should they not? In other words (actually values): 1*100 / 1 / 100 =100/0.01 =(100/0.01)*100 =10000/1 = 10000 All ten-thousand values are stored in values 1-10000, then then accessed with the division by one. This is the easiest way to count to a number, i.e. 10/1= 1,2,3,4,5,6,7,8,9,10 ...is it not? The most exciting values I have found are with regards to the family tree: Each one comes from two parents: forming three. My initial work regarded three divided by two, however I have found a much simpler version. Given that each one comes from two, and those two come from two (four in total) etc. we may trace the family tree and calculate the population of the world (or at least our position (or TIME)) within the world, with regards to our introduction to it as follows: 1 / 2 / 4 / 8 / 16 1 1+1 1+1+1+1 1+1+1+1+1+1+1+1 1+1+1+1+1+1+1+1+1+1+1+1+1+1+1+1 (or 1/1024) This traces one to grand-parents. Interestingly the population is greater in the past...! How can this be so???

Because humans come from earth and earth itself comes from two recursively sexed other planets ... that's where all that extra population came from. Logical right ? [=