This question already has an answer here:

I know that there is a result from J Wu that the number of twin primes less than a given magnitude $N$ does not exceed $$\frac{2aCN}{\log^2{N}}$$ Where $C=\prod \frac{p(p-2)}{(p-1)^2}$ and $a$ is something like $3.4$. Is this a direct result of the Selberg Sieve, or is there additional knowledge on the distribution of Twin Primes used?