Well, I suspect it was me. I used that very expression (near impossible) about HTTPS in an answer about hacking a few days ago. And that was an answer to a question you asked.
What I meant is quite simple: such securities are theoretically not just hard to hack. They are impossible to hack. Theoretically. But there could be a breach (you can't assume there is one. Hackers are not magicians who find breaches where there are none. That is why they usually don't target a single tech. That is why "hacking this" can be impossible, while "hacking something, whatever" is always possible: you just have to find a weak target)
So, lot of protocols that are supposed to be secure, happen to be not that secure. Because of bugs. Because of naiveté in their coding.
But it is very very unlikely that it can be the case of https. Because this thing is used everywhere. It has been (unsuccessfully) attacked more than anything. It has been checked constantly by literally thousand of experts.
So it is quite simple indeed: to find a breach in https, you have to find something that probably isn't even there, you have to find something that thousand of experts checked wasn't there, you have to find something that thousand, maybe millions, of hackers failed to find before you.
Where as to find a breach in the account management of CoolAndroidGame [fake company name. Sorry if it is a real company I am not aware of], you just find a breach that one tech guy at CoolAndroidGame says he has made impossible (probably being out of his field, since I doubt they hires cryptography expert), and that nobody really tried to hack before you did.
It would be simply less surprising that you find a breach somewhere, when nobody had already tried before you did, than finding a breach somewhere when zillions of people failed to find one before you did.
That is why, nowadays, CoolAndroidGame tends to not rely on home made security layer. They just rely on SSL, HTTPS, or other very improved and tried techs.
So my comment was just about that simple fact, that SSL, HTTPS, ... have far more man.hours of work to make them safe, and far more man.hours of unsuccessful attack attempts.
It was not about the algorithm itself. Algorithms that make decipher impossible are older than computers anyway. Not saying that there is no hard science in this (especially when you have to communicate with short keys, and with people with whom you haven't exchanged any keys prior to the communication). But that is generally not the weak part. It is plain old bugs that are the breaches, usually.
Of course, for https, you can try to convince user to bypass the security.
With a man in the middle attack (which is quite easy. My students do it in on single 3 hours session) you can easily read communication. But the user has to click "ignore" on the warning that the authentication key is not valid (which some users tend to do, because there are some legit websites that uses invalid keys, or self signed keys. So, people tend to click "ignore" without worrying and even reading).
You could also try to convince people to install your own authority in their browser. Which then allows you to sign fake keys that their browser will accept without a warning.
You could probably do harm with extensions.
You could bribe an authority (one of the companies that validate https certificates)
You could distribute you own altered version of firefox with your own authority embedded.
So, there are some way "around" https. But they rely on the ignorance or recklessness of the user. On social engineering generally speaking.
Maybe https itself has a breach. But if it has one, it is a hard to find one, that zillions of hackers failed to find so far.
So, it is easier to go for social engineering.