LOL at the notion of RNG's being "cracked" so easily.
As a software engineer myself, allow me to provide some technical insight and dispel some misinformation as to how modern RNG's work, hopefully at a level a layman can understand. A basic, out of the box RNG such as that supplied with most modern programming language libraries (such as the standard C runtime, .NET library, Java JRE, etc) typically is seeded by an application before use. A seed is a hopefully non-deterministic (i.e. not easily repeatable or predictable value, in fact a pseudo-random number itself) that "kicks off" the RNG and gives it something to start calculating from. Even basic programmers will take something pretty unpredictable like the machine uptime in seconds or milliseconds, which is *technically* exploitable but incredibly difficult to do so, because you'd have to determine the exact same seed value down the second or millisecond (and this is relative to when the computer was booted, not the absolute date/time). And due to the way RNG algorithms work, you have to be exact, there is no being "close." As a simplified example, seed values 1111111 and 1111112 will produce wildly different RNG results, so if you don't get it exactly right you might as well be a mile away.
Now any serious RNG algorithm, certainly one where millions of dollars and the integrity of the company behind it is at stake, is going to take this several steps further than grabbing a single value. They take multiple non-deterministic values, possibly timestamps but also various entropic (chaotic) values that cannot possibly be predicted, such as mouse movements, network latency timings, etc. They hash (mix in) those values to come up with what is for all practical purposes, a purely non-deterministic value. Combine that with the fact that many of these applications write their own RNG algorithms rather than using the "out of the box" one that comes with their development library, and those are typically trade secrets that nobody else is privy to. Also, many re-seed themselves at random intervals as well, so they're never running forever on the same seed.
All of these layers within layers of RNG initialization and calculation virtually guarantee that any serious RNG is not practically "crackable," at least not without a lot of inside knowledge.
Now, there have been poker site cheating scandals in the early days of poker, and most recently of note, with the UB/AP scandals, where hole cards were known to players. However these exploits had nothing to do with RNG prediction. Early site network protocols actually sent hole card info across the wire to every client, which made it trivial to exploit once that was discovered. Of course that was quickly remedied and now you only receive your own cards and those that are shown down. In the UB/AP situation, several players (employees or former employees) had access to a backdoor login to the site's servers that allowed them to monitor hole cards on the server itself. This backdoor was unknown to the new owners of UB/AP at the time, and once they realized it, they closed this down as well.
So anybody telling you they've "cracked" the RNGs of even one poker site, let alone "all of them," are unequivocally SCAMMERS preying on the greed and technical ignorance of players who will do anything to cheat.