everything encrypted today can be considered cracked if it is not time sensitive.
Yes, true for all asymmetric cryptography, which can be broken by a QC running Shor's algorithm. Symmetric cryptography on the other hand is far less vulnerable, AES256 for example is quantum-proof against the best attack (Grover).
To break bitcoin's ECDSA it would take 2
128 operations on a "normal" computer to derive a private key, whereas for a quantum computer running Shor that drops to a much more manageable 128
3.
But for symmetric cryptography, the attack vector has to be different, and the exponent only drops by 1/2, so something that would take 2
128 operations on a normal computer still takes 2
64 on a QC. And if we do move up to AES256, that's still 2
128 on a QC... the same number of operations to break ECDSA classically right now.
IBM already has 53 qubit machines. Once they hit 128, goodbye normal SSL. :/
Whilst IBM have certainly made some impressive advancements, I am somewhat skeptical and believe that their achievements have been overstated. Media articles tend to be full of breathless hyperbole and are overly simplistic, as if 'number of qubits' is all there is to it, when clearly this is just a headline figure. I'll believe that IBM have achieved something truly special once they can prove that they have robust error correction.
53 qubits is not the same as 53 fault-tolerant qubits. Decoherence is a huge obstacle in quantum computing, and if IBM are leading people to believe that simply throwing more qubits at it will solve everything, then they are being at best very disingenuous.