Yes and No.
Yes, everything increases in difficulty but the increases in difficulty are asymmetrical.
The difficulty of reversing a computation (e.g. reversing a hash or decrypting an encrypted message) grows much faster than just performing the computation (e.g. hashing a message or encrypting one).
That’s the basis for encryption to begin with.
It’s also why increasing the size of the problem (e.g. the size of the hash or the size of a private key) makes it harder to crack.
The threat posed by quantum computing is that it might be feasible to reverse much larger computations than it previously was. The caveat on that, however is that they have a hard limit of what problems they can solve based on the number of qbits they have.
So for example, let’s say you use RSA for encryption and someone builds a 1024 qbit quantum computer. All you have to do is increase your key size so that it would require 1025 qbits to crack, and then that quantum computer wouldn’t provide an attacker any benefit at all.
(Of course, they’d still be able to read your old messages, but that’s also a fundamental principle of cryptography; it only protects you for a period of time)



Obviously, this nonsense is meant to try to prop up their stock price until they have an opportunity to cash out.
However, even if it was true it would mean their company is over valued.
They don’t have any sort of “moat”, so if AI models were truly that good then we could use them to replace Anthropic and render their entire company worthless.