Whenever I read articles from "security experts", I cringe over how they keep relying on probability and computational complexity rather than proven methods that cannot be broken given infinite time and resources. Server admins keep SSH open to the internet just because they might need it when working from home some time, didn't have time to write a custom protocol or just hope that nobody borrowing their computer will steal their password-less SSH key, yet everyone's surprised when hackers found yet another backdoor into the system.
IoT is just a plague of idIoT devices where anyone can get root access to security cameras and such, by just trying a few common default passwords. Most chat applications boast about how hard their security is, yet in a few seconds I can log into my account from a different device without even being asked for my password. Even accidentally logged into my bank without being asked for my password (due to a bug that they have now patched). I keep my payment card in aluminium foil because turning off NFC payments don't stop skimmers from making a full copy remotely before making a regular purchase.
So much focus on authentication rather than privilege, minimalism and attack vectors
Why don't we have dedicated hardware with proven correctness for common server protocols to prevent instant erasure of file history? Even if someone gains access, there should be limits to what an administrator can do remotely, so that hackers cannot cover their tracks, encrypt the content or overwrite the server's operating system. If hardware needs patching, just use an FPGA.
We had unbeakable encryptions for many years, so no need to wait for mainstream quantum connections
Why don't companies give their employees unique read-protected microchips with unbreakable single-use 256GB true random symmetrical keys (that can be flashed at the office when consumed) instead of weak passwords (less than 120 random characters) or easily stolen SSH keys (from which enough quantum computing power can reveal the passphrase)?