The capability to make a decentralised system is there, but not one that can't fail. The problem is that Web3 is a system where - unlike normal examples of decentralised systems - there is an incentive for the system to fail.

If you linked together say, a hundred temperature sensors in your house, that's a decentralised system that *could* fail but is unlikely to (and has no motive to).

Web3 on the other hand - subverting consensus is very lucrative, there is every reason to do it. Decentralised systems are very weak against this. As such this isn't really a problem of machines or algorithms but a problem of making a system where the failsafe (humans mining with their machines) has a conflict of interests.

I don't doubt the professor's knowledge, but I doubt his benevolence. Making a decentralised blockchain secure isn't possible unless you can secure consensus, and short of mass mind control, that can't be done. On the other hand, there is a great deal of money to be made from these systems.