I don't think it's realistic to think that human judgment can leave the picture after the first deploy. New systems are fragile, requirements change, and mistakes happen. A certain level of trust is required in the founders of a project. And generally, they have long-term incentives that are aligned with your own.

So it implies the crypto maximalist takes on true trustlessness are not proving workable. At least in practice with lots of real users.

The systems can be trustless and upgradable.

As we get proven/tested/formally verified code that implements governance protocols we can use that code to build system of allowing upgrades while allowing those who disagree to exit the system.

For example, one of the ZKRollup chains has a mechanism for upgrading that requires a small number of approvers for the “break glass” scenario. When an upgrade is approved, anyone can exit the network before the upgrade takes affect and move to a system with rules they support.

Governance models will evolve as we figure out what works and what doesn’t. Some people may be ok with a super majority being able to revert the state of a contract in response to a bug while others will want immutability at all costs.

Blockchain/cryptocurrency is about each participant deciding how much trust they are ok with, and having an open ecosystem where assets can be moved to systems that people want to use.

Do you have any examples of these takes? The concept of "true trustlessness" is obviously lacking in nuance and probably impossible. Nothing in the world can exist without some form of trust. However, we can certainly use smart contracts to decentralize trust and limit risk exposure.