tl;dr: "free will".

There's this picture floating around the net these days:

It's presented in the positive light, like replacing social consensus with technical consensus is a worthy goal, something to aspire to. I think it's wrong, at least on large scale and in areas of life where anyone gives a damn or gets excited about, and for two reasons.

People can (and will) decide which rules they follow

Even the greatest totalitarian regimes attempted so far did not really succeed at preventing human beings from thinking differently and disagreeing. In contrast to machines which literally must do what they are created (or programmed) to do, we always have the choice of disobeying. There will always be the "resistance movement" - it might be small, it might be suppressed, but it will always exist. If a system is deemed to be too rigid, it will simply be abandoned.

A case might be made that a kind of "technical consensus" could be implemented in a very specific area, where the vast majority will just "go along", mostly to enjoy more convenience or to avoid difficulty (which are often one and the same motivation). Such an area could be "the monetary system" and this is what cryptocurrency enthusiasts mean when they talk about "social consensus" being bad, i.e. susceptible to crime, corruption, cronyism, etc., but when you come down to it, just replacing banks with cryptocurrency does absolutely nothing to direct or govern how the money might be used. To do that, ideas such as the DAO came to be, with the goal of (among other things) making a technical solution which deals with exactly how to direct the flow of money through an organisation, and to decide which goals the organisation should pursue. An implicit part of such a system of technological governance is a kind of a voting mechanism, where the stakeholders would come to an agreement about what the organisation as a whole will do. But, unsurprisingly for anyone but the die-hard under-the-rock IT geeks, processes such as inter-personal communication, understanding and empathy, as well as persuasion, bribery and blackmail will always happen off-chain.

I've been a part of an organisation where we voted for what the organisation will do next, and I can tell you that being on the side which won by 51% does exactly nothing towards actually implementing the course of action. Even a much smaller minority (Nassim Taleb says it's around 3%-4%, and he knows his numbers) if not handled with gloves can de-rail a large organisation.

This is why social problems cannot be solved by adding another level of technical indirection between people.

Real life is hard and people actually like to be creative

If a truly all-encompassing technical system which influences human behaviour on a large scale is ever developed, it will also need to be extraordinarily malleable, changeable and adaptable to new circumstances and new requirements - or it will be abandoned.

At one extreme, we're talking about ethical dilemmas which may arise through the use of such a system, easily illustrated by the "trolley problem". If such a system gets into the position to decide something really important, and by "important" I mean emotionally important not logically, and does the wrong thing, it will be boo-ed into oblivion and abandoned. There was a highly publicised legal case around where I live a couple of weeks ago where a high court ordered, with a completely sound legal reasoning, completely logically within the framework of laws and practices, that custody of a certain child should be taken away from its mother and given to the father. Despite the legal consensus being completely clear, the public opinion blew up and people's support for the government and the confidence in the legal system took a sharp dive. Now imagine a DAO, a computer program, announcing its decision, backed by completely logical, traceable, and predictable decision to not transfer some sum of money to a family with a sick child or something, and the child dies (maybe the father didn't hit the targets set for his job in the DAO? Doesn't matter, there could be many reasons). The first time it does it, it will be boo-ed and criticised. The third time it does it, someone will rise and create a parallel currency, "with a human face", even if it means using rocks, seashells or nuggets of gold as the method of measuring value, and returning the decision back into the fallible hands of human beings.

Because that's exactly the point: the "bad" points of "social consensus" compared to "technical consensus" can easily be summed up in a single word, "adaptability," and adaptability exists for both good and evil.