First, I think we must accept that right now we do not have the ability to reach consensus. We lack Trusted Leaders and we lack Chairs. Without these we are afloat in our various pods of bias, preconception, personal history, etc. Not only can we not reach consensus, we don’t even have the ability to bring the various factions of bitcoin developer mindshare together to find out if there are technical objections to any particular change. Consensus changing code is rarely proposed, rarely reviewed, etc.
On one hand Trusted Leaders and Chairs can certainly be helpful processing consensus changes. On the other hand, with time they become single-point-of-coercion, cf. future risks like Watch: How Bitcoin is Fueling the Climate Crisis and a Software Change Could Clean It Up - Greenpeace USA On this “coercion” risk concern, I think Jeremy Rubin had some interesting thoughts in the past.
Getting “chaired” consensus change was my attempt with the Bitcoin Contracting Primitives WG:
If someone wants to keep it further, I don’t have time to maintain it anymore. (I must say it was maybe a bit too much ambitious in scope).
Technical consensus has also been found on many bad ideas, or ideas that were not compelling. Mainly that happens by an idea simply not gaining traction. Examples from Optech newsletters since taproot activation: two-digest BOLT11 invoices (#256), relay of annexes before a defined need (#255), LN QoS bit (#239), global LN reputation tokens (#228), (at least some versions of) LN > capacity-dependent feerates (#219), perpetual subsidy (#209), and fee accounts (#182).
You have weak and strong technical consensus. Let’s all remember OP_EVAL: OP_EVAL doesn't stop recursion · Issue #729 · bitcoin/bitcoin · GitHub
For now, I think it’s good to let the organic evolution of consensus changes moving out of Core / BIPs. Things we’re seeing with bitcoin-inquisition / bananas.
On review of proposal when they’re mature, the 2019 taproot review sessions was good practice. Note best reviewers in this space have always hands full with maintaining current stuff alive.