I think the difference here is between a consensus about a scientific question and a consensus about a social question.
It's not though. Consensus is about facts not questions. In the case of "when life begins" it's just a definitional fact. In the case of AGW it's actually two linked facts, one of which is data driven (1) Earth is warming, and the other of which is "established" by process of elimination (2) people are the primary cause.
Earth is Warming. I guaranty we don't know Earth's current temperature, not on the surface, not in the atmosphere generally, not from the surface to the edge of space, and certainly not in the entire volume thereof. What I can't know is how accurate our estimates are and what the real margin of error is (the calculated margin is nonsense).
I guaranty we don't know what Earth's temperature was in any of the those volumes as a historical matter (and some are completely impossible to have measured given the way we indirectly measure historical temperatures). What I can't know is the margin of error on the estimates, or whether the indirect measurements we do are less accurate than the direct ones we are attempting now, or even if they are more accurate. It's literally possible that the indirect historical measures actually encode more relevant data than the real time measures, of course it's literally possible they're completely noise.
I guaranty, the way we have built the models using micro data research is heavily biased towards that which we easily understand, can get too or have researched, which is a heavy bias. What I can't know is what else is out there, or if we by random chance have pieces of it correct.
Putting that aside, let's assume that we are in fact warming (and there are plenty of people that make reasonable cases this is not the case, and others who make nutty cases), then that leads to the question of what's causing it.
Caused by People. That's our second "consensus" item. It's caused by people. This by the way is almost completely reliant on a single correlation. As the human population's technology and pollution have increased the "measurements" we are able to do appear to have increased. That's it. There's no experiment that confirms it. So how do we get from correlation to causation? Well honestly we can't.
However, we can approach the limit by eliminating other potential causes (again assuming we are even correct about the temperature). As with any purely observational Science, we become more certain of a theory by eliminating other potential causes as a source. And that's the whole lot of what is going on. If the correlation breaks the theory falls apart (which is why things like a "pause," a medieval warm period, and historically far higher carbon concentrations at lower temperatures are so damaging that they have to explained away).
Modelling though? Total red herring, and literally not a science experiment. It generates zero data. It runs zero experiments. It settles zero factual questions. What it may be good for is generating hypothesis, and if we get tuned enough correctly making predictions. I personally don't believe it's close to there yet, but reasonable people can disagree.
Why is it dangerous to treat this "consensus" as meaningful? Well mostly because we only know of a correlation and have not shown a causative effect. That means that contrary evidence will either "break" the model completely, or lead to "narrative explanations" of things that altered the course (whether these are actual refinements, or literally just covering for a broken hypothesis is in the eye of the beholder). If it's just a false correlation, and we implement ridiculous and damaging policies, they will (a) do no good, and (b) be credited with any correlative change.
So if we do nothing, and nothing happens. Proof it was a mere correlation. If on the other hand we completely rearrange the world, destroy all economies and kill have the population, and nothing happens. Proof we "saved" the world.
On the flip, if there is a causation and we do nothing it gets worse. But if we "do something" and it's the wrong thing it will also get worse and maybe much worse (which is why I'm so opposed to international treaties that punish the most efficient producers).
But let's not kid ourselves. Consensus on AGW is literally consensus on a hypothesis, and literally one that's incapable of being tested.
This all shows a basic misunderstanding of what is meant by "consensus". Yes, consensus is not part of the scientific method. It is, however, a goal of the scientific method.
I think others have addressed this, but there's no truth to this statement.
What it is not, and what critics often mistake it for, is some kind of end-state after which dissent is no longer welcomed. It's a handy strawman often used as a crutch by dissenters.
Is it? What are you using it for? It appears to be to try and bolster a hypothesis that can not be tested by asserting only a crazy person wouldn't be convinced because it's a "consensus" of Scientists. Again, most of the scientists involved are now statisticians not experimental. GIGO applies to everything they do, though I'm not asserting that what they are putting in is garbage, it's just a fact that they are running logical models that can not generate a new result. If the actual environmental conditions are not an option of the model then the computer can not output them.
This strawman is often used to argue against taking action based on a given scientific consensus - for example, climate science - because dissenting research is no longer "allowed" due to the consensus. However, that is not the case. Of course, findings that go against large amounts of previous research will face headwinds in changing conclusions - which is as it should be, for obvious reasons.
And your strawman is often used to avoid explaining why implementing policies that would harm the environment is something we must do on an emergency basis "for the environment."
It's not a strawman to acknowledge that climate science is not experimental.
The primary driver of the "consensus" strawman, however, is the "democratization" brought on by the internet, where people can find support for just about any pet theory, and can simply avoid dealing with contrary views. Flat Earth theory, anyone?
The primary driver for the "consensus" strawman is political motivation to try and tag unpopular and highly damaging policies as "above legitimate question." It's been noted above, you don't find the same "consensus" hammer arguments being used where they aren't being used as an appeal to authority in a political debate.