That hasn't been a consensus on that for a good thirty years at least.

Of course people always look for it, since if it does actually work that way it's easy to understand so finding the cases where it does is particularly rewarding. But ever since the early days of QTL mapping and especially since GWAS (> 15 years old now), not only has it not been the consensus, but rather the consensus has been more that it often isn't this way.

However, this doesn't mean that there's no mechanism, or that the instructions in the genome aren't absolutely central. It's just that while billions of years of "whatever works" produces things that work, it doesn't always converge on the same style of answer that we like to have when designing an airplane. That there are a forbiddingly large number of almost indiscernibly small effects, in some cases, doesn't mean that anyone's world is "about to crumble".

It's just like with cancer. People hoped cancer was simple. Find the right drug--cure cancer! But it's not simple. Once you grapple with the horrific complexity while still accepting that cancer is a fully mechanistic biological disease you can make progress. If it were actually incomprehensible complexity, then it would be hopeless. But so far, a lot still remains comprehensible, as long as we take things as they are and not with the degree of simplicity that we wish there would be.

Finally, I will happily drop sarcasm if you drop confident proclamations about the self-delusion of entire fields of sciences ("The youngish scientists are soooo sure it's not this way[; ...] they do not know that their worlds are about to crumble.").