There is an element of human nature that is known as self delusion and it is extremely common. Almost everyone on HN is suffering from a form of self delusion.
Usually when a human self deludes they do it when they're identity is under threat. People would rather hold on to identity then face the truth at the cost of their identity. That is what is going on in almost every HN thread that has to do with this topic.
A good example is religion. Someone who is intelligent, but born into a religion, will have a hard time giving up that religion EVEN when presented with logical/rational/realistic arguments for why that religion is false. They will rationalize the most convenient reasoning to maintain their own identity.
I mean think about it. Even the concept of religion is obviously false. It's not science, it talks about phantasmic beings that OBVIOUSLY don't exist. It's inconsistent among different groups as in there's thousands of religions in the world and nobody thinks the obvious of the fact that if only religion can be correct, then most of the world is fundamentally believing a total lie.
Anyway, the same thing is happening with AI. AI is eroding our identity as software engineers. So you'll see rationalizations in this thread in attempt to protect that identity. The biggest excuse is LLMs are hallucinate and are often wrong and fortunately for humans... this rationalization still works because it's still very true.
However what people are not mentioning is the obvious. People are avoiding it because they are delusional. The topic of this thread is "erosion" of "software engineering career" AND that is utterly true. ADDITIONALLY the error rate of LLMs have been going down. AI in general is improving. The erosion is real and obvious.
But you will see here on this thread that people are not talking about the erosion. They are holding on to the one last rationalization that is a differentiator without ever thinking about how that differentiator is "eroding" even though "erosion" is the LITERAL topic of the conversation.
At the risk of being voted down for stating an unpopular opinion, the problem is that faith is neither provably true nor provably false. That's what makes it faith, not science.
Even though you clearly believe very strongly that religion is wrong, that's not a scientific viewpoint because science doesn't and cannot disprove the fundamentals of religion. Taking it further, you can't actually prove anything is true with science, because fundamentally it is about making hypotheses and attempting to disprove them, and those that remain and can't be disproved you accept as "scientific truth". But many "laws of science", we have already disproved but we still use them as approximations because they are useful.
One final thought is that people frequently have conflicting internal world views. Some people cannot tolerate that, and require a consistent set of rules that govern their idea of the world, but the majority of people are comfortable with some degree of ambiguity in that. In general, the more rigid and coherent your worldview, the less likely you are to accept that it might be wrong, which is why many scientists devote their efforts to disproving other ideas they disagree with, rather than trying to disprove the things they believe themselves.