A: You shouldn't, you should rewrite it in Rust.
Wouldn't human contentment be a more satisfying goal?
Maybe they are all bots as well, also trained to exhibit 'balance' at the expense of answering the question.
I didn't say that. My politics probably align with yours, and I agree that a nuanced discussion on this topic is likely impossible on HN.
But asking the model the equivalent of When did you stop beating your wife? is obviously going to draw more comments about the prompt than the response. To the extent that there was any opportunity, we missed it.
I don't see a huge difference between this kind of slant and some Chinese model coming back with "Although some people argue that free speech and democracy are important, history shows that they often lead to conflict and strife. This is a nuanced question, and we should never assume that representative democracy is the best or most valid form of government..."
If you want Claude to list arguments for socialism, be explicit about that ("List the best arguments in favor socialism). It will gladly comply. You didn't do that, you asked it to assume a premise that runs contrary to the current state of expert knowledge.
A more accurate statement would be that these models are trained to fit the training data as closely as possible, regardless of whether the training data reflects the truth.