Computational or mechanical analogies are not totally wrong but they're dangerously over-reductive.
I never used to understand this position, until I saw right wing snowflakes getting triggered over a woman looking too manly at the Olympics. They confidently explained XX and XY to whomever would read their little meltdowns, and even rode back on their usual line that genitals determine gender - suddenly they care about hormone levels.
So I'm having to come around to the idea that over-reductive can be a real thing. Although I'm not sure that the sort of dumdum that dive off the deep end with these things would do any better.
A biology professor of mine put it this way: pretend you are reverse engineering alien technology. The stuff you are trying to understand was not designed by human minds. It was designed by evolution over billions of years, literally an alien intelligence that doesn't "think" anything like you do.
A big difference between biology and human engineering is that humans like one-part-one-function and linear chains of cause and effect. Biology isn't like that at all. It's an analog causal matrix where everything affects everything else to varying degrees and most parts have multiple simultaneous functions.
With biology, the code is running against nature, so the instruction set includes assumptions like "but if these two get close enough, then this other reaction occurs first". I suspect it can get so complicated that the only way to find out what it does is to run it.