"what DNA (or evolution) really wants" "Eventually, some pre-living ribosomes would have associated with bubbles of fatty acids" - as if they have wants and desires, or ability to act with intent.
I think evolution is a kind of alternative religion for many people.
Using that language with events that always happen doesn't lead to problems. But as soon as you use it with events that only happen sometimes you cause your thinking to be skewed.
There is also a difference with complexity level - using "want" for single step events is not the same as using it with multi-step complex events, that only as a final result "want" something. Doing that leads you to personify the system.
> The problem usually is people who don't know what they are talking about misinterpreting it as though atoms/molecules/cells/etc. have feelings.
Exactly. Which is why you should not do it.
I don't think scientists make that mistake, although some philosophers have argued that they (rarely) do. It is very likely that laypeople reading such articles might, though.
Yudkowsky's "An Alien God" talks about this: http://lesswrong.com/lw/kr/an_alien_god/
In many many areas, we use personification to easily abstract complex behavior and rulesets. It's a signal to the reader/listener that we are deferring discussion about exactly why or how something happens.
The oxygen "wants" to bind to the iron to create rust, because we're glossing over electron orbitals. The T-cell "thinks" it has found an intruder, because we're not tracing it's internal algorithm. The computer "guesses" when to retransmit a packet, because the details of exponential backoff are outside the scope of my explanation...
Most of the time your kind of descriptions of the "how" are not presented thus ppl. mis-conceptualize.
Sometimes I think it would be best science used its very own language partly like Latin in medicine is used.