Good questions, Margarita.
There are two kinds of cool young stars. One kind is the small, low-mass young star. Such low-mass stars are born in the Rho Ophiuchi region. In the upper right quadrant of today's APOD there is yellow-orange light penetrating the thick dust. At least one source (near a particularly dark cloud) is undoubtedly the light of (low-mass) star formation. (Small yellow-orange stellar objects are scattered throughout most of the picture, although many of these sources might easily be background objects strongly reddened by dust.)
Low mass stars, M- and K-type dwarfs, don't create an appreciable amount of dust, as far as I know. But medium-mass stars die moderately young (say, at the age of 6-7 billion years) as Mira variables
, and they produce a lot of dust as they die. Even younger, more massive stars, that started out as blue main sequence stars but have evolved into cool red giants or supergiants, produce a lot of dust towards the end of their lives. (On the other hand, still very hot evolved young stars also produce a lot of dust at the end of their lives
.) Note that today's caption talks about "large, cool, young stars", so it clearly refers to evolved giants.
Perhaps today's caption refers to the fact that dust is scattered when a star dies violently in a supernova explosion. But even stars that die more gently
to produce planetary nebulas create and scatter dust.
Note that supernovas themselves not only scatter but apparently create dust:
"At day 868, the last time Gall’s team observed the supernova, the amount of dust had increased to 0.04 of the Sun’s mass, or 830 Earth masses,” says Cowen at Nature Magazine. "If the increased dust production continues, in 20 years SN 2010jl will have produced the equivalent of half the Sun’s mass in dust particles, similar to the amount observed in the widely observed supernova SN 1987A.”
This means that if there were a number of supernovae active around the time our Universe was still young, and they were producing star dust at the same rate as SN 2010jl and supernova SN 1987A, they could certainly have provided all the dust we know was present in the formation of our Universe.
As for the confusion of what is a pre-main sequence star and what is an evolved giant, no star can, to the best of my understanding, be a pre-main sequence star and an evolved giant at the same time. A star whose classification is B2III/IV is an evolved, aging giant and can't possibly be a stellar fledgling that has barely begun its life as a star.
However, a blue giant star which is old enough to have exhausted its core hydrogen has one thing in common with a pre-main sequence hot star. Both are larger than a main sequence star of the same spectral classification. Therefore a pre-main sequence star may be misclassified as giant star. An example is NGC 6727 and new-born, possibly still pre-main sequence star HD 176386. You can see it here (at 6 o'clock) with another star totally engulfed in blue nebulosity
. At the upper right of the binary baby star there is some very dark dust and low-mass star formation. (Here is a large picture of the entire area of star formation around HD 176386.
It is clear that HD 176386 has either just entered the main sequence, or else it is still a pre-main sequence star. But my software Guide classifies it as a B9 IV star. Obviously this classification is wrong. So I'd say that HD 147889 (SAO 184376), just like HD 176386, is a pre-main sequence star which is larger than a main sequence star (precisely because it hasn't yet entered the main sequence) and therefore has been misclassified as a giant star.
A pre-main sequence star can certainly have reached a temperature corresponding to a B2-type star. As for HD 147889 (SAO 184376) it is deeply buried in dust in such a way that it most certainly looks as if the star is in the process of being born in there. A pre-main sequence classification seems very reasonable to me. The star is deeply reddened, but it could well be quite hot.