Hebrews 2.8 contains the first bracketed text in Hebrews in NA27 - an indication that although the committee decided on AUTW as the text to print, they were sufficiently doubtful as to signal their doubts with single square brackets. This isn't the same as split decisions where the committee was not unanimous (these are not routinely marked and can only be discovered by reading Metzger's commentary carefully). This is a committee decision to codify their own doubts within the printed text (this is defcon 4 on the the scale of committee uncertainty [defcon 3 is to signal that NA25 had a different reading; defcon 2 is to provide evidence for and against the text reading; defcon 1 is to note that a variant exists - these are NA27's subtle indications of degree of doubt and well worth being familiar with]).
The text and apparatus presents this as a choice between the presence and absence of AUTW. The apparatus puts the witness of P46 in brackets. This means that the editors regard P46 as a "kind-of" witness to the absence of AUTW, but the brackets signal that things are not as simple as that. The editors of NA27, being careful about such matters, knew that readers would be interested in what P46 actually does read here, so they have a cool appendix where (all?) all the bracketed witnesses are explained. It is the place to go when you wonder what the complication is with P46 here. It is appendix II (pp. 721-749) and is a very valuable collection and on p. 745 re Heb 2.8 we discover that P46 lacks not only AUTW but the whole phrase AUTW TA PANTA.
Armed with this extra information we might notice that NA27 here over-simplifies the textual situation as a simple choice whether to read AUTW or not. In fact P46 offers a third reading altogether. Occasionally when you notice this in NA27 (it is a characteristic of NA27 which is the negative side of the editorial decision to include as much information as they possibly could in a compact and abbreviated manner - it is not something to complain about, but to be aware of) it is worth checking what else may be happening in the passage. A quick look at Tischendorf reveals the interesting information that D E syr and cop have TA PANTA before UPOTAXAI. This gives a fourth variant:
1. UPOTAXAI (P46)
2. UPOTAXAI TA PANTA (B etc.)
3. TA PANTA UPOTAXAI AUTW (D E etc.)
4. UPOTAXAI AUTW TA PANTA (01 A etc.)
Once you write them down like this you can see that there are more issues here than the presence and absence of AUTW, but also about the presence and placement of the (arguably redundant) anaphoric TA PANTA. Incidentally in some cases like this it can be helpful to check the apparatus of the UBS 4 edition (although not on this occasion, since here it is worse that NA27).
I'm not going to attempt to solve the general textual problem here, more illustrate how to get the most out of the NA27 edition while being aware of its limitations, but it is worth noting that even with the extra bracketed information about P46 we haven't yet got all the relevant information about how this verse reads in P46. Along with the shorter reading here, there is a different placement for the second AUTW and the lack of the article before PANTA (when it does occur later on). P46 has its own, rather elegant, way to present the thought of this passage. Interesting.