user959 has a good answer which tells me that I should probably spend more time reading the translation committee's commentary. Having said that, textual criticism is quite interesting.
While using text criticism, we look at two primary areas of evidence to support a reading: external evidence and internal evidence. With external evidence we evaluate the sources based on things like date (with older generally being preferred), distribution (with widely-distributed being preferred), and agreement across these dimensions.
Internal evidence focuses on why the variation may have occurred in the first place. This focuses on things such as transcriptional evidence ("scribal tendency"), intrinsic evidence (which reading fits the context of the passage the best), as well as the difficulty (preferred) and length (shorter is preferred) of the reading.
So, let's check out Jude 5 beginning with the "accepted" passage (note: I'm using UBS 4):
Ὑπομνῆσαι δὲ ὑμᾶς βούλομαι, εἰδότας [ὑμᾶς] πάντα ὅτι [ὁ] κύριος ἅπαξ λαὸν ἐκ γῆς Αἰγύπτου σώσας τὸ δεύτερον τοὺς μὴ πιστεύσαντας ἀπώλεσεν,
NA27:
Ὑπομνῆσαι δὲ ὑμᾶς βούλομαι, εἰδότας [ὑμᾶς] πάντα ὅτι [ὁ] κύριος ἅπαξ λαὸν ἐκ γῆς Αἰγύπτου σώσας τὸ δεύτερον τοὺς μὴ πιστεύσαντας ἀπώλεσεν,
NA28:
Ὑπομνῆσαι δὲ ὑμᾶς βούλομαι, εἰδότας ὑμᾶς ἅπαξ πάντα ὅτι Ἰησοῦς λαὸν ἐκ γῆς Αἰγύπτου σώσας τὸ δεύτερον τοὺς μὴ πιστεύσαντας ἀπώλεσεν,
That's quite the change, so let's look at the variants.
First, the UBS 4 has chosen the reading above and marked it as "D" (which is something like a confidence interval) which means "... occurs only rarely, indicates that the Committee had great difficulty arriving at a decision." (USB4 p. 3).
VARIANTS
πάντα ὅτι [ὁ] κύριος ἅπαξ This reading is attested to by א (Cx. Sinaiticus), Ψ (Cx. Athos), C* (Ephraemi Rescriptus - a rewrite [note that the article is omitted in these major uncials]), mss 1505, 1611, 2138, and a Syriac text.
πάντα ὅτι ὁ θεὸς ἅπαξ This reading is attested to by mss. 1243, 1846, lectionary 596, the Syriac Philoxeniana, and a few other texts.
ἅπαξ πάντα ὅτι ὁ Ἰησοῦς Attested by A, B (Cxs. Alexandrinus, and Vaticanus), mss. 33, 81, 2344, the Vulgate, and Jerome.
ἅπαξ τοῦτο ὅτι ὁ κύριος Attested by the Byzantine and based on Cxs. K, L (Cxs. Moscow and Paris ... with a slightly different word order), mss. 436, 945, 1067, 1175, 1292, 1844, and a few lectionaries.
(note that between 2 and 3 and 3 and 4 I skipped a few variants. Overall, there are 7 major variants in the UBS).
External evidence:
Cx. Sinaiticus dates at 4th C., Rescriptus to early 5th, and Athos to 9-10th. The remaining manuscripts date to at least 7th C. The distribution of these texts is localized to Egypt, with Athos originating from Caesarea.
No major uncial support. Again, the manuscript evidence is late, (at least 6th C.), but the Philoxeniana dates to early 5th C.
Cx. Alexandrinus and Vaticanus date to 5th C. and 4th C., respectively. Sinaiticus and Vaticanus are the two premiere codices, and the addition of a second of the "great uncials" (Alex.) lends some weight to this reading. Additionally, ms. 33 is a complete manuscript, but dates to 9th C. Other manuscripts date from between 9th C. through 11th C.
No major uncial support. Moscow and Paris date to 9th and 8th C., respectively.
Internal Evidence
Internally, there are no minor scribal tendencies (transposing letters, etc) surrounding the difference between lord, God, and Jesus. About the closest we could come to an error would be if Ἰησοῦς was omitted due to homoeoteleuton. However, Ἰησοῦς in any line before, is in the accusative case and not the nominative.
The readings are all around the same length which would eliminate any later scribal elaboration within the text. The most difficult reading is πάντα ὅτι [ὁ] κύριος ([ὁ] being a likely scribal addition to the originally anarthrous κύριος) with later clarification by scribes who read in Ἰησοῦς or θεὸς for κύριος.
It looks like the NA28 group weighed the external evidence as superior to the other arguments in their decision. The external evidence is pretty convincing, but the internal evidence is also pretty damaging. I appreciate that the group appears to have attempted to leave theology out of the textual criticism process, but I also recognize that this would have been a tremendous form break for Jude. As such, I agree with the UBS4 {D} reading, but find the third option (which NA28 took) to be almost convincing.
The Byzantine tradition is in line with the UBS4's preferred reading, though it includes what appears to be a clarifying object τοῦτο to the text (a minor scribal addition).