New language classification idea from Blench: what if Modern South Arabian is derived secondarily from Ethiopia, before partial phonological levelling of modern Ethiopian Semitic to a more Cushitic shape? This would explain why MSA is nowhere around in the Old South Arabian period.
I’m not competent to tell if there actually are enough shared innovations to positively support this, though, and I know better than to take Blench on face value on this. As usual he is more excited by archeological correlates than nitty-gritty linguistic details. Bare phonological inventory comparison has some uses in narrow dialectology, but is close to worthless within larger families. Irish is not closely related to Russian despite both having palatalized consonants aplenty; nor Greek to Spanish despite both having a full spirant series /f θ x v ð ɣ/ plus simple /i e a o u/ vowel system.
The tangent on a “Southern Afroasiatic” group of Semitic + Cushitic + Chadic is almost too unargued-for for me to even mention it, but I am amused by how almost everyone seem to have their own pet sorting of the five major AA groups; Newman’s Berber–Chadic vs. Egypto-Semitic, Fleming’s Cushitic vs. Nonethiopian, Diakonoff’s Egyptian–Chadic vs. Berber–Semitic–Cushitic, Bender’s Chadic vs. Macro-Cushitic… and now Blench’s Egyptian and Berber vs. Southern AA, apparently.