(1) There are sensible things with efficient causes.
(2) If a thing has an efficient cause, it has exactly one efficient cause.
(3) Efficient causes are prior to their effects.
(4) Priority of efficient causes is irreflexive.
(5) Priority of efficient causes is transitive.
(6) Every sensible efficient cause has an efficient cause.
(7) Infinite regress of efficient causes is impossible.
What would surprise someone who has a fair degree of acquaintance with Aristotelian views of efficient causes woud be (2), which is a very un-Aristotelian thing to say, at least if we are talking about medieval Aristotelians. (Aristotle himself has no notion of efficient causes at all; the concept requires a higher level of abstraction and generalization than we find in Aristotle's discussions of causes.) And it's not generally in the spirit of Aristotelian accounts of causation, either; it makes efficient causation intransitive, which eliminates the possibility of (genuine) remote and proximate causes in the order of efficient causes. (Sobel actually recognizes this, but treats phrases like 'proximate efficient cause' as redundant phrases used for convenience. This might well be a modern thing to do; it's not the sort of thing you would expect of medieval scholastics, who tend to want a specific and substantive rationale for every term they use.)
When we look at Sobel's account of this premise, we see that he has been misled by translation. Sobel says (PDF),
Aquinas writes that it is not possible for "a thing...to be the efficient cause of itself" (loc. cit., emphasis added), from which I gather that things that have efficient causes are to have unique ones. He envisions for a sensible thing x, a sequence of efficient causes that lead to it. If, for example, there are three causes in such a sequence to x, then I read him as saying that exactly "the ultimate cause" is a cause of x, while "the first is the cause [only] of the intermediate cause, and the intermediate is the cause [only] of the ultimate cause" (ST q2,a3 p. 22).
Aquinas's Latin, however, has no definite article: the only options available are the (potentially more ambiguous) default case, where there is no article or article-like expression, and the use of an article-like circumlocution (e.g., of the sort that would be translated by phrases like 'a certain thing'). St. Thomas's Latin here is Invenimus enim in istis sensibilibus esse ordinem causarum efficientium, nec tamen invenitur, nec est possibile, quod aliquid sit causa efficiens sui ipsius, with no article-like expression, so there is nothing to bear the weight Sobel places on the definite article in the translation (quod aliquid sit causa efficiens sui ipsius, what Aquinas is saying is impossible, simply means, "that something should be an efficient cause of its very self" or, as we might also put it a bit more colloquially, "for something to be a cause effecting itself"). Likewise, the interpolation of 'only' into the sentence about ultimate causes pretty clearly makes it say something different from what Aquinas means, and something Aquinas would pretty clearly not accept, for a number of reasons.
But Sobel is aboveboard from the beginning that he is only considering an argument along the lines of the Second Way, and not necessarily Aquinas's own version of the argument.