Almost all books/articles on the S-M Gambit are rather one-sided and tend to underestimate the chances of the other side. So I assume - note the title! - that Taylor's book/article suffers from the same disease.
Again and again I must state that 6...a6 is as good as most other defenses, but not a clear cut refutation. A thorough Black preparation certainly will pay-off, but stating that 6...at is an easy way is a sign of underestimation. This is exactly what White hopes for.

Well, 1.e4 c5 2.d4 cxd4 3.c3 Nf6 4.e5 Nd5 is exactly the same as 1.e4 c5 2.c3 Nf6 3.e5 Nd5 4.d4 cxd4. In the time I used to play the Smith-Morra, 5.Nf3 was considered best.
But when you compare 1.e4 c5 2.d4 cxd4 3.c3 d5 4.exd5 Qxd5 5.cxd4 with 1.e4 c5 2.c3 d5 3.exd5 Qxd5 the first is inferior for Black compared to the latter.
On the second question I can only answer yes. After Qxd5 5.cxd4 Nf6 6.Nf3 e6 7.Nc3 Qd8 White's best for instance is 8.Bc4 transposing to the Steinitz Variation of the Queen's Gambit Accepted.
Another transposition is 1.e4 c5 2.d4 cxd4 3.c3 g6 4.cxd5 d5 5.exd5 (5.e5 is not appealing) which can become the Panov Variation of the Caro-Kann with g6.
So if White wants to avoid theory the S-M Gambit might not be the right choice!

I will not deny, that this setup is playable - it is as good as many other defences. But Lubo's opponents probably did not play the plan with Rf1-d1-d2 and Ra1-d1 (with Bf4 in between).

Yes, the Finegold is no easy defense to play, as black has to be aware of many tactical subtleties to defend correctly.  In fact, I have never tested it in an OTB game, simply due to the fact that it requires a lot of memory work for one sideline.  The first defense I listed is much easier to play, and is usually more than sufficient to get the edge needed to play for a win.

@MNb:

I don't understand (1).  I have 6...Be7 in the move order.  Maybe you think I meant 6...Nf6?

Also what do you think is white's best try against the Finegold?  The main line seems more difficult for white than most Smith-Morra lines.

I noticed after i posted it sorry

In the hands of a strong tactical player, I think this line is very dangerous up to 2100 level.  But to play it well at the upper end of this range, you really need to know what the critical lines and how to be tricky.

There are two lines that I have studied that I think cause white a number of problems:

1)  First, there is the line recommend by Gallagher in Beating the Anti-Sicilians:   

1.e4 c5 2.d4 cxd4 3.c3 dxc3 4.Nxc3 Nc6 5.Nf3 d6 6.Bc4 a6 7.0-0 Nf6 

I think if you play the Smith-Morra, you will have a lot people play this against you, partly since it is recommended in a well-known book.  Also, this system is fairly simple to play for black, once you have learned the reasons for this exact move order.  Black has to be aware of a few tricks, but many of the moves are very natural and usually not too difficult to find, even in a speed game.  I think white has a couple of tricks that Gallagher does not mention, but I am not aware of a way for white to get full compensation.  If you're white, you should probably consider 7.Bg5 and 7.b4!?.  7.b4!? is not mentioned by Gallagher and is a little tricky.  This would be very likely to catch someone off guard as a surprise weapon.  In two games I have played against experts (rated about 2100 USCF), they both preferred 7.Bg5.  Gallagher recommends 7.Bg5 e6 for black, and if black plays 7...Bg4, white can play 8.Qb3!, winning back the pawn.

2)  You probably won't have too many people play this against you, but there is a line named after Ben Finegold, called  the "Finegold Defense" by Bob Ciaffone, the writer of a self-published book entitled Smith- Morra Gambit Finegold Defense.  It goes:

1.e4 c5 2.d4 cxd4 3.c3 dxc3 4.Nxc3 d6 5.Nf3 e6 6.Bc4 Be7 7.0-0 Nf6 8.Qe2 a6

Black has to be very accurate in his move orders, but when he gets it right, white just looks dead worse.  Once again, it is probably unlikely to run into someone who plays this, but if someone is very determined to refute the Smith-Morra, you've got to watch out for this possibility.  The book is self-published and unusually written.  It is neither written in a tree or illustrated games format.  The author simply discussess analysis of the primary lines, with a given starting position in the first several chapter.  It isn't until page 76 that the author thoroughly explains the reasons for the move orders chosen in the first several moves.  I think this book would be very difficult to understand if you didn't have a good command of English, as the text is very wordy.  Pretty much, you have to read the whole book (it is only 140 pages though with fairly big print and several diagrams) to understand what is going on, but the author has put a lot of effort in the explanations.