There are several objections to this argument. One is that since it is estimated that something like 99.9% of all species that have ever existed have gone extinct that your failures are mostly right there in the fossil record. It is just that they aren't things like mastodons with an equivalent to
anntenapedia but things which through a failure to adapt, in many cases to a radically and violently altered environment, died out
en masse. These obviously won't look like grossly morphologically abnormal mutants.
You, or perhaps more accurately Josh Greenberger, are clearly thinking more along the lines of mutations like antennapedia or others causing gross morphological mutations we naturally consider deleterious. One important concept to bear in mind here is that fossilisation is really a very rare event and the long term survival of such severe mutations is
also often very rare, not to mention the rate of such mutations being relatively small. So despite only having a small subset of all of a species that was ever extant in the fossil record you expect to see a small subset (surviving mutants) from an even smaller proportion of them (individuals with gross morphological abnormalities).
A lot of seriously detrimental mutation are what are called 'embryonic lethals' in other words the embryo never develops into a full viable organism, the window for such embryos to fossilise is minute and our ability to distinguish a viable from a non-viable embryo highly doubtful.
In fact this might be true even for full grown adults. When we look at vertebrate fossils we can determine a progression in the developing fin/limb where the number of digits reduces in many lineages. We also see polydactyly mutations occurring in many vertebrate lineages. Is it possible that some polydactylous mutants might be misidentified as ancestral forms which all had more digits? I don't know, it would be highly dependent on the specific morphological taxonomies being used.
This is natural selection in action. To compare it to the monkey typewriter scenario it is as if every few years or so someone puts the meaningful pages in a fireproof box and sets fire to all the rest. The action of selection is what means we don't end up with a room/world full of nonsense pages/hideous mutants. The mutants die before getting a chance to fossilise.
TTFN,
WK