For anything to become a factor in evolution there has to be survival advantage. If the faster texting teenybopper was more likely to win his girl (or her man) then faster texting might be a trait carried to the next generation. The problem is that technology moves much faster than evolution can hope to - the next generation of teenyboppers might text by hand gestures alone and all those super-fast thumbs would go to waste!

I think this question has evolved since I answered. I really must start copying the question into my answer before I start typing.

Well, the QWERTY keyboard has been around for over a hundred years and, although we have developed strangely curved keyboards, we have not developed strangely curved hands. The SMS has been in heavy use for less than ten years so I don't think that there is any need to worry.

 As for the broader question; yes, technology is increasingly the means by which we come to create and define our own environment, rather than merely living in it. And that naturally will alter the course of evolution, but I doubt that we'd get stupider. It seems that primate intelligence correlates with average group size (NS), and as we tend to now live in larger, more complex societies, then, all other things being equal, and given that such societies remain viable over at least tens of thousands of years, which may seem doubtful, we should get smarter.

 Another issue is that of artificially accelerated evolution, designer babies and the like; and whether and to what extent we'll eventually merge physically with our inventions.

 Still another issue is how evolution happens; in fits and starts, aka punctuated equilbrium, as Stephen Jay Gould believed; in which case a truly serious global crisis would be required for anything significant to happen; or whether it is a smooth, incremental operation.

This question raises themes of enormous scope, so everything I say (as well as, in my opinion, everything everyone else says) is subject to a great deal of context and qualification. Quite reasonably Kin doubts that we'd get stupider and mentions the correlation of intelligence with average group size, and the fact that our societies grow larger and more complex societies. However the actual situation seems to me a great deal more complex than that suggests. The ideas of HG Wells concerning the Eloi and Morlocks or the castes of Selenites might regain relevance in a different context if our increasingly specialist technologies continue to reduce the demands on the minds of users, while increasing the demands on the brains and skills of the developers. I do not expect to be in a position to criticise our progress when some of these dilemmas drive us to terminal decisions, but I seem to see some nasty brick walls across certain of our future paths. I have already in the past remarked on our inferiority in certain respects to the caste system and community commitment of eusocial creatures such as termites. When it happens (if as a community we last long enough to see it happen) that our survival depends on truly large-scale projects, then we shall depend terminally on our ability to overcome aspects of our nature that so far we have shown no interest, let alone capability in dealing with.

Kin also spoke of the ultimate abomination and unthinkable: “...artificially accelerated evolution, designer babies and the like..." Such things lead to the question of, not so much whether we fuse with our own inventions, but whether we are merely a stepping stone in the development of a future teleological evolution of mechanisms. A large future, and a larger subject.

The question of punctuated equilibrium is a totally different matter, simply one aspect (certainly an important aspect but just an aspect) of naïve views of traditional Darwinism. The so-called punctuation event is not particularly abrupt, and the most interesting aspect is not the event but the apparently stable intervals between. These reflect selection for compatibility with the status quo, and anyone who has tried to battle with the status quo under stable conditions will understand the power of the status. In either case, the upshot could be either increased or decreased intelligence in or between such events.