People need to learn the difference between meaningful data and noise. Not in the image quality of a sensor, but in things like product release announcements. Everyone is rushing to make statements about "what is Nikon doing" and "where is Nikon going" and "Nikon has given up on high end DX" and all other sorts of insane speculation based upon one random data point, the D3100 specs. This is akin to betting on the long-term value of the Dow or S&P 500 based upon the daily fluctuation of a single stock. (If you want an in-depth analysis of why that is, read Taleb's book "Fooled by Randomness").
The broad trend in digital is that Nikon has done well in pushing the envelope, especially at the pro end. That started with the D1 and continues through the D3s/D3x today, with the main hiccup being the D2h (which was still better than most people gave it credit for).
When the D2h came out we heard the "I'm switching to Canon" chorus at full volume. On one short-term data point, people made large-scale economic decisions for the long term (selling off their Nikon gear to buy Canon gear). Then the D3 came out, followed by the D3s. What do those new data points say? Why, sell Canon and buy Nikon. This is classic sell low, buy high foolishness.
Single product introductions, even multiple ones, can be meaningless noise in evaluating future prospects. I love how Taleb puts it in his book (again "Fooled by Randomness"): run a Monte Carlo simulation on the same investment strategy and look at the probability of being successful on different time scales: 1 year, 97%, 1 day 54%. In other words, by getting too caught up in the day-to-day noise of individual events, you'll be less satisfied with the same portfolio the more you try to analyze it on short-term events (the day's stock prices, or in the case of Nikon, today's product announcements).
There are trends in Nikon's D3100 announcement that are observable: the D40x->D3000 line continues to evolve, video is becoming more important and getting more features with each new release, neither the basic Nikon button-and-dial user interface nor the menu structures are changing. But to start discussions or send me emails with "the D3100 is now the highest pixel count DX camera Nikon makes" is nearly absurd in terms of using it to predict the future. More likely is this: if Nikon's lowest end camera has 14mp, the future higher end cameras are likely to have more. And the pro cameras will continue to have better pixel data (less noise, more dynamic range, etc.). But if you try to predict the future from a single data point, don't expect to beat the market ;~).
Put another way, the D3100 announcement is only meaningful in terms of where it sits in the low-end DSLR market historically. First, there's a progression (D40x -> D60 -> D3000 -> D3100). Second, how it stacks up against similar competition (low end Sony, Canon, and Pentax cameras, for instance), for which we'll need to see some real performance data, not just a feature set. Third, that Nikon continues to spend a lot of time integrated video at the expense of not doing anything much on the still side that's innovative. In other words, not a lot of "news" in the announcement. Everything may still pretty much be the same as it ever was in low-end Nikon DSLR land: slightly laggard feature set, slightly better performance, (eventual) aggressive price. Thus, I see nothing that indicates Nikon will suffer much with this new camera. Just the opposite: the sane conclusion is that the D3100 will simply take over from the D3000 the role of being one of the most popular DSLRs in terms of unit volume.
What's really happening with all the angst over the D3100 (and probably over the upcoming D90 replacement) announcement is that people are displacing their frustration that they still can't buy a 24mp FX body at a reasonable price with the Nikon name on it. If you can't get what you want, then bitch about something else. Wow, that's a mature approach to life.
One curious aspect of the D3100 is the sensor. "Nikon developed" appears in the marketing materials, and the sensor is indeed not the same one as in the Sony NEX5 (as originally speculated) in several key ways, one of which is physical size, the other pixel count. The Nikon 14mp sensor has slightly more pixel count in a slightly smaller area than the Sony. Indeed, the area is no longer what I'd call DX. The D5000, D90, and D300 have a sensor area of 23.6 x 15.8 (28.4 diagonal), while the D3100 has a sensor area of 23.1 x 15.4 (27.8 diagonal). This represents a loss of about 2% in image circle coverage needed. We're now at 1.55x crop as opposed to the old 1.52x. It's an unusual change, but the bottom line is that the Nikon and Sony 14mp sensors are indeed different in important ways. But I'll repeat: the proof is in the pudding. We need to test the new sensor on live subjects before making any useful pronouncements about it.
Update: I've now heard even more details about the D3100 sensor and from a few sources closer to Nikon Japan. It does indeed appear that Nikon has moved to producing their own sensors. They may be licensing or cooperating on some of the underlying technologies, but it appears that Nikon has decided that Nikon DSLRs will in the future have Nikon-controlled and Nikon-exclusive sensors. That they haven't chosen to make a bigger deal of this is strange. And some parts of Nikon appear not to have gotten the memo on this, as I've seen quotes from Nikon personnel in some subsidiaries that say "Sony sensor."