I have used a lot of Nikon-format lenses over the years, including the Nikkor 18-105 VR, Nikkor 16-85 VR (twice), Nikkor 18-200 VR, Tokina 11-16 (three times), Tamron 17-50 (non-VC), Sigma 17-50 OS, Sigma 10-20/4-5.6 (three times), Nikkor 50/1.4D, Nikkor 35/1.8G, Tamron 70-300 VC, Nikkor 70-300 VR, Nikkor 105/2.8 VR Micro, Nikkor 55-200 VR, Nikkor 85/3.5 Micro, Nikkor 24-70, and probably some others that I'm forgetting. I've also used the older Canon kit lens as well as a ton of MFT lenses (my second system).
Despite all of this, I've never used the Nikkor 18-55 VR. But while I haven't used the Nikkor 18-55 VR, I know its reputation and stats are worse than the 18-105 and 16-85, and arguably even the 18-200. Therefore, I am basing my experiences on the 18-105/16-85/18-200 and figuring that the kit lens is no better than those lenses in low light or in portraits, which I think is a fair statement, especially since all of them start at f/3.5 and the kit lens has the cheaper version of VR, not the more effective version.
I wasn't happy with the 18-105/16-85/18-200 as a general purpose lens. Not good enough in low light many times, and not thin enough DoF on the long end. Therefore I probably would not have been happy with the 18-55 VR lens, either.
Was it worth it to get the 17-50 OS instead? It depends, but for me, the answer was yes--barely. CONS: A multi-lens kit would cover more focal lengths, and a good alternative lineup is a 18-55 VR + UWA + telephoto + 35/1.8 for low light. And yes, 17 isn't much wider than 18. PROS: Every little bit helps, whether it's 6% or 12% extra width. An 18-55 + UWA covers more wide angle focal lengths, but it also involves switching lenses and carrying two lenses instead of one. Plus I like the convenience of a zoom that can go from landscape to portrait in seconds and can take stabilized shots at f/2.8, which gives greater DoF than a 35/1.8 wide open, plus it's actually effectively FASTER than the 35/1.8 for stationary objects. Field curvature issues are overblown as far as landscapes are concerned; your DoF is huge anyway. And if you're taking shots closer in, esp. at f/2.8, your corners are likely out of focus anyway. Lastly, the lens is popular enough that I expect to be able to sell it easily at a reasonable price.
That last part pushed me over the edge: resale value. The best place to skimp is on the camera body, which depreciates quickly, hence why I got the D5100 instead of the D7000. (Any lower than the D5100 and I think the hit to quality isn't worth the savings.) Lenses depreciate slowly, so as long as you can front the money and don't damage the lens, you should be able to resell. You might even have a net negative or zero expense, if you bought the lens at a good price--maybe you bought it used, too!
Does this mean OP should buy a 17-50 OS too?
Not necessarily. It depends on how much he values convenience and money, among many other things. I agree w/ you that, given his stated needs (wide angle, low light, midrange focal lengths), he may be better served by a dedicated UWA zoom + kit zoom for stationary low-light objects + 35/1.8 for moving objects in low light and to act as a portrait lens in a pinch (despite perspective distortion problems).
On the other hand, that's a lot of lenses to carry around, since "hiking" was one of his needs, which implies that he'd be interested in portability. UWA + kit + 35/1.8 is a lot bigger and heavier than a single 17-50. Furthermore, since he has a tripod (and presumably a head capable of panning), he could simply stitch his way to wider angles. If he has an L-Bracket or GigaPan, he could also do vertical stitches almost as easily as horizontal stitches. And he retains the convenient 17-50, constant f/2.8, and stabilization.
I agree that there isn't a linear relationship between price and performance. We see this in everything from camera gear to video cards to automobiles where every little step up costs more and more. I also agree that oftentimes people get recommended stuff that might be overkill, especially if they are on a budget. So it's a personal decision as far as how much one is willing to front on lenses, because even if lenses didn't depreciate at all, they still costs a lot of money up front.
Everything is a tradeoff. Price vs performance, size/weight of two lenses or more instead of one (hence why superzooms are popular).
I think between you, me, and everyone who replied, and online references, OP should have enough information to make a decision by now.
Personal preference on the wide angle lenses, but I have owned a couple UWA lenses (Sigma 10-20, Samyang 14mm on full-frame) and I did not find that extra width to be significant.
I've owned 3 copies of the Tamron 17-50 (non-VC), plus the Tamron 28-75 (for 5D), and the Sigma 17-50. And I have the 18-55 VR now. I compared shots I took wide open with the f/2.8 lenses to the wide open shots I took with the 18-55, and there is little difference even when viewing at 100%. In many cases the shallower depth of field negates any resolution advantages. Maybe the f/2.8 lenses are sharper when stopped down to f/3.5, the max aperture of the 18-55 (2/3 stop), but that's not a fair comparison.
I'm not saying the 18-55 VR is as good as the Sigma 17-50 (of course not), but it's not a big jump in image quality as reviewers on the net would lead you to believe. The Sigma and Tamron lenses have quite a bit of field curvature so the corners wide-open are not much better than the kit lens. And yes, the f/2.8 lenses are better for portraits than the kit lens, I'm not debating that.
It's a testament to marketing that every photo forum site on the net tells new DSLR owners that their kit lens is crap and they need to upgrade to a 'much sharper' piece of glass. Yet photos taken with the kit lens by experienced photographers shows that the difference becomes really negligible compared to the lighting, composition and post-processing skills of the photographer.