Manufactured material is a waste of energy - that needs constraining to a minimum amount of lifts. For anything significant to be constructed up there then it's seriously going to require harvesting materials from up there, and using them to build manufacturing facilities.... Look between Mars and Jupiter, almost every element in the periodic in them rocks, and squish them together there's a mass of about double Earth - before we're even ½ way through harvesting that we can be shipping material back from out in the Oort. There's really more than we can use, even accounting for exponential expansion, lifting it from the floor is just... restrictive.
Yes, smaller is "easier" to launch. That's the point of moving about ½ way up before launching, requiring a lot less fuel. You can do that with a much much smaller rocket and then still carry similar payloads - the 2.5Tonnes I'd guessed at payload capacity is actually quite tiny. I wasn't thinking of breaking orbit, but yes, the same design with less payload would easily. And no-one really launches 1KG, it's mathed like that because more weight needs more fuel - but the fuel isn't the major part of that cost, it's the fact that 90% or so of the rocket is disposable, it's taken a lot of man-hours and commonly with high grade materials machined to high tollerances... By cutting off the lower stages of the rocket - which is where most of the fuel is wasted - you should be able to drive costs down significantly.
As to the rail launch system... Not a lot has changed in 50 years with physics, all the rules are still the same. Lets assume you've no magical way to do this in vacuum that also allows to exit at rediculous velocities. The principle "fail" is going to be wind resistence. 7.5Km/s in atmosphere isn't happening, most likely. To take something like the Bugatti Veyron, that's fine up to about 150MPH, or only a dawdling 67.056 m/s but after that the exponential increase in wind resistence is what eats up all the engine power, charicterised by a noticiable drop in the steepness of the accelleration curve, it becomes flat compared to say, 30MPH-90MPH. Lets ignore that you'll create friction enough to become a fireball well before 1800m/s, taking it past 2000m/s will take significant energy even with "perfect" aerodynamics. To avoid getting lift at this stage, either removing from the track or providing friction against, it'll also be useless for controlled re-entry. Once you get up past 20,000 meter the air is maybe thin enough to think about hitting those sorts of speeds. It'll still flame 2000m/s+. Then, somehow, if you do manage to acheive these speeds then the curve as you move from horizontal to vertical for launch would generate unacceptable G. The only mitigation being to reduce the incline, and increase space giving you about thirty miles of gradual increase in elevation, and even if you can build that big, you'll of lost most energy before you'd left the end of the rail. Fire up KSP and try to build such a system, I've not managed to kill any with G so that's not going to be a factor. Space might be an issue, building a multiKM rail might be out of the question but doctoring the physics on an electrical motor might allow to hit velocities before edge of runway - which could be remodelled into the curve... Would allow you to at least see what I mean with air resistence and gravity stealing all your momentum before ½ way, and the temperatures generated in the attempt.