This is a way to launch the entire human population into space, safely and at low cost per kg. It sounds like the dumbest idea ever, except for the many similar ideas that have been proposed before it:
The idea is to have a 2500-km evacuated tube 10 meters in diameter, initially running along the ground, but with the last 450 km steered by kites, ending 30km up in the air. Launch spaceship through the tube by railgun, exiting the tube at a speed that will put them at escape velocity (11186 m/s), or maybe faster or slower. Add 2000 m/s for air drag while escaping the atmosphere, so about 13000 m/s. Speeds below escape velocity will naturally return through the surface of the earth, so the ship needs to use fuel during the first orbit to adjust the trajectory to not hit the earth. Speeds above escape velocity will also return through earth's orbit in one year, but solar sails could handle missing the earth, no fuel required. If you are launching the whole population off the earth, I assume the goal is solar orbit, not any type of earth orbit.
Rail guns use externally-supplied electricity rather than internally-supplied fuel to accelerate things. Their challenge is you have to hold the rail very steady at the end, because bumps get more severe the faster you're going. This idea calls for going 13000 m/s at the end, which is very fast.
Average humans can deal with about 5 gravities of acceleration for a few minutes. 4 gravities is 39.2 m/s2. 2000km = 1/2 (39.2 m/s2) (t2), t = 328 seconds, about 5.5 minutes. v = (39.2 m/s2)*(328 seconds) = 12854 m/s. More than 4 gravities of acceleration gives faster release speeds, fewer gravities gives slower release speeds. Alternatively, more gravities allow for a shorter tunnel. Pointing the tunnel east gives you an extra 463 m/s due to earth's rotation. There's a certain time of day best for launching into solar orbit, too, but I'm not sure what it is.
You want the tunnel to angle up as much as possible at the end to reduce total atmosphere drag and reduce the length of the tunnel supported by kites. So first you accelerate sideways in the tunnel on the ground to release velocity, then you accerate up to increase the angle of the tube. Accelerating straight up at 3 gravities (plus the 1 gravity down due to actual gravity, giving 4 gravities total), 30000 m = 1/2 29.4 t2, t = 45 seconds, that's 45s*13000m/s = 580km of kite-supported tunnel. But note the earth is round. A tangent to the earth's surface will be 30km above the surface 617km away; you'd actually have to accelerate down at 2 gravities (3 really, but earth supplies 1) to keep to the surface at escape velocity. So you get effectively 5 gravities up instead of 3, and need only sqrt(30000/(49/2))=35 seconds accelerating up, so only 35*12854=450km of kite-supported tunnel. The ship would leave going 13000 m/s sideways and 1700 m/s up. And the earth keeps curving. The dense atmosphere would be left almost immediately and the atmosphere would be left entirely in under 40 seconds.
There is a jolt when the spacecraft leaves the evacuated tunnel and hits the atmosphere. Heavier and more streamlined spacecraft will get less of a jolt. At 30km up, the atmosphere is 1/100th as dense as at the surface. Beyond 100km up, the atmosphere doesn't exist. 2000 m/s extra may be more than needed to account for punching through the atmosphere. The seats in the spaceship need to rotate: you feel gravity backwards in the tunnel on the ground, then down in the kite-supported section of the tunnel, then forwards when the spaceship hits the atmosphere.
The kite-supported tunnel can be tethered to earth. The kites need to be actively steered to counteract winds. Stratospheric winds above about 30 degrees latitude consistently blow from west to east, pulling on the kites and keeping the tube in tension. The tube is the tether. The kites can steer to quickly compensate for wind direction, and they can provide aerodynamic lift despite being heavier than air. Railguns cannot tolerate bumps at high speeds, so a very challenging requirement is to keep the railgun inside the tunnel to within millimeters of the right place no matter what the wind is doing. It can drift by kilometers over time, but its location for the few seconds of each launch can vary only a little and only smoothly and almost predictably. Another challenge is that people cannot maintain this, it has to be robots.
Could you use balloons instead of kites? Probably not. Helium balloons are expensive. Hydrogen balloons are cheap but they blow up. If you surround H2 by CH4, and surround that by N2, and surround that by atmosphere, then adjacent layers are not mutually flammable. This could plausibly be made to tolerate nature. However, any structure supported by balloons relies on people not trying to attack it. If people are attacking it, don't try to launch any spaceship with it. Some of the power needed by the elevated tunnel can be generated from the wind. You need kites anyways to steer it, and they can provide enough lift on their own. Adding balloons seems like a liability.
The tunnel itself, however, is sort of a balloon. It is an evacuated pipe. If you make the weight of the pipe small enough, it becomes self-supporting due to being a balloon. And any weight you can cut off the pipe is weight you don't need to add kites to support. Make the pipe out of some 3d-printed carbon latticework rather than bulk steel, with big beams supporting fractally smaller and more numerous beams until it reaches the surface of the pipe to support a thin solid (mylar?) airproof sheet.
The end of the tunnel is difficult. A ship hitting the atmosphere at 13000 m/s will cause the atmosphere to spontaneously combust. Although the tunnel is evacuated, the atmosphere isn't, and the tunnel has to be open where the ship escapes, so the atmosphere will rush into the tunnel while the end is open. The atmosphere that gets into the tunnel will ignite and hit the tunnel with severe sonic booms as the ship goes through it. These can both be helped by having a small tunnel exit, but a big bulb at the end of the tunnel for incoming air to expand into (and to keep the walls further away from the ship). (Explosions would also be harmful if balloons were used for suport.)
If launches are more than 10 seconds apart, it makes sense to have doors that close when there is no ship. It takes 3 minutes to slow a ship down so it doesn't hit something broken at the end of the tunnel, so safety calls for launching no more than once every 3 minutes. There are probably 2 hours a day that are good for getting into solar orbit, so that's 40 launches a day. With 70 people per launch, that's 1 million people a year, which is 1000x too little to launch the whole population. If you made a thousand of these, that could lift the whole population.
If launches are 1 second apart, there is no point in closing doors during the launch hours, just add more pumps and bulbs to evacuate faster. A broken end of the pipe means hundreds of shiploads of people will die. However it would allow 2*3600*365.25 = 2.6 million launches per year, and 380 people per launch would hit 1 billion people per year with just one tunnel, which is enough to lift the whole population. However, there is recoil from each launch, and I'd be surprised if the track could be sufficiently stabilized in a second.
If instead of one tube you have two, and the tracks can optionally cross over from one tube to the other, and launches in the tubes are alternated, then a broken end can be avoided by any car that has not yet passed the last crossover. That can be in the kite-supported region, ten seconds back, allowing two launches every 10 seconds: 263k launches per year. It would take 3800 people per launch to hit 1 billion people per year from one such paired tube, which like sounds too much. Ten of these could launch the whole population in 10 years, though.
A smaller prototype of this idea could launch satellites with 50g accelerations. Almost all the tunnel would be supported by kites.
The main advantages of this approach are it gives you a long inclined railgun without digging any holes or strategically placing mountains, and it avoids the atmosphere until it reachs a point many times higher than any mountain. Unlike a space elevator, it can be built with existing materials.
ISEC proposes supporting a space elevator with a series of ambits, a term they've co-opted to describe a semicircular arch high up, supported by heavy bolts continuously shot at very high velocity through an evacuated tube up to that arch. These ambits would be several thousand kilometers up. Keeping the tube evacuated and steady is very similar to the evacuated railgun proposal here, except that the tube is much smaller and levitating the tube is done by stealing momentum from the up-going bolts. Getting the bolts up to speed has to be done at extremely high accelerations at ground level (they propose doing it in the ocean, which gives them a track up to 3 kilometers down pretty easily.)
The least likely bits of the railgun proposal are having the atmosphere support it beyond about 5 kilometers up, having an open end in the atmosphere, and the ship blasting through the remaining atmosphere. That can all be solved by using ambits to support the end of the railgun, all the way up to 100km where there is no more atmosphere. But, it requires supporting about 300km of additional evacuated railgun, which is probably 600 such ambits rising to heights between 5km and 100km. Between the ambits would look like a suspension bridge, with the ambits as the towers and the railgun tube as the road.
Firing bolts to keep the ambits levitated takes continuous energy. Since the ambit tube is much smaller than this railgun, the effect of wind is less, but the requirements of holding it steady are more stringent. I don't know if adding ambits to this levitated railgun helps or hurts overall. I am doubtful of the ability of a thin ambit tube to stay steady enough against gusty winds.
This essentially puts an airplane cabin full of people into orbit without killing them. However, once they escape earth, they still need to breath, eat, go to the bathroom, and eventually exercise. They start out in an orbit (position + velocity) that intersects earth, which is dangerous. So I don't expect them to approach anything worth approaching for weeks, unless someone is going out of their way to meet them. They could probably do small orbital adjustments to clump together 24 of these containers into two rings, spin them up so one goes one way and the other goes the other, then there's a somewhat long way to walk plus gravity. But they'll still be as crowded as an airplane cabin, with no supplies beyond what they brought. (The obvious solution to that is to pack fewer people and more supplies in each launch.)
Making worthwhile places in space for humans to live is a different problem. My best guess is humans won't colonize space. Robots or modified humans with more space-friendly bodies will. So lifting all existing humans off the planet will never be an issue.
There are several pieces of this that could be built and tested independently with relatively small funds.