Railroad to nowhere? Are you kidding? Do you realize the size of the economies and amount of existing commuter rail in LA and SFO? MetroLink in LA removes about 250 Million annual highway miles from the roads -- at a cost of about $0.28 per mile, oddly enough, less than the federal mileage deduction. In 20 years it has grown to 512 miles of track, about the same as from LA-SFO, but the vast majority is in heavily populated areas. Yes, the trackbeds were almost all already there (as, indeed they are or could be selected at least for this route), but damn near 100% of it has been replaced (and much of it doubled or quadrupled in parallel tracks from the existing freight lines) explicitly for increased Amtrak and Metrolink traffic. Flying between LA-SFO is a serious pain in the ass and at the end of it is not that much faster than driving. Sure, it's only 2-3 hours curb to curb (people have a silly tendency to not count the time milling around terminals, sitting on tarmacs etc. as if "flight time" is it). But, unless you live in El Segundo and are going to Burlingame, you only save maybe an hour flying from what it takes to drive door-to-door. I recall one trip where, having driven the exact same route less than a year prior, and despite living less than three miles from the airport with a destination fifty yards from a BART station, flying saved me a grand total of fifteen minutes over driving. You can't reasonably add any more flights between LA and SFO/OAK -- or hell, add a lane to I-5 -- without a comparatively huge infrastructure investment and the possibility of expanding surface traffic on either end is about nil. Given the timeframe involved, the fact that flight is for the foreseeable future pegged to the cost of kerosene, which ain't going DOWN in price, barring depopulation, something will have to be there that is economically scalable.
Yes, the budget and timeframe given the landscape of the route are both wildly eyebrow-raising, but the idea is hardly a "railroad to nowhere."