That parasitic vending machine has another parasitic vending machine next to it, which has an ever better prediction model, such that it buys from the other two machines only when it knows that the second one hasn't bought enough to meet future instant demand. For instance, perhaps it knows the schedule for a nearby convention, when the other parasite only knows average daily purchase volume. So 2 hours from the end of the headline event, the third machine puts in a big order from the first, and then two hours before the end of the expected surge, buys out the second machine so that it can't undercut using its available stock. It then sells out at a higher price and waits for the next opportunity.
So the only time anyone ever sells a drink for $2 is when an unexpected busload of thirsty tourists drops by and buys out the entire stock of cheap, instant beverages.
In your example, you have also forced my hand. What happens if I instead buy a 2 hour ticket, right in front of the guy, and ask what his price would now be to trade my ticket for his?
A couple of interesting (and non hypothetical!) things to ponder. What is the value that motivates people to buy DVDs that are a bootable install and/or live image of a free OS for $9.99 at Fry's? Similarly, what is the purchase calculus that someone does when they buy a paperback book in an airport bookstore which is a book they already own and have a perfectly readable copy at home?
We don't have Fry's here, but depending on your 'net connection it can be cheaper / more time effective to buy a physical version instead of a download. See also physical console games.
I see it rather differently, in part because I think this statement conflates two things. One is information value and the other is how we have tried to stuff information value into goods economy ideas. Perhaps an example will help clarify my understanding for you.
This example comes from a friend of mine at Facebook who was getting his MBA from Berkeley at the time;
Imagine you have a soda machine, it is sitting there with a selection of ice cold beverages. Now anyone who has traveled around knows that the "price" for a single beverage from a soda vending machine varies, often by a lot, from a low value when it is in a company cafeteria to a high value when it sits in the floor vending nook of a high priced hotel. But its the same soda. Our imaginary machine is different. It has a variable price right in the machine! If you give it $2.00 it will immediately dispense an ice cold soda, if you give it $1.00 it will give you a ticket that will dispense your selection 60 minutes from now, and if you pay it $0.50 it will give you a ticket that will dispense your selection two hours from now. The purpose of the machine is to capture as much value for the soda as possible, from the "I don't care what it costs, give me a soda" crowd, to the "Ok that is the same as I would pay for it in a grocery store" crowd. All from the same machine.
Now you're sitting there looking at the machine and counting out $2.00 when someone walks up to you and asks what you're going to buy, they offer to sell you their ticket, which they bought nearly two hours ago, for that soda which is going to drop in a few minutes for $1.50.
Think about that transaction for a minute. What did you buy? Did you buy a soda? Or did you buy the information that a soda would drop in a few minutes? There is no copyright here, no patent, there is instead two of the fundamental ways that information gains value, it is timely and rare. Timely, in that you want a soda now and this person knows when that soda will be available, and rare in that there isn't really any other way to get that information, only the person who bought the ticket (and presumably some timer inside the machine) has it. It was always a $0.50 can of soda, but that extra $1.00 value came from information about its availability which you bought from this stranger.
There is an interesting experiment you can run which shows this effect pretty clearly, in the experiment you have three operators and a number of test subjects. Operator #1 is offering to buy the information about where a queen is positioned on chessboard in a building several hundred yards away. You get a ticket for the information that is wanted, walk over the building, show it to Operator #2 who is standing by the chess board, they look at your ticket and move the queen to the place on the board, you return to operator #1 and collect your fee. Now you have operator #3 stand somewhere in the path between operator #1 and operator #2, they offer to sell you the information about the position of the queen for 10%, 25%, 50%, or 75% of the fee value.
You can use information economics to understand the natural value of information (vs the imposed value) and when information is "intrinsically" valuable or merely "temporally" valuable, and when you can create value out of collections of "free" information.
The takeaway though is that it is the information does have a value (again if it did not then people wouldn't miss it if they didn't have access to it) but individual assessments of value differ (why will some people subscribe to the WSJ and others only read it by going to the library and reading it for 'free'[1]).
And as a PostScript I specifically picked a DVD full of zeros because as you point out a DVD full of 1's is a blank DVD that can be programmed and a DVD full of random bits is an entropy source which can be used to improve the strength of one's cryptography, so really only a DVD full of zeros has the least value.
[1] Travelling to the library has its own opportunity costs and costs of execution (getting there, meeting time schedules, keeping a library card in good standing, Etc.)