Earlier quoted context omitted.
"someone can provide good evidence that it actually is environmentally beneficial" I don't think anyone is, just that its less bad than the alternatives. You don't drive, presumably you walk or cycle then. Each of those activities have negative environmental externalities compared to not doing them, after all you have to consume extra calories. So while each is infinitely better than driving, they aren't environmenta…
"to me sustainability is about how long you can keep doing something [...] glass is infinitely recyclable, the plastic isn't" Exactly - it's about reuse (ie. infinitely recyclable), and the shift in thinking that comes along with that. Burning plastic releases toxic byproducts into the local environment, and requires us to use more fossil fuel hydrocarbons to make new plastic. On the other hand, instead of transporti…
> regardless of how much energy it requires (up to a point, and assuming clean energy sources)
The world isn't yet powered through clean energy sources in any significant way. Electricity production runs mostly on fossil fuels, and then electricity is currently only a part of the equation - heating and transportation is still mostly powered by burning hydrocarbons pulled from the ground.
It is not just valid, but very important, to consider sustainability in context of system at the whole. Greedy algorithms don't work here. A glass bottle may be "infinitely recyclable", but if doing the actual recycling involves using energy from unsustainable sources, it's not a sustainable behaviour in itself. Sustainability is an end-to-end thing; if any link isn't, the whole chain isn't.
Now that we've established that glass recycling isn't currently a sustainable practice, we can still consider whether it's better or worse than alternative unsustainable practices. Regular "single use, thin plastic bottles, destination: landfill" approach seems idiotic. But alternatives involve "thick, refillable plastic bottles" (something I recall Germany used to do), or "single use, thin plastic bottles, destination: power plant" which isn't as dumb as it sounds - if it could be made the primary fuel source for said power plants, it's a step up. I reserve judgment on the alternatives, as I have no data. And it's kind of a problem today: regular people don't have an easy way to review end-to-end performance of such strategies. So they vote for convenience, or for whatever looks ecologic.
Burning hydrocarbons has still place in a sustainable future - they're a great high-density energy store, after all. But to make the chain end-to-end sustainable, we'd have to be using clean and sustainable sources to take back all the emissions and to manufacture more hydrocarbons.
EDIT: And to be clear, I'm always biased in favour of reusing and (then) recycling. But the energy calculation has to make sense, because sustainability is in big part an energy management problem (the other big part is habitability management problem - not doing net damage to the ecosystem; but, assuming the will to do it, it also quickly becomes a subset of energy management).