Helium is only 5ppm in the atmosphere. Extracting useful quantities of it that way will probably never be economically viable. In other words, if for some reason we can no longer get helium from natural gas wells then it will be cheaper to just let patients die instead of doing cryogenic distillation of helium from the atmosphere to run MRI machines.
We are already separating out the majority elements from air via ASU plants, so we should compare the abundance of helium in what is left from typical extraction. And that looks quite technically viable, if obviously uneconomic at present.
This is a very good point.
Oxygen, nitrogen, CO2 and argon make up 99.94% of the atmosphere. The remaining 0.06% has 5ppm is nearly 1% helium. That's up 200x from the original concentration and is well above the 0.3% that is sometimes quoted as the limit for economic extraction of helium (and well below the 7% of some natural gas).
Furthermore, the leftover gas is also already cold. It is absolutely true that 85K isn't very close to the boiling point of helium, it is a lot closer than starting at the temperature of gas at the well head.
The gotcha is almost certainly going to be that an ASU probably doesn't liquify most of the gas it takes in. That means that the exhaust gas will only be slightly enhanced.
In a world of extremely cheap solar electricity pushing grid prices negative, a lot of things might be a lot more economical then conventionally thought though - particularly when you factor in the desire to get a full return on industrial manufacturing of panels.
For me personally, this is one of the most promising aspects of solar that I hope to see in the future. There are many, many things we could do but currently do not because the energy cost is not worth it. Push the energy cost to zero, or even below, and it will be interesting to see what new things become abundant.