>>it is hastened by unionization
True in a technical sense, but misses the point. Sure, there's technically a price point at which automation is "worth it." If labour costs more, demand for labour is lower. That might be because of automation, or any of many ways that demand reacts to price: efficiency, decreased production, increased capital expenditure, outsourcing, whatever.
So really, the true generality is not that "the biggest thing working against unionization is automation." The biggest thing working against labour is wages. Higher wages, less demand for labour. This is also true about anything, not just labour and it's somewhat banal. If cornflakes or netflix cost more, people will buy less. That doesn't mean you always have to price to lose money as a business. Same for labour/unions.
In any case, the bigger point is that if some job is almost "worth automating" at minimum, market or union wage... That's not worth defending. It will probably be automated soon anyway, and there's little to gain by defending it.
>> Unions are about minimum wages.
This is very far from true. For one thing, most people don't earn minimum wage. Most union workers don't. No one wants to. In fact, neglecting the lowest rungs is a habitual mistake of unions.
Starting from the premise that unions are about minimum wages misses most of why the current amazon stuff is even interesting. Depending on what kind of union(s) they form (getting back to my original point) unionisation could mean a pretty wide range of outcomes. Amazon has a big pay scale. There's nothing to say that unionisation means nothing for those at the $200k point on that scale.
The basic premise of an amazon union is (1) amazon is very profitable (2) Negotiation power/leverage is always relative to the size of the pie. A union is what you negotiate with, profit is what you negotiate for.
Bluntly, amazon have $100bn in profits and 1 million employees = $100k per employee. Rarely has a plumper prize been seen on the horizon.