I would guess the point is that those companies are probably very few. If your business can't extract $12 an hour from an employee then perhaps it should fail.

The harm done by these job losses is less than the boost to the economy by putting more money in low income consumers pockets.

I'm struggling to think of an example of labour that wouldn't produce $12 per hour in value and can't be reasonably cheaply automated.

If you are going to shut down some businesses, where would more money (to low income workers) come from?

Would depend on the effect of the rising costs (of labour) of the firms. Do they make less profit or increase the prices of their goods?

Hopefully the extra money comes from an area that has a higher rate of tax than the tax rate of low income earners thus putting more money back into the "market economy". This should be likely in a country with progressive taxation.

You're talking about disturbing a system with billions of variables by changing a few thousand of them. Almost any such change will have unintended consequences of the same order as the intended effect. For politicians, this is beneficial, since they can be surprised by the consequences, ignore their causes, and advocate new changes to "fix" them. For someone who actually intends to improve things, such as yourself, the fact that poking the economy with a stick is likely to backfire should be cause for concern. :)

Yep, economics is basically equivalent to executing random scripts concurrently as root and when they crash you get to argue about what the stack trace was.

They can work independently and set their own pricing.

Yes. In this modern world, it is better to give unemployment checks to pay people to study or clean the streets than to do something no one would give a tenner for.