The "Stanford-Binet" test was originally devised for children, but the revisions upon it--that make it the modern IQ test--coming from Stern and Terman in large part, seem to be statistical modifications to do exactly what you're saying it can't. So, even if (your position) the test is not meaningful, the intent was to make a generally comparable measure of intelligence.
You said "IQ was never intended as a general measure of intelligence." The argument I'm making is that it seems as if that was exactly the intent of creating the IQ test.
In response to your other questions:
Do you believe that a 110 IQ person is 10% smarter than a 100 IQ person?
Unsure, but I would say no in terms of thinking about it as some kind of normalized distribution.
Smarter in what way? What does it even mean to be smarter or have a higher "intelligence quotient"? Note the use of the word quotient.
I'm not sure anyone has ever proposed that IQ is the ONLY measure that matters for ALL aspects of life. But, there's very strong evidence that whatever IQ measures has a statistically significant correlation with a number of life outcomes (job performance, income, health, educational performance, etc.).
I heard a lecture years ago where the speaker talked about IQ in terms of mental plasticity / learning flexibility. The example given was being a cashier at McDonalds. This is actually be a fairly mentally taxing job. You have to be able to listen to customers, understand special orders, know the buttons on the register to press to get the desired order in, accept multiple forms of payment, make change from cash, multitask with getting different orders to different customers, etc. Almost anyone can do this job (though as anyone who has been a fastfood customer can attest, some workers ARE better than others!). What impact does IQ have the ability to perform this job? Not a lot. The impact of IQ is in the ability to learn the computer system, learn how to handle new and unexpected order combos, etc--the learning portions.
My own crank theory is that human intelligence almost all boils down to pattern matching on steroids and that genius is pattern matching on a level that most people can't understand (or, differently, recognizing different kinds of patterns from what most people see).