Add to this that the side project problem isn't relevant for many people, and people can't predict 5-10 years out whether it will be relevant to them. How many 22-year-olds expect to write a $20,000/month iPhone app, some time in the future, and aren't already trying to do it? Also, if you do a side project on your own equipment and time, most of the time you can hide the fact that much of the work occurred while under a surrender-all; if nothing else, just rewrite the code outright after you leave, since the code will be better and take much less time to write on a second write. This means, from a practical standpoint, that most people don't worry about this situation until/unless they end up in the 0.1% it actually affects and it's too late. This law is only relevant to extreme edge cases; for the rest, it merely makes a statement about the meaning of employment, and in the New York case, it makes a disgusting one.
I used to have this attitude as well that I couldn't ask for stuff and no one would agree, I gave it up, started approaching things earnestly and honestly and it's amazing the things people will do if you ask them in a respectful and polite manner.
I actually think it's best, as a general principle, not to work on side projects while employed. If the side project really is a side project (i.e. it's unrelated to your work and therefore something you justly own) you'll do a much better job of it if you work on it full-time. The side project and the day job usually both suffer if you try to serve two masters. But there are cases to which this principle doesn't apply (i.e. an iPhone app that doesn't take much time to build but becomes very popular) and it's for those that we need (in NY) better laws.
You do have to push hard though, its extra legal work for them and so they'll deter you mainly for that reason.
Libertarian arguments like this which stress contracts often seem to fall flat by the following reasoning:
1 We seem to agree on what the desired outcome is.
2 If contract theory fails here, the outcome is undesirable.
3 If contract theory succeeds, the outcome is identical to what we would have achieved with legislation.
4 A legislative solution doesn't carry with it the uncertainly that a contractual solution would.
∴ Legislation seems preferable.
You have to know what outcome you want to effect of course, and be reasonably sure your legislation will achieve that. In this case we see (or at least strongly suspect) that when companies are not allowed to own IP produced by their employees independent of company resources and unrelated to their business, the financial benefit to the employee and the benefit to society in economic growth far exceed the benefit to the company (in most cases this benefit is zero). If there were a wider range of solutions I would not support legislation (or possibly I would support narrower legislation), but that doesn't seem to be the case. So I support legislation.
You're absolutely right that companies would be better off in attracting talent by revising portions of their contract. You're absolutely wrong to suppose that they're necessarily going to do that.
You just need to mention it when they give you the agreement to sign. By the time there's a piece of paper printed out with your name on it, they're committed. And on the other side of the table from that piece of paper is the guy who hired you. A reasonable person that can recognize a silly clause when you point it out.
Just cross out the paragraph, initial next to it, then sign and hand it back. If they want to make a big deal about it, it's them making the big deal, not you.