I support the government acting as a solver of coordination and lack-of-information problems.

To reuse an example I brought up in another discussion, suppose that a company is using a chemical in some manufacturing process which is highly toxic, and that toxic chemical is making its way into the population in harmful quantities. 0.2% of the population knows about this and understands the danger, and of these, all who do not work for the company oppose the practice. The remaining 99.8% of the population has no opinion.

In such a situation, a boycott is highly unlikely to be useful (getting a boycott to work even under favorable conditions is a formidable coordination problem, and it's much worse in a situation where most of the population is unaware of the relevant information, since any attempt to raise awareness has to compete with every other source of information jockeying for the target audience's attention.) However, if the concerned parties can go to the government and say "this is the evidence that this manufacturing process is harmful, we all agree that it's too dangerous to allow," then the government can review the information and decide whether the process should be banned or not. By having a body which can engage in full time review of public concerns, the population can address more issues than if individuals had to research all the issues that might be relevant to them all the time (they have other things to do which put constraints on their time,) with a greatly reduced opportunity cost compared to every member of the public having to address all those issues for themselves.

I do not think that rights, negative or positive, are a particularly useful way of framing what the government should or should not treat as within its purview. I think that there are some ethical injunctions the government should follow against certain actions even if they may seem like good ideas at the time, when we know that there are certain things that tend to appear to be good ideas at the time and then lead to bad consequences anyway, thus warranting a policy based on the outside view. But I think that framing issues in terms of rights is a bad way to sort out what are and aren't good policies to pursue. I'm going to be the second person to reference Yvain's work here, and quote from his Non-Libertarian FAQ (rights and heuristics section)

Third, when push comes to shove the Non-Aggression Principle just isn't strong enough to solve hard problems. It usually results in a bunch of people claiming conflicting rights and judges just having to go with whatever seems intuitively best to them.

For example, a person has the right to live where he or she wants, because he or she has "a right to personal self-determination". Unless that person is a child, in which case the child has to live where his or her parents say, because...um...the parents have "a right to their child" that trumps the child's "right to personal self-determination". But what if the parents are evil and abusive and lock the child in a fetid closet with no food for two weeks? Then maybe the authorities can take the child away because...um...the child's "right to decent conditions" trumps the parents' "right to their child" even though the latter trumps the child's "right to personal self-determination"? Or maybe they can't, because there shouldn't even be authorities of that sort? Hard to tell.

Another example. I can build an ugly shed on my property, because I have a "right to control my property", even though the sight of the shed leaves my property and irritates my neighbor; my neighbor has no "right not to be irritated". Maybe I can build a ten million decibel noise-making machine on my property, but maybe not, because the noise will leave my property and disturbs neighbor; my "right to control my property" might or might not trump my neighbor's "right not to be disturbed", even though disturbed and irritated are synonyms. I definitely can't detonate a nuclear warhead on my property, because the blast wave will leave my property and incinerates my neighbor, and my neighbor apparently does have a "right not to be incinerated".

If you've ever seen people working within our current moral system trying to solve issues like these, you quickly realize that not only are they making it up as they go along based on a series of ad hoc rules, but they're so used to doing so that they no longer realize that this is undesirable or a shoddy way to handle ethics.

In general, I think that the government should act according to a decision process of "what, within the ethical injunctions we're restricted by, are the most positive impacts we can make on society, according to our best understanding of the public's preferences should they have the information available to us?"