while he will just take a few days off work to deal with potential customers, slap together some marketting text and call it a day.
If that's how you really feel then you've got a long and painful road ahead of you. This is the wrong way to look at things, and it's likely to lead to your failure, unless you change your mentality.
9/10 times you will need both technical and business skills to make a startup successful. And they should both be thought of as equally important. It's common for technical folks to think that writing code is the only thing that matters. However, there are just as many business folks that think the idea is the only thing that matters, and that developing the software is easy and unimportant.
If that's how you really feel about the non-technical tasks involved in a startup, then you are unlikely to treat your co-founder with respect, which will lead to many problems. How would you feel if your co-founder said:
He will just take a few days off work to slap together some code and call it a day.
Doesn't feel so good. You'd tell him that he doesn't know what's he's talking about, and that developing software is way more complicated than he realizes. Well the same applies to marketing, sales, etc. Just because you don't know or understand the work that goes into successfully marketing a software product doesn't mean it's trivial. Here is a great article that discusses the many tasks a non-technical co-founder should be doing. As you can see, it's much more work than one person can handle.
To answer your questions
Is it usually worth it to partner up with a non-developer in an IT startup from the get-go? I definately see the benefit of bringing in some tier-2 MBA types, but what about the initial founders?
It depends. It depends on what the startup is and on what your goals are. If you want quick growth and VC funding, then you should get a business co-founder to market the heck out of your product and spend time courting VCs. If you are content being a Micropreneur with a small, niche product, then you may not necessarily need a business co-founder. If your goals fall somewhere in between, then in most cases you should consider a business co-founder.
Note, however, that regardless of which direction you want to take, business type tasks like marketing are necessary and crucial. It's just a matter of whether you want (and have the knowledge) to take on the responsibility of tech and business all by yourself, or whether you want someone else to work on this fulltime.
It seems to me that it would almost always be a better idea to partner up with another dev and then share the non-dev responsibilities rather than take on 100% technical burden because such a setup spreads the risk/burden evenly. Am I wrong?
If this is an equal partnership, you are spreading the risk evenly. The roles each person has isn't going to push more of the risk on one versus the other. What matters is what the outcome for each person would be if the startup fails. In an equal partnership, if the startup fails, you both fail equally.
The both of you need to sit down and hash out in detail what each person's responsibilities will be. None of this vague "you handle tech and I handle business" nonsense. Break these roles up into smaller components. (This will also serve as a roadmap for your startup.)
In doing this you will probably see that his role is a lot more involved than you realize. If after having this discussion it's clear that one role will require a lot more work than the other, then you guys can negotiate an equity split that is fair based on the amount of work each will perform. If after all this, you still can't come to an agreement, then walk away.
What others think
Bob Dorf thinks every founding team needs three members: a hacker, a hustler, and an artist. See here.
Noam Wasserman's research found that only 16% of tech startups in the US are solo-founded. See here.
Y Combinator is not a fan of single founder startups. From their FAQ: "A startup is too much work for one person."