Forget about C++. Learn a high level language that has the libraries you need for building an application quickly. Someone suggested Ruby, and he's not wrong. Python is another good choice, though it doesn't have an obvious "guided path to success" the way RoR works for Ruby. PHP will allow you to find a lot of potential co-founders, but it's probably not worth it...at this stage you don't have the know-how to recognize a good PHP developer, and a bad one will simply kill you (PHP is perhaps the best language for writing horrible code--the majority of stuff out there written in PHP is really frighteningly bad). That's not to say there aren't bad Ruby and Python developers...but the barrier to entry seems high enough to keep out most of the riff raff.
Find a co-founder who likes your ideas and trusts you'll get up to speed on whatever bits of the equation he/she doesn't want to handle. In a startup past the very earliest "get-to-prototype" stages there's a roughly even split between the sucky non-tech stuff and the actual development...you could take the non-tech stuff and leave the hacking to your co-founder. I'm quite technical and have developed a lot of software in the past ten years, but I found an even better developer and took on the sucky stuff while leaving him to handle most of the development. It's working very well (though I often find myself wanting to get back to real "work" when I'm doing the boring accounting/marketing/taxes/funding/support/etc. stuff--I have to remind myself that if we reach our goals I'll be able to hack anytime I want, in between rides on my robotic dinosaur, of course).
There're two possible questions in the title, and I'm not sure which one you mean. Are you asking how a non-hacker can start a tech startup, specifically one focusing on the web? Or are you asking how a non-hacker can start any startup, be it biotech, finance, auto-body, housepainting, or some other industry entirely?
For the first question, you really either need to be a hacker yourself or convince a top-notch hacker to work for you. You can't expect to start a successful tech startup if you don't have good tech. I listed some ways that entrepreneurs repel potential tech cofounders here - http://news.ycombinator.com/comments?id=13000 - and also suggest some things you might bring to the table that might convince a techie to work with you.
If you're just looking to start any startup and don't care about the field, the rule is always "Concentrate on your strengths." Basically, you want to go into business being more passionate about the particular project you're working on than anyone else on earth (or in your location, if you have geographic barriers to entry). That's what'll convince customers to buy from you rather than a competitor.
If you're not a hacker and you're trying to start a business that requires a lot of hacking from the outset (such as anything computer-related), you're going to have to let go of the idea that this is going to be "your" startup. The beginning or a company's existence is the most risk-laden period, and it's also the period where the value you can provide is far less than the value the people actually writing the code will provide. You have to convince hackers to come on as co-founders (not employees), and compensate them accordingly.
The other option, as others have suggested, is to learn to code yourself. I doubt that you'd be able to do a good job of coding something up yourself without anyone with formal computer science training backing you up, but learning never hurts. If you know that starting a tech business is something you want to do someday, I'd say you should start learning how to code now. Personally, whenever a non-hacker tells me about something they want to start, I can't really see what value they're bringing to the table. Most ideas aren't that original (and don't necessarily have to be to be successful), so if they can't actually make the product, what good does working under them do for me? Business-types are definitely useful, but not until there's at least some sort of prototype.
I've never actually done a startup myself, so this is really just my personal intuition of the way things should work.
One of the first American Billionaires once said he could answer any question asked. He called in a meeting and ask each journalist to come armed with top notch questions (of course industry related ), which they did. The day of the meeting they were surprised to see 16 phones at the Great Entrepreneur 's desk. For each question asked he pick up a phone and call someone who then, somehow, provided the proper answer.
If you cannot code, need to start a startup and do not know the moral of this story, then you probably should go sit back and think for a few weeks before jumping into the middle of the ocean.
It's a little different when searching for technical experts, however, because it's often not possible to tell a technical expert from a technical nincompoop without a solid background in the subject. There are a lot of folks out there who can play buzzword bingo with the best of them, but can't actually sit down and code something.
Of course the problem is, not being technical you are always prone to the 'Warren Harding' error. [0] You know, someone who dresses badly, uses a mac notebook, talks real fast about technical things that has all the attributes of the hacker except they are crap at building product.
Some suggestions for finding a co-founder can be found reading, Students guide to Startups. [1]
Reference
[0] Warren Harding was 29th President of the United States and had the distinction of being pushed forward because he had the perceived traits of a leader and turned out not to be so good. I picked this up reading the Malcom Gladwell book 'Blink', 0-7139-9844-X.
The billionaire is Henry Ford. The version I was told was that he was on the stand, and that was his reply to one of the prosecutor's questions. Jobs also liked to call himself the "Henry Ford of Personal Computing".
Well back in 1984 four guys decided to start a software company called Nantucket. The two developers first ever C project was writing a dBase compiler. That compiler became a full fledged dBase dialect that even spawned an active third party tools market. Not to bad for a first project.
i am working on a startup at the mo but am not technical, we got someone to build a proportype (by paying) but realise if you wanna iterate and create a decent app, you cant keep paying people. I wonder how kul and harjeet YC Alumni managed (as they are not coders)
the solution to my problem is simple
a) im learning how to code from the ground up xhtml, css, php, mysql and then prob move on to ruby (this will take many months if not longer)
b) im looking for a technical cofounder, problem is, they all have the 9-5 mentality.
I majored in Politics as well, maybe i need some Joe Kraus magic!
As someone with a bit of technical skill: when the greatest motivation to work on a project is the money you'll pay me, I have a 9-5 mentality, because that's what you're paying for.
When the greatest motivation to work on a project is the insane coolness of the project itself, things are different. Then, the money is just a bonus.
Can you figure out something about your idea that hackers will find cool?
yer, definitely, and thats how im sellin it, in fact ive sold it to one technical person and he is willin to come onboard. prob is our site is php he is a c# man. i think he will be able to pick it up though. its quite strange really, i work for a technology company but there are no coders! lol
I am in the same position as you. All suggestions here seem reasonable but my advice is don't wait until you have a hacker cofounder. Use rentacoder and try to create a prototype.
I'm using eLance - this is just to get something started and then I'm hacking the features I want myself as and when I learn...oh yeah, and then get a co-founder with an unfeasibly large brain who can help you out!
Rentacoder and eLance are neat ideas, but they have the same problem that the job market as a whole does: you want good code from good hackers, and if you're not technically inclined yourself, you will have a very hard time determining who is a good hacker and who just talks the talk.
If you want a technical startup, you need a technical cofounder. "I'll just outsource the development" will doom you to failure. I know this because I am an employee of a startup that thought that way, and learned the hard way that execution matters; and spending umpty-million dollars in the first year on outsourced projects that you wind up needing to pay umpty-thousand-dollars in the first year of operation to replace -- well, let's just say that, even with venture capitalists involved, burn rate matters, and money you spend stupidly still counts as an expense.
Find a co-founder who likes your ideas and trusts you'll get up to speed on whatever bits of the equation he/she doesn't want to handle. In a startup past the very earliest "get-to-prototype" stages there's a roughly even split between the sucky non-tech stuff and the actual development...you could take the non-tech stuff and leave the hacking to your co-founder. I'm quite technical and have developed a lot of software in the past ten years, but I found an even better developer and took on the sucky stuff while leaving him to handle most of the development. It's working very well (though I often find myself wanting to get back to real "work" when I'm doing the boring accounting/marketing/taxes/funding/support/etc. stuff--I have to remind myself that if we reach our goals I'll be able to hack anytime I want, in between rides on my robotic dinosaur, of course).