Is experience or skill more important when creating a web startup?

Hi, I am interested to know what the people here think about this topic. I have added some context below to frame the discussion around something. It is a long. Apologies if you fall asleep before getting to the end ;) I have been having a debate with some grad-students/MBAs at Oxford regarding a web startup they want to do and about the need for experience vs skill in the developer(s). About them: "We are a group of postgrad students (Oxford) and we're currently loking into setting up a web 2.0 startup. It will be a social networking portal, focusing on a user base of about 20million in the EU plus about 20million in the USA, which has not been covered yet. The portal will also include a number of novel features and will be unique in its composition. We are about to finish the business plan and already acquired high-profile consultants for the board and established first VC/BA contacts. Due to the central importance of the role (CTO) we require something along the lines of 3-5 years of experience in PHP/MySQL and scalable architectures." I can't talk about what it is they will do or the novel features etc. so please don't ask :) I thought I would play a counter argument just to see thier stand on this position with the experience!=skill approach stating for my $0.02 I'd go with someone who is passionate and interested in what they do and is able to learn fast and deliver rather than the fact they have a few years experience. Although experience is important in many respects (essentially because you know people will have "practised" what they preach and learnt from past mistakes and teachings), in the startup tech world it does not matter as much. Making things people want and having rapid develop cycles where you can prototype, learn, test, and implement features on a fast (day-build) basis and based on an easy to use architecture are more important. The top tech companies in the world have been started by people how had little experience (and were actually working with the latest tech at the time with which no-one had much experience over) and in many cases still are. They replied: " I agree on your point w.r.t. experiencs!=skill, but I have to add that in an time-critical development process experience just adds thats extra advantage that might lead to your product reaching the market just before the competitors' does. Experience in a relevant field just means you can actually employ the knowledge gained from previous mistakes - particularily in the fast growing web 2.0 environment no project is like the other so there are many mistakes to be made. Also, especially the understading (and subsequent realisation) of scalability is key in our project - we just can allow for our service to break down because of an inflexible and limited initial architecture...when it grows, it will grow exponentially! In the beginning you might be fine to run an immature architecture till, say 100k users, but as soon as the viral growth aspect of SNS kicks in you will have garanteed downtime and most likely you'll have to redesign your system and shut down your service for a limited time - giving your competitors a lead... " I agreed with this as time-critcaly kinda changed things and replied: " Experience is advantageous but not a neccessity. --- Absolutely. If you do not learn from history you will repeat its mistakes. But... thankfully we are in 2nd/3rd cycle of internet app development and hence have many a case study, best-practice guide, design pattern, tried and tested toolkits and more to help us. We can now stand on the shoulders of giants. But, just a thought, interestingly enough many a leading tech company still do not weigh experience too heavily when hiring/building. This seems especially true in recent high profile tech startups that are doing very well. The simplest solution (in those cases) seems to be to blend experience with inexperience to bring fresh thinking to problems and challenge assumptions. This seems to be a pattern to making disruptive technology? Scalable Architecture = it goes without saying. --- Everyone has faced the same problem at one point or another, and whether you go with a .NET or J2EE or LAMP stack (or other) , there are common patterns that resolve many of these issues and are tried and tested by many a startup/company that went from 0-millions and beyond very fast. Only when you are doing something truly new, will you not find a comparable. " Although I was being devils advocate to an extent I have been thinking about this more deeply this morning and thought I would post here (albeit in a long-winnded way but-accurate-to-the-conversation way) to see what others in the same boat have to say. Thanks, Babul.