Graham talked about the ways in which the obstacles to starting a new business have declined. (He has written up the talk on his own website, here). Anyone with a smart idea, some technical skill, the Ruby on Rails library and an internet connection can create a start-up business, as the costs for hosting and marketing tumble. There will thus be more, younger and wackier start-ups on the market. Not all of them are going to be successful: the world doesn’t need 10,000 new internet products. So what will differentiate the survivors from the also-rans? We might hope that it is quality, but other factors are likely to have an influence, one of them being geography. From Bobbie Johnson’s account of Graham’s talk in the Guardian Tech blog:

Will we still need Silicon Valley? Yes. It’s true that you can start a startup anywhere, but you also have to make it succeed - which is more likely to happen in a hub. The increasing cheapness of starting up might actually make it more necessary to be in a hub. The benefits are face to face meetings; visiting a friend down the street; happenstance. The question is not whether you need it, or whether it offers any advantage at all. Your competitors can take that advantage. This may be an uncomfortable idea.

But, a startup that’s just a couple of people can be moved more easily. We make people move for Y Combinator, the advantage of working face to face for a few months outweighs the cost of moving. This means that seed funding can be a national business, not a local one (unlike venture capital). Is seed funding international? If it is, it’ll be harder to set up new Silicon Valleys. All you’d end up with in your local one are the ones who couldn’t be driven enough to move to the real Silicon Valley.

Graham’s ideas are related to Michael Porter’s theories around industry clusters. If people in the same town are all doing similar sorts of things, or things to meet the needs of those people, then they will be more successful. Start-up resources have a network effect - the more they are connected and the closer they are connected, the more power and value they will have.

Interestingly, this is one reason NMK was founded ten years ago. Because London is a digital media cluster, and we believed that we could create events and a knowledge-bank that would allow more knowledge-sharing to take place between companies in this sector, to support the natural networking and interaction that would happen organically.

I am not entirely sure whether London is an internet start-up cluster, though it’s something that we’re keen to support. The recent successes of companies like last.fm and the emergence of start-up networking events like OpenCoffee would suggest that there is reason for some optimism.

On the other hand, while London’s venture capitalists (e.g. 3i, Ariadne, Index, Advent) seem keen to invest in home-grown talent, my impression is that the same is not true of the Angel community (characterised to me by more than one entrepreneur recently as "white-haired men who want to play at still being in business, to have something to talk about at the golf club"). This is why the Seedcamp initiative has been an interesting development recently and its development into a home-grown alternative to Y Combinator is well-worth pursuing. If Graham is right, and start-ups need seed funding not venture capital (and they all do, in the first instance, at least) then this is surely the industry’s highest priority.