There are a lot of non-valuable ideas out there, and the ideas in Mike’s presentation are good examples. However, that doesn’t mean that there are not also valuable ideas.
Making an idea valuable is mostly a matter of detail. The idea needs to be thought through enough that execution becomes trivial. If hiring a good team of developers would produce a successful business, the idea is valuable.
Mike’s ideas are the sort where hiring developers won’t get you anywhere, because there is a nearly impossible step 1 to get past before you can get to the benefits of the idea.
Community algorithm development on sites with established datasets has a lot of potential, but step one is to have a site with a better algorithm. The better algorithm doesn’t exist on day one, and because existing sites are good enough for most users, you won’t get enough users to make algorithm improvements worthwhile. It’s worth noting that the reason for Google’s initial success was not so much the improved algorithm but the uncluttered search page - the better algorithm isn’t immediately obvious, but anyone can see a fast loading page and come back for that. Then of course there’s the whole evil users thing to deal with.
Personalizr has different problems - but it’s still the same issue of day one value and user adoption. You start out with no value (because no data has been added to the system) and probably bad publicity for privacy issues.
Support.com is somewhat closer to a valuable idea, although it would take a lot of investment to get the dataset up to a useful standard. I generally don’t think much of services that can only make money from advertising, though I do like the idea of targeting ads at people who are annoyed with a particular product.
The other idea linked in the Venture Hacks post is the shadow market - It’s got potential, but there’s no way of proving the idea is any better than existing funds management methods until well after launch. A large amount of money is needed to start, which nobody will provide - those who usually invest in managed funds won’t give their money to someone with no track record, and VCs, who don’t mind the risk, won’t be interested because they expect higher returns than you will get even if it goes well.
Fortunately, non valuable ideas can be turned into valuable ideas - basically by adding more ideas. Maybe you come up with a search algorithm that is better for finding support information and you can launch support.com right away without having to put together the data manually. Maybe you find a way to start a managed fund without needing anyone to trust you with lots of money. Maybe you design a service people will pay for that complements the free service that was your original idea.
Once you have that valuable idea execution is the easy part. It may not quite be trivial, but there is at least plenty of advice available on how not to screw it up.