Valve: This makes the least sense to me. You shouldn't enter a highly competitive market (like games) as a side project.
Twitter: I don't think all that many people care about the real time web. But even if they did the question becomes "Why couldn't Google create a search engine that crawls Twitter every few minutes?" They don't need to own Twitter to make a real time search product from Twitter's data (which is all public after all)
StumbleUpon, Last.fm & FriendFeed: In all these cases the author assumes there's some value in knowing what the users of these services find interesting. But the truth about all of them is they have a very small niche audience. Meaning if Google used them they'd be skewing their results to a small audience while possibly alientating everyone else.
Also, even if StumbleUpon provides targeted advertising, it's still quite likely to be very poor-performing traffic. If you're exposing people to things they were unaware of before, then, by definition, they're going to be very early in the sales process, and the conversion rate will be abysmal.
Because of StumbleUpon's very low costs ($.05 per unique visitor) for targeted traffic, their ads, which aren't even marked as such, can be a really good deal depending on what your site or product is.
BS. I think Google is going to slow down on the acquisitions. Based on the ratio of projects they buy that die, I think they're going to begin pushing for more internal development and less acquiring outside technologies.
Of course if a site is a sensation and has an established community in a short time that may be worth purchasing but overall I don't think we'll see Google going after any of these.
Valve: This makes the least sense to me. You shouldn't enter a highly competitive market (like games) as a side project.
Twitter: I don't think all that many people care about the real time web. But even if they did the question becomes "Why couldn't Google create a search engine that crawls Twitter every few minutes?" They don't need to own Twitter to make a real time search product from Twitter's data (which is all public after all)
StumbleUpon, Last.fm & FriendFeed: In all these cases the author assumes there's some value in knowing what the users of these services find interesting. But the truth about all of them is they have a very small niche audience. Meaning if Google used them they'd be skewing their results to a small audience while possibly alientating everyone else.