Except I think startups are collectively figuring out that creating a company solely around an API plugin isn't a very solid business plan.
I define a power user as someone with technological sophistication. For example, one of the features Matt Cutts mentions is filtering your mail using regular expressions. That requires technological sophistication.
Being more comfortable with computers, technologically sophisticated users are more likely to google around to find free or cheap alternatives to a product they see. Thus, companies in a market with technologically sophisticated users compete more on the basis of price and less on the basis of advertising than other companies. This makes it harder for these companies to make a profit.
It's true that technologically sophisticated users will be more likely to see the merits in Matt Cutts' startup idea. But that's a defect, not a feature.
[edit: a famous bit of company-is-a-feature carnage for reference: STAC vs. Microsoft: http://en.wikipedia.org/wiki/Stac_Electronics ]
Which makes me wonder why Matt can't get this championed internally or for someone's 20% project?
Startups need to solve difficult problems or they are nothing more than marketing firms trying to create a hype around hot air. Filtering email intelligently, including automatic anonymization of sensitive parts within the email, can be a difficult problem to solve, and I think there are many types of email integration that a startup could tackle based on the approach Matt suggests.
It doesn't seem like it's terribly impacted the other players, at least that's my perception. (could be wrong - someone with hard stats pls correct me)
My point being: that fear alone while is a rational one, should not be the only criteria to definitively say that an idea should not be pursued b/c it guarantees that you will fail in your startup
The danger is that when your app can be absorbed into something more popular. They don't compete with you — they just destroy the market for what you do.