Same thing happened in the earliest days of the automobile industry: wide experimentation with body types, shapes, transmissions, engines, and so forth.
Over time, biological and technological evolution sharply reduced the diversity of these early days. Creatures and cars settled into modest changes on the most effective patterns.
The same thing is happening today in search -- Yahoo's new search platform and MSN's recent AdCenter launch are incredibly similar to Google. Not a little similar. Not moderately similar. Very very much similar. What do I mean, you ask? The way Y and M handle adgrouping. Ranking by quality scores. Creatives. Bidding. Quota. Services. The WSDLs.
The industry has rapidly "evolved" (converged? infringed?) to the Google model, for better or worse.
In many ways, this convergence is good -- it makes life easier for agencies and advertisers alike. The platforms are increasingly similar under the hood: this simplifies things for the engineers, for the client services staff, for the accounting folks. (Though beware Google's T&C's: "Thou shall not co-mingle Non-Google with Google" and "Thou shall not place Non-Google on equal footing as Google", see Secs. 2c, 2d, 2f)
But, looking at the situation more broadly, I find it curious that Yahoo and MSN opted to follow Google, rather try something new, something different, something more innovative.
Yahoo and MSN are both significant enough that whatever crazy scheme they devised, the industry would certainly implement the API and learn how to use it effectively. Sure, less risk to follow convention, but potentially greater upside gain by challenging it.
The standard standard search advertising model (the Google model) is certainly solid and working well. I just don't think search, search advertising, and search APIs are solved problems.
When will a major engine next take the risk to bring diversity and experimentation back to search?