Reverse engineering Google thought process in developing Google Web Accelerator

What was Google think-tank smoking when they developed GWA?
They developed a product which provides further speed to already very fast users ( in Bengali we say it is akin to applying oil on a well oiled head ), doesn’t anonymize (anonymization adds value to a proxy) well, over-optimized without concern for destroying web analytics, loading web servers or breaking web applications etc.

Lets reverse engineer the thought process that went behind closed doors by the famed think tank (lets not forget the Phd’s) in Google. A “thought experiment” as Professor Einstein would say.

I think the key meeting went something like this:
Big Boss: We need to capture users detailed surfing habits, everything they do, when they do etc. Lab demands it for targeted search, advertising and other ideas (hush hush).

Architect Phd: As we cannot install a Trojan, being “do no evil” [smirk] and all… the only way I can see is to route all their traffic through our proxy. We provide them a free proxy service, maybe some anonymity.

Marketing: Only pseudo-anonymity. Anonymity makes bad press. At least provide a way for eager webmasters to find out the actual users.

Sales: What does it buy them? Why would they use it?

Architect Phd: Umm… faster speed of web access. We can provide faster speed to dialup modem users by compressing data, caching get requests, pre-fetching (interrupted by marketing as it was going way over his head)

Marketing: No way. Dialup constitutes a declining segment of the market. It wouldn’t go well with our investors. Google should target the emerging and rapidly growing sector, broadband users. Why don’t you make it faster for broadband users?

Architect: But they are already way too fast…