It does seem like some older browsers don't... but then I'm curious about spiders and email image proxies and the like. I'd be very curious to see what breaks if you just completely stop all HTTP 1.0/1.1 traffic.
I plan on conducting a field experiment to determine if it's a net-positive or net-negative activity. My hypothesis is that HTTP1 is largely a risk with no benefit when coupled a bypass for search engines. I feel uncomfortable making claims until I can prove its the better move.