Search Me's idea will work about 10 years from now when everyone has a fast enough computer for it.
Unlikely - it will always take longer to look at screenshot of webpage to decide if it is the right search match, this means spending more time on search site rather than quickly finding the right stuff and going there.
This may work for obvious searches that produce obvious result (ie: google), but try that with long tail and it won't work well.
Spending a lot of money on servers and buying in traffic was not exactly long term idea.
Then again, if you don't do that with all hype then you won't probably raise $50 mln. This ought to join the same category as boo.com (check Wikipedia)