I know mouse capture seems like a security concern to many browser vendors, but just have a dialogue to agree to mouse capture and a clear use of ESC to exit it. If they don't do it, some plugin will and we'll be set back into another X years of flash dominance in online gaming again.
At least with WebGL you have multiple groups working on solving the major issues. I don't see Flash fixing these things alone.
There also Silverlight that will come out with their own solution but like with Flash I don't see it as an easy win over WebGL.
I'm not sure this falls under the "advantage" column in this scenario. Number of "groups" doesn't count for anywhere near as much as "number of people", and there's no particular reason Adobe can't field as many people on their one side as the other interested groups can in total, unless one of those groups really commits. Having a lot of big names on the list doesn't guarantee than any given name has more than "a guy we told to subscribe to the mailing list and scream if somebody proposes something that doesn't serve our interests". (That's by way of example. I know there is non-zero real commitment, my point is that it isn't at all the case that this group must outweight Adobe's ability to solve the problem.)
The real competitor here are native mobile applications.
Also, I have to wonder about the security implications of WebCL (not the C not a G), like malicious sites using idle GPUs for password cracking. That sort of thing.