I waited almost 4 hours for the response...

I recall seeing it when I registered AFTER I read their email...  

And since I am paying from my own pocket, I will think about paying an extra $200...

Absolutely. I've upgraded mid-course so I'll be getting my courseware on the 21st and I'm getting my new labs allocated on the 21st as well. Given there's additional focus on Buffer Overflows, Web Apps and modern operating systems I definitely see the benefit in upgrading.

If it's any issue, I've been warned that Core Impact won't be installed in the new Version 3 XP boxes, which implies it may have been dropped from the syllabus.

I quizzed the sales team as to when the V3 syllabus would be published and they said it'll hit the web site next week.

I don't think it will really be missed if it is dropped from the course. The course really only gave an overview of it as a pentesting framework.

http://www.offensive-security.com/blog/backtrack/pwb-v3-information-security-training-at-its-best/
its recent.

also has demo video.

Are you asking why metasploit stuff might be necessary, or why they removed it?  Not having taken v2, and preparing to take v3, I can't speak for why it'd be removed.  It's good knowledge to have, as you really need to understand, for instance, how to take code from a 0day or other exploit, and compile and use it, with meterpreter or others, against your targets.  

Based on them saying 'kernel payloads and db_autopwn' were what was removed, I'm not so sure they removed metasploit, entirely.  The kernel payload section... I'm not sure what exactly he used to cover, so maybe it just tended to be something a bit harder to grasp and took more concentration away from other topics, etc.  Or it could be that with many of the other attack vectors that are coming to be, muts and company wanted to put attention and focus on other things.  I won't speculate.  

db_autopwn would take some of the learning out of the picture, by automating steps and functions that you really SHOULD understand, to be an effective pentester, and not always relying on pre-canned tools and scripts.  The same would go for the removal of CoreImpact.  As KamiCrazy mentioned, Core is so fully capable of many things, that using it in such a class could very easily lead you to become complacent / lazy, and not truly get the value and understanding that comes with manual testing procedures.  

I know that I, personally, would MUCH rather understand the how's and why's of what happens underneath such tools and scripts, than to lean so heavily on commercial tools or pre-compiled scripts, etc, as it's much more beneficial and useful, in cases where, for instance, a customer doesn't allow you to bring your own machines and code into their network, but rather, you need to scrape together 'available' tools and ideas to do your tests.  You'll be a better pentester from the experience, by learning things the hard way.  Once you truly understand these things, then using other commercial tools and scripts becomes a timesaver, not an exercise in laziness, and you can better test and validate findings from those tools, by using other knowledge and techniques.

They do have http://www.offensive-security.com/metasploit-unleashed/  (love it)
I hope they have full version of this course soon.