That said, I'd prefer if they had used plain C, which is even more portable.
First, blaming a tool is lazy. While there are problems with C/C++ much of this is due to practices and code complexity. There are tools to help us humans with both of these things. Additionally, do you trust Cython? Why is that (or any other language written in C) better than C? Have you audited that code yourself?
Second, if the developers are as diligent about testing this new browser as they have been with everything else they've done, I would trust it as much as I do Google's Chrome (where I don't trust Google at all) or Apple's Safari (I feel them to be only slightly better than Google).
Third, and perhaps most importantly, if it doesn't work for you that is completely fine. It's not your computer or your data. Why denigrate the project and those who use it or work on it?
For those who think Mozilla isn't about money:
In zamadatix's reply to your message he said that C++ is only being used to bootstrap the language and that it's certainly not the plan to require C++ forever. I imagine if Jakt thrives, he will be right and Jakt will probably start compiling into a lower level language instead. If they do target a lower level language though, I imagine this will result in Jakt being less portable than C++ which has been around much longer and is more established.