Disagreed.

The old introduction may be obscure, but at least it is informative. A visitor can follow the links, read 5 minutes about biases (if it is required) and then he gets some understanding of what this site is actually about. The new version is much more vague. Who would like to think more clearly, improve their lives and make major disasters less likely? Well, pretty much everybody.

I don't think that minor changes, like rephrasing introduction or adding disclaimers about criticisms to FAQ will have any noticeable effect. To attract significantly more people, you should write actual content, like HPMoR or the sequences.

Personally, when I recommend LessWrong to friends, I tell a few words about what it is, so they don't actually have to read introduction at all. Or, even better, link directly to the sequences.

Anyway, remember that one data point is one data point. People in marketing have their A/B tests for a reason; now, it's unclear whether the new introduction will attract more people because of its apprehensibility or less because of its vagueness.