Why don't you A/B test your theory?

It is tempting to think that a better mouse trap does not need marketing. But a better mouse trap + marketing beats a better mouse trap every time.

Landing pages are the "salesperson (or salespeople, sometimes it makes sense to have multiple)" for the product. They are absolutely needed in some form or another for a large majority of products - "products" that many of us are interested in don't need them as much, we've heard enough about them elsewhere to know we at least want to take a test drive. The majority of the web population isn't like that. They need to be convinced to take that test drive. The landing page can be a "used car salesperson" kind of page (and I generally hate those, but they work for a lot of types of products) or a "soft sell informational" kind of page; either way, they need to have good closes and other hooks.

The difference between a good and bad landing page is the difference between success and failure.

The product is the thing, and the product can be viewed differently to different people in different regions with different motivations. A landing page shouldn't dramatize anything. Businesses need landing pages to allow for segmentation and to continually refine your message based on the audience. I have yet to come across a marketing agency that recommends against the use of landing pages.

I don't know if that works for all types of products, though. Some, perhaps, but not all. A B2B landing page, for example, needs to be able to convey the benefits of using the product, solutions, etc.

Many times on HN someone will show off their thing, and I have no clue what it does.

(I found the best way to deal with this is to simply suggest they make it easier.)

I can't imagine what, say, Facebook or Twitter would be like if you were just dumped into a "temporary account" profile before you signed in. It wouldn't make any sense.