Seriously?! If search engines had AI, I can see it being cool/innovative. With the current state of search technology, it's a usability disaster.

Problem #1 - language. Is their portfolio page called "portfolio"? Or "companies we've invested in?" Is their contact page called "contact"? Or "Get in touch?" Do they have a blog? Searching for "blog" seems to indicate that they don't-- unless it's called something other than "blog". Search is hard and they are making their users play "guess the right search query" when a scan/browse action would be easier and less error prone.

Problem #2 - scan-ability. So say I AM interest in learning more about their portfolio. Roughly what types of companies do they back (what's the search query to answer that question?)? What have they backed most recently (again, what search query?)? Who works there (a search for partners suggests "Energy (partners)")? Scanning is a huge part of web behavior.

Problem #3 - discoverability. What about what I'm NOT searching for? They have some GREAT content around how to pitch VCs, what ideas are VC-worthy, etc. With a content-free page, there is a 0% chance of me learning about stuff that I didn't know I was looking for.

Problem #4 - SPELLING. Take a look at their partner page (if you can find it). Are there any names there that you might have trouble spelling/searching for if you heard it at a dinner party last night?

We don't have to guess with usability. What they SHOULD have done is write down 5 tasks that people might do at their site and had a few friends run a few informal tests.

I'm willing to turn in my UX designer badge if my guess is wrong and this UI increases task performance speed and reduces errors.

Hrm, as I consider it-- I guess your comment doesn't dispute all that. It probably can be cool, innovative, AND a usability disaster.