The USV Blog Search Engine

It has always been possible to search AVC. You click on the search icon in a desktop browser or you click on the menu button in a mobile browser.

But there is another way to search my blog posts, both here at AVC, and also the ones I have written on

With our recently launched refresh of, there is now at tab at the top called “Writing.” It looks like this on a mobile phone:

You can search by type (USV blog only, team member blogs only, or all), topic, author, and date.

This search engine includes writing by many USV alums on the USV blog and all of the current USV team members who blog regularly. It is quite a library of content, mostly on tech, venture capital, startups, and that sort of thing. But naturally it veers into many other topics from time to time.

If you want to read what USV team members (current and past) have to say about something, there is now a resource to do that. And we hope to make it even better over time by improving the metadata and search functionality around this large library of content.


Comments (Archived):

  1. jason wright

    The Strange Death of Europe, and The Madness of Crowds, both written by Douglas Murray.Possibly worth reading when in Europe…on a wet and windy afternoon.

  2. William Mougayar

    Is the tagging done manually or automatically then reviewed for accuracy?The content is very structured and that yields more accurate search results.

  3. jason wright

    ‘You are what you read’.’You are what you blog’ is the new?

  4. sigmaalgebra

    Looks okay up to comparatively good!Of course, people read and post for meaning, but the search is via words. It’s a big step from words to meaning.E.g., let me guess, if one does a search on key words”CO2″ and “climate change”then they will get lots of posts by Albert, posts with the meaning that human sources of CO2 might in the next 100 years have a significant effect on climate. How might one use key words to search for posts with the meaning that the main causes of climate change have nothing to do with CO2 and that in the next 100 years human sources of CO2 have no chance of causing significant climate change?The problem is, for the two searches, agree with Albert or don’t, the meanings are very different but the key words are much the same. How ’bout that!That is, *meaning* is what we want; key words are what we give; but key words are from a crude down to a useless way to characterize meaning. And this situation holds, as in this example, even for Internet content based just on text, that is, words!For Internet content based not on words, e.g., still images and instrumental music, along with some others, the “situation” (especially including discovery and recommendation and not just search taken narrowly), i.e., the weakness of key words, is worse. And when the *meaning* is tough to put into words at all, e.g., *meaning* of art or personal artistic taste, the “situation” promises to be still worse. And for recommendation and discovery, we tend to be short even on input of even just key words. Hmm ….Disclosure: No way did I write this post just off-hand; instead, for a solution, i.e., search for *meaning* of Internet content, math derived, code running, alpha test on the way. New office is looking good, but just had another unpredictable, exogenous interruption in the work.

  5. Richard

    When a VC website uses 20 yr old search techniques, you begin to understand that the last 20 years of VC success was likely more about luck than skill.

    1. jason wright

      What techniques would you be using?

      1. sigmaalgebra

        For “techniques”, as below, try out my alpha test once I announce it.

        1. jason wright

          I will. Below? Where?

          1. sigmaalgebra

            At my post…The alpha test is coming but, sadly, is not ready yet. I’m still working on my recent move and office.I intend to announce the alpha test here at AVC, on BRC, on Hacker News, via a friend on Facebook, etc.The applied math appears to work as intended; same for the corresponding software for the math (users won’t see any math) and the Web site. Have the first server running.Need to get interruptions out of the way, finish my new office, a lot better than the old one, add some data, e.g., on newsletters and blogs, get a static IP address and domain name, …, just routine things.Maybe people will like it; stranger things have happened!

  6. JLM

    .You may have some kind of circular short circuit. When you start from the left to right and pick a topic, and then an author, the selection of the author makes the topic box go to “ALL.”The search function is very fast.JLMwww.themusingsofthebigredca…

  7. awaldstein

    I subscribe to two USV blogs via email, yours being one.And only infrequently search the USV site.Just tried it–pretty damn quick.

    1. sigmaalgebra

      “Quick”?The whole AVC collection could be stored in what, 1 GB (gigabytes, billion bytes)? 10 GB? If so, then, SURE it’s “quick” — the whole thing is in main memory!Okay, okay, maybe the search engine doesn’t just read in the whole thing and, instead, uses some relational database software, maybe MySQL? But it’s long been standard for relational database software to cache heavily, be a total hog on use of main memory to cache.But for more, the database is nearly all just read only so could be on solid state disk (SSD) with no concern about wearing out the disk, and relational database off all SSD stands to be a whole new world. And, keep in mind that for some months, maybe a few years, could get 14 TB (trillion bytes) in one small form factor, maybe 2.5″, not more than 3.5″. And long standard functionality of relational database is to permit even single database tables to span more than one disk drive. So, have a tower case with, say, 5 of the 14 TB SSD puppies — might get ALL of Disqus in that!!!I know; I know; some of us do a lot of typing into Disqus, but 14 TB is a LOT of keystrokes by each of a LOT of people over a LOT of weeks.Yup, “quick”, just from routine usage of current hard/software.

  8. sigmaalgebra

    > One of my co-founders told me it’s [e-mail newsletters] his sole place for getting news.Terrific! It’s happening!!!! Use the Internet and focused media to split, finely partition, and pull out the audience of the one-size fits all big media!Terrific — I’ll make a special effort to have such newsletters covered in the database of my search engine! Thx! AVC and Disqus just paid off!

  9. Mike

    Thanks, good to know on search capability. When is the book coming out?

  10. Matt A. Myers

    Inspired by Techcrunch?