Posted by Josh Webb

In business and technology alike you often are thrust into a project, engagement, or dynamic that is wholly foreign to you. You won't know what you don't know and you may not have the foggiest clue about what line of questioning or research might help mitigate your ignorance. This situation calls for a topical deep dive. The goal is not to become an expert but rather to gather enough information to:
  • Understand some of the basic vocabulary
  • Develop the basis of an intuition for how the new microcosm might function
  • Apply pattern-matching from previous deep dives and areas of expertise to make informed assumptions
  • Identify people in your network that might be less ignorant than you
  • Clearly articulate what you do (and do-not) know and what items from the latter list you would like to move to the former

Because the 5 bullets above are generalized concepts, we'll walk through it once with something specific. Let's take a butterfly valve for example. (HAHAHA, yeah right! not a chance)... "Deep learning" is good example of something we can research. It is a complex concept, in a very dynamic space, poorly understood by laymen, and generally a dark art for all but the most dedicated software geeks.

Let's see how we do with an exampling diving into the technical frontier of ‘deep learning' (follow along):

A Google Search for Deep Learning suggests a Wikipedia page (Click the link and check it out before proceeding; Try to Interpret what you see.)... Uh-oh. It's a tough one:



But let's persevere and see what a lay person can glean from this:
  • We see the phrases "Machine Learning" and "Neural Networks" and "Algorithm" repeatedly (all three of which have far more useful Wikipedia page introductory paragraphs (you clicked on those right?))... And since we're researching; resolvable curiosity links get opened in a new browser tab!
  • The Applications section of the Deep Learning Wikipedia page seems to be readable by someone without a Ph.D. Real-life usage examples are great; this "Deep Learning" algorithm seems to be good at a certain class of problems. Problems that you intuitively understand, even if you do not understand the how just yet.
  • The Applications section also mentions some company names. Microsoft, Amazon, Google, Apple, Baidu... good company to keep.
  • This is a business class so let's look at "Commercial Activity"... wait a second, that creepy Facebook auto-photo-tagging feature uses this?!?!
  • One Facebook-related article (reference #157) leads to another (in new tabs, of course) and suddenly we have some intuitive understanding for what this technique does, how it is different than other tech, a timeline for the tech, the names of some of the champions, who they work for, and what they're working on. You can do an awful lot with this data and some strategic googling.

At this point, you can probably feel safe bringing this topic up at a cocktail party or perhaps even asking pointed questions of an expert without being offensively ignorant, but what's next?



You take the blue pill—the story ends, you wake up in your bed and believe whatever you want to believe. You take the red pill—you stay in Wonderland, and I show you how deep the rabbit hole goes.

Do you know about Google's Patent tool? You should. Whether that was for something like Deep Learning (as linked) or for some other application... Hypothetically, if you had a patent/technology that you were learning more about, you might search it here to understand the novelty of the invention then look at the patents that it references to see how it distinguishes itself from those and where it builds upon them. Similarly, you might investigate those patents and see where they are novel. It is beyond the scope of this post, but the architecture of patent law is based upon novelty and incremental inclusions of functionality on prior art. You can use this to understand patent-based IP in a deeper way.

You are currently enrolled at a school. Schools have libraries; but more importantly, they have librarians. Librarians get a bad rap for being Dewey-Decimal-System-crazed-book-re-stockers and quiet-place-shushers. The reality is that librarians are stone-cold, ice-water-in-their-veins, query-crafting research ninjas. They're also wildly under-utilized and paid for by your tuition. They have access to resources you never knew existed... all for the asking. Hell, for a $5 Starbucks gift card they'll probably commercialize the technology for you.

When I deep dive in a new field, my focus is on comparison, contrast, and novelty. My goal is to understand how this specific unknown fits into its surroundings and how that little microcosm relates to what I do know. With enough research, you can build bridges from that little island back to your knowledge mainland.

Other resources:

  • Startups? - Crunchbase, AngelList
  • Technology? - HackerNews
  • Network? - LinkedIn
  • Did Wikipedia leave you high and dry in terms of next steps and context to continue? - Reddit
  • Maybe you aren't asking a new question? - Quora
  • Is there a professional or academic society/association associated with the topic of your research (try googling one of those words with the name of the topic's discipline)
  • Financial filings and other public company dealings? - EDGAR
  • Industrial players, channels, products, and vendors? - ThomasNet
  • People. Talk to people and leverage their wisdom. Be bold to get a meeting, respectful during it, curious as it closes, and appreciative after it. If you get only one thing out of this course, this is it.