Compounding knowledge is a term stemming from the compound interest equation and even in some way based on it. It's the gradual, long-term accumulation of knowledge that improves one's decision-making abilities, thinking and general knowledge. Going wide or going deep is an extension of compounding knowledge and an interesting debate. Is it better to focus on a very specific area early on? Or is it better to have a wide range of knowledge which serves as a foundation for further specialisation?
Compounding knowledge is very similar to long-term financial investing, only that knowledge is harder to measure, based on an entirely different medium, usually harder to grasp and hold on to. The core principles of building on previous investments remain completely the same. In practice, compounding knowledge means searching for valuable resources, summarising and remembering them, connecting the gained knowledge together and making use of it. Sometimes in a whole another area by building a unique perspective. The goal, just as with financial investing, is to enrich the information widely available into something new, something more valuable than when it was acquired. Since the imagination of knowledge through numbers is impossible and incredibly unpractical, it's perhaps better to think about it in terms of The Snowball Effect.
The Snowball Effect
The Snowball Effect is a metaphor for something that builds upon itself as time progresses. Usually, the term snowball appears in research, when researchers are dealing with sampling. Snowball sampling is a technique, where already recruited participants recruit new participants to be included in a research. The number of them quickly multiplies. A snowball rolling down the hill starts as a small one, but eventually grows to great dimensions to the point where it could be life-threatening (an avalanche). However, applying snowball effect to knowledge is nothing to be afraid of, it's even incredibly beneficial. It's just another representation of the compounding effect. A snowball that rolls down the hill is a sphere that grows larger with every roll. This means, from a purely logical standpoint, that its surface is getting larger and larger. But the volume experiences an even greater increase. One could imagine successful compounding in this way. Every piece of information connects to a previous piece of information that acts as a pillar. For both successful memorisation and learning.
The gradual acquisition of different pieces of information is key to new ideas. Sometimes it happens that you read something, a book perhaps, and then suddenly you have an epiphany of where all you could apply this newly acquired knowledge. This information, however, needs to be beneficial. It's of no particular use to consume information that you forget a week after consuming it. There's no use to add snow to the snowball that melts the very next day. It's the same with consuming information that has no application in the long run. That's why it's objectively not smart to consume daily news. Almost no one remembers what happened a week ago that was told to us by the news. Apart from that they affect our mood and make us informed, not smarter. Books, on the other hand, are a completely different story. They contain valuable and tested knowledge that is much more beneficial to your development than some other sources.
Remember, the compound equation works only on the long-run and it's no different when it comes to knowledge. Or a snowball.
Wide or Deep?
The question arises whether to dig deep or to dig wide. Do you accumulate a wide range of knowledge and build on it by connecting and thinking about it laterally? Or do you accumulate a large amount of knowledge of a specific area where you want to work and thrive? There are multiple perspectives on this.
Warren Buffett, the most successful investor of all time, is perhaps one of the best wide-known examples of "going wide". He accumulated enormous amounts of knowledge about how companies work, how stock markets work and how business works. This allows him to make decisions when it comes to everything about making money. Berkshire Hathaway (his company) is buying other companies only after Buffett makes the decision. And how does he do that? He reads every annual report of a business he wants to acquire and makes his decisions based on raw data and his vast amounts of knowledge about this field. How did he accumulate such surreal amounts of knowledge? First of all, he made every opportunity to learn about business count. And secondly, he has a great memory. Over the years, he has built his own library of information on this relatively specific area. Today, he still reads about 500 pages of books per day. 
On the other hand, there are arguments that we all should postpone specialisation and try to accumulate as much knowledge as possible from as many areas as possible. Take Roger Federer, one of the best tennis players of all-time, for example. His mother made him train several sports along tennis, even though his peers were specialising into a specific sport. Perhaps all these different skills from different sports accumulated and helped him on his way to becoming one of the greats. There's no uniform truth to this. But Federer, on the long-run, acquired all these skills, which just may have allowed him to spend less time specifically training tennis. He maybe had a knack for tennis, but all the subsidiary skills definitely helped him get where he is today. 
The conclusion of this debate is that yes, specialisation and compounding in a specific area is necessary sooner or later. But in order to find where to go deep, you first have to go wide to find where you truly see yourself. Compounding knowledge, however, is going on in both. Buffett found his passion early and started accumulating knowledge about it from an early age. Federer needed some time to find where he's the best, but once he did he started compounding in tennis where he still does today. So objectively, it's hard to know your passion if you haven't experienced it. That's why it may be better to first go wide and then go deep.