Data size, quality, machine learning and Apple

A conceptual inference struck me the other day as I was considering buying additional Apple stock for my ‘invest and almost never sell’ portfolio – there’s a 5-10 year horizon for this capital.  This go around I sold PUT on Apple stock and if it gets assigned, I will be content – there is no FOMO as i can repeat this.  But maybe in the inference path I used will be helpful for thinking about weighing investments leveraging competency in machine learning which I believe will be a major investment thesis over my target time horizon – 5-10 yrs.

Here are my assumptions wrt machine learning capabilities

  • The larger the data set available for the machine learning and correction, the more accurate the machine learning – basically the faster and better it learns and is then able to execute
  • The cleaner the data set the more efficient, effective and timely the same machine learning

If you buy those, then looking at Apple there a couple of like companies with such a broad and large data set that can be leveraged to provide better and better (i.e., demanded) user services and products, e.g., Apple, Amazon, Google, Microsoft are the most relevant.  There are others that may have large data sets but I do not think they have the same breadth and quality, e.g., Facebook, Comcast, ATT, Verizon.

Apple’s data may not be the largest, but i think it is the most controled and hence the cleanest.  Microsoft is probably the second cleanest data set but not as broad across device types, usage models and connectivity modes.  While Google and Amazon probably have the largest and most diverse data sets, neither have the same control over data quality that Apple and Microsoft have.

Based on this inference thread and my current position in Apple, I will continue to add to my position until the narrative changes or my allocation of Apple exceeds 10% of my managed portfolio.  Microsoft will be added to my target list with an upcoming deep dive on 5-10 year investment thesis and entry.

Leave a Reply

Your email address will not be published. Required fields are marked *