By Categories: Company News, Tech News, Tips & TricksTags: Comments Off on How to Train an AI

Over the past few days, those who are beta-testing our AI-driven cloud service have noticed some odd things.  I’d like to go over the top 3 that some think are bugs.


One of the most frequent reports I’m getting right now is songs being added to users’ playlists.  Our AI makes decisions based on our tastes and choices.  For instance, if you skip through songs on your playlist that are fast-paced rap songs and stop on R&B, the AI begins to understand your mood is more centered toward R&B and will intentionally add some popular R&B songs based on your tastes and albums.  So if you have a lot of R. Kelly albums you will find new R&B music from R Kelly from various media sources and allow you to stream it from their services.

This is not a bug.


Another report was submitted; names are being assigned to faces in photo albums.  Once you add your email address, the AI will search social media with that face and will take a stroll through the friends and if it finds a face and name that matches the faces in your photos, it assigns a name.   The AI will not generally do this by default unless you indicate to the AI that you are struggling to find certain photos of people.

Again, this is not a bug.


For example, a friend of mine looking for a job just got an interview and dropped the schedule in his calandar and then almost instantly had a “Interview Points” dropped in his root folder.   He thought this was a bug until he read the points and saw an outline of topics to cover in the interview.  The AI decides what you’re attempting to do in your file structure.  Sometimes it may drop notes, and suggestions or even create a simple business card template in a “company” folder to aid you in a recent job search or resume or interview schedule.  The AI is very intuitive and looks at all aspects of the information you upload from emails to calendars, to contacts, to photos, and documents.  The AI makes decisions and if something is within it’s power to do in an effort to aid you, it will attempt to do it.

Once again, not a bug.


AI is extremely new to all of us at our company and no one here is an expert in AI.  This started as a fun little project to see how or what the new advents of AI would do with a few of our files and then noticed how well it learns.   This quickly became applicable to our clouds and our own cloud was born.  Our AI runs from 3 neuro networks that make decisions based on consensus.  No engine can act alone and needs consensus from the other 2. That means every decision made is as error-free as possible and means that when the AI does make a decision it’s very meaningful because the AI is pretty sure it’s a decision worth making.