Leaving the Cart Behind The Horse: Why I’m Not Focusing on Voice Technology (For Now)
My wife came into my office this weekend and politely grilled me with “I don’t think Calvin is listening to Pollexy. Is it working?” Well, when it comes to Calvin at night, she’s correct, and for now that’s OK.
Right now, my priority is understanding the problem, not providing a solution.
Calvin wakes up and can’t get back to sleep at certain times of the night–just like us.
Calvin wakes up and has to go to the bathroom sometimes–just like us.
Calvin wakes up and probably has his own internal worries–just like us.
Think about yourself for a second–if I had a camera on you, and you woke up at 2 a.m. worried about your doctor visit the next day . . . and I called you to say, “go back to bed,” would that help? Probably not.
Unfortunately, we can’t ask Calvin why he’s awake, or what he’s thinking about. But we can gather metrics, and use machine learning to try and understand patterns–and that’s my priority right now. In the meantime, I’m verbally communicating with Calvin over the Nest Cam myself if necessary.
Some of the metrics I’m gathering are:
- What did he do the day before and how long?
- How much screen time?
- Did he take a nap? When and how long?
- When did he eat?
- When did he take his medicine?
- When did he shower and how long?
- How many bathroom visits and when?
While I can’t promise we’ll solve his sleep issues by looking at charts and graphs, observation is the only tool we have right now until Calvin can use spontaneous language to communicate with us to tell us why he wakes up. When we start finding trends, then we can potentially start working on a voice solution.
For the curious, this chart shows the past 24 hours if anyone sees any immediate patterns that jump out:
- Top chart = bathroom visits
- Middle chart = when he was in his bedroom walking around
- Bottom chart = same as the middle chart, but the previous 24 hours for comparison