an IBM coffee drone, fun and fluff
The story has it all; a prestigious company, a respected business publication, drones, coffee and machine learning. There could hardly be a trendier list of credentials.
There are some whimsical aspects to the Financial Times’ article and there are some missed opportunities. In the end it was a bit of fluff and a bit of fun.
Why am I here? What do I think was missed? Drone delivery surely has to be the pinnacle of tech. In short, they missed Claim 1.
fun and fluff
IBM’s US patent 10,040,551 a.k.a “Drone delivery of coffee based on a cognitive state of an individual” issued on August 7th. I could try and describe the article, but I think it is better to simply quote from it. I will start with a sentence that sounds quite serious: “IBM is one of a series of companies looking to harness the power of drones and smart technology”. I like the last bit here … “harness the power of drones and smart technology”. I wonder what dumb technology is, but I digress. “The technology … which the filing said could be used in offices to keep employees alert or by coffee shops to increase sales … would use the sensors to scan for people who have asked for a drink”. “They will also be able to learn an individual’s preferences, such as what time someone likes their third flat white”. That sounds vey impressive. It would be even more valuable if one could “dispatch caffeine to flagging employees and thirsty café customers even before they ask for it”.
All of this sounds pretty revolutionary and there was some straight faced reporting of the story that seems to drink the cool-aid, so-to-speak. Beyond the nifty ideas there are more than a few practical questions. Do people want drones flying above their head all the time, let alone during a coffee or networking break? The Times’ article indicated that the coffee will arrive in plastic bags to prevent leakage. That does not sound too appetizing, particularly if you have to open the bag and pour out the coffee when it arrives. I don’t think this would fly, no pun intended. In many ways this particular application, or use of the technology does not seem all that practical to me.
the serious bits
At this point I admit I did not take the time to read the whole Specification. Instead I read Claim 1, the sole independent claim, to get a quick notion of the technology in the ‘551 patent. What does it tell us? What does it say at face value?
Claim 1 refers to a “sleepy cognitive state” several times. The concept of being sleepy, and it being recognized does not come up in the FT article. Related to this, there is one clause that jumped out at me;
“accessing sleep data pertaining to a sleep cycle of the individual determined by motion detection of the individual during a sleep period, and adjusting the confidence level in accordance to the sleep data”.
So, the system or computing bit collects sleep data of an individual during a sleep cycle. They have to monitor each individual while they sleep. One can hypothesize this becomes a baseline against which the system can learn and determine if a person’s actions indicate they are, or are about to become, sleepy.
I can not image the employment agreement or conference that requires an individual be monitored during sleep prior to joining the company or going to the conference. This is a steep request to be able to receive coffee in plastic bags. It must also be asked if the collected data would be used for anything else. Would the system monitor how often it determines an employee needs coffee during the day? Would this be communicated to HR? What if there are other indications from the collected data? What if new software finds that the specific sleep patterns indicate a predisposition to stealing? Would the employee be monitored more closely?
Yes, I know this sounds a bit big brother-ish, but that sure looks like the slope along which we are sliding. Technology is fascinating and important to society. We do however need a clear picture of it to decide where we want to go and what we want to achieve with it.