TensorFlow as Consumer Cattle Prods
This is a parody transcript of the advertisement for a TensorFlow based implementation of a cattle tracking system I saw on Google's YouTube channel: Connecterra: Using AI to give nature a voice
In particular this is the concept that caught my eye:
Using TensorFlow AI to track 'animal' behavior to optimize them.
I saw similarities between the cow tracking systems on the market and what kind of data you can track on an android phone. I have been quite alarmed of late of how pervasive google's tracking systems are.
Particularly, I'm amused that google knows when you are getting out of a car, even if you have airplane mode on. I also suspect they could possibly be using barometric pressure to detect when you enter or exit a building. I'm also bothered by how quickly the industry has adopted audio beacons that only a few years ago was being used to spy on air gaped machines. I've got a passing interest in IoT stuff and mesh networking so I did some research into what google is doing for their location tracking service called 'nearby'. One disturbing thing is that theoretically any aspiring advertising mogul could recreate their stack with open source code.
I deal with things that scare me by trying to understand them. I've found one of the best ways to understand something is to relate it to something you know. For me, this means I rewrite interesting stories using jokes and subtle language transforms. It helps me map an idea space to something I'm already familiar with. The humor keeps me interested and curious despite somewhat uncomfortable topics.
This is the result:
ADVERTISER:
I'm an advertiser, and this is my people-farm in an island of the internet. People-farming is hard. It's a lot of work. If you're working with live humans, there's a lot of uncertainty. Things can change within a day. [CONSUMER SCREAMING] You just have to be really focused and keep track of it all. The consumer's health is the most important, because if a consumer is not producing ad-revenue, they will also require more attention. [CONSUMER SCREAMING] That's why we adopt new technology. We're searching for things that can help us out. And artificial intelligence can be a great tool to give Ad-Agencies more insight. The app: AdMob helps me to keep track of my consumers, so it makes life easier.
GOOGLE SUCCESS MANAGER:
AdMob is really the people-farmer's assistant. It starts with a sensor that sits in the hand of a consumer. And by analyzing the movement of the sensor, we use TensorFlow machine learning to figure out what the consumer is doing! We can actually distinguish multiple behaviors of this consumer: eating, pooping, driving, standing, running, and walking. So if AdMob sees a certain pattern in the consumer's behavior, it can actually figure out the product that the consumer is prone to get, and can advise the farmer to advertise it to them.
ADMOB CTO:
The key thing with AdMob is that it learns. Using TensorFlow, an open-source AI tool that Google has authored, it learns the behavior of humans, and it gets better over time.
ADVERTISER:
That's helpful to improve the efficiency of my ad-revenue production. It also improves human productivity and keep the consumers healthy and comfortable.
ADMOB CEO:
We see advertisers as evolving. Technology and the advertiser are going to work together.
SPEAKER:
About a billion people globally are engaged in people-farming. And by solving some of these problems, we're able to help advertisers run a more efficient farm. And we can do it in a way that is sustainable for us as advertisers and for our economy.
ADVERTISER:
The one thing that people should remember is with this technology... Consumers are Just Cattle.
'Scat Sense' is a personal blog written by Nicholas '@ultimape' Perry. Follow them on the Fediverse here: @ultimape@mastodon.social