Snobots, shoveling and changing the way we think about things

Today is our second snow day this week, and yesterday the schools had a delayed opening due to weather. I grew up in this area so I’m used to the snow, but lately I’ve been thinking about it differently.

Live look out my office window

If you live in an area where it snows regularly you’ve probably developed your own approach to snow removal. You can ignore the next part and skip down to where I ask you to think. Our driveway is one that the plow truck operators refer to as a blower only drive. So before you tell me to “get a plow” it’s not an option.

If you’ve never dealt with snow, let me explain a few things. Snow is like a living organism, it’s constantly changing. Temperature, sunlight, wind, kids, pets, you name it and it has an effect on the snow. This means clearing the snow, so that you can get on with life, has an intuition aspect to it that’s hard to quantify (sounds like something A.I. would struggle with).

On Tuesday, when we last had school canceled, it didn’t start snowing until mid-morning, around 9:00. But then it snowed all day, dropping 4-6 inches. It was cold, so the snow was light and fluffy making it relatively easy to clear. BUT! Wednesday was forecast to warm up and possibly rain. So I had to clear it. Otherwise it would have packed in and frozen into a sheet of permafrost, not to be removed until May.

I timed my shoveling to about an hour before the snow stopped falling; roughly 4:30. This allowed me to get it done in the last drops of daylight and while it was still light and fluffy. On Wednesday morning there was a dusting coating the driveway and walk but I knew it would melt with the sun and rising temps.

Today we’re scheduled to get 10-16 inches with temperatures in the mid to upper 20’s. I just got home from running some errands and the snow is wet and heavy.

There is a huge difference between shoveling or snow blowing 5 inches of wet snow and 10 inches of wet snow. The effort curve is non-linear. Dealing with double the snow requires more than twice the effort. Instead of planning to go out close to the end, I have to figure out when we’re about half-way and get out there.

In fact, I like to wait until we’re a little past half way. That makes the second pass a little easier than the first. This is a good thing because by the time I’m doing the second pass I tend to be tired and grouchy.

Here’s where I start the thinking part.

I can clear up to 6 inches of snow from our walks and driveway in about an hour. If it’s really wet and heavy it might take 90 minutes. The issue us more about how much ground there is to cover than the snow itself. Walking the snow blower up and down the driveway simply takes time.

So even if I went out to clear snow every time it accumulated to 3 inches it would take me an hour. On a day like today thats 4 or 5 hours outside in the cold and snow. It wouldn’t bother a Snobot, but for me it would bring new meaning to the words grumpy.

That means that all the snobot has to be able to do is clear the maximum amount of snow in an hour – 6 inches if you don’t want to check.

But the other challenge of clearing a foot or more of snow is where to put it. In years when we’ve had multiple storms the driveway and walkway get more narrow with each snow. Three and four foot high snow banks are not unheard of. Lifting or throwing snow up and over a four foot high bank requires some muscle.

And this is where I remember we have to think differently. The problem statement is simple – how do I keep the walk and driveway free from snow?

The challenge isn’t about automating the snow blower or creating a robot to use one of the snow shovels. The snobot could be a tracked electric vehicle with a heating element underneath. It could be a flying drone that relies on the wash from it’s rotors to blow the snow away. Or it could be some crazy combination of heating, pushing, blowing and throwing that I can’t name.

A snobot would be connected, getting real time updates on both hyper-local and regional weather conditions, as well as smart. It would understand the ground temperature and topology and it would be tireless.

Artificial intelligence and automation doesn’t simply mean having computers take over existing machines. It’s about re-thinking how we solve problems.


Artificial Intelligence, Automation and jobs

As you dive into artificial intelligence it’s hard not to come across news and opinions about how it will be taking over jobs. While there are concrete examples like automated beer delivery if you compile a report, follow a route or repeat a task A.I. will impact your life.

I can remember once when I was like 8 years my mom yelled at us about offering to help. She said that if we walked  into the kitchen and she was working on making dinner of cleaning we should ask if there was anything we could do to help. It was the least we could do to contribute she said.

Later, when we were older she yelled at us about asking if we could help. “You’re old enough now to know what needs to get done. The best way to help is to just do it. You don’t have to ask me!” she told us.

Simply following instructions didn’t add a lot of value, we had to think and act. That’s what artificial intelligence is starting to do and what we as humans have to get back to.

Image courtesy of Ilya Pavlov via Unsplash

There was a recent NY Times article about Siemens not having any jobs for high school graduates. They need people who can creatively solve problems and work with technology. It made me think about the kids who are in high school now with plans to go into a field that will be automated out of existence in the next 5 – 10 years.

For example, right now there is tremendous value in the ability and willingness to drive a tractor trailer truck from Sacramento to Chicago. Motor freight is a cost effective way to get a product from point A to point B. But the human is the most expensive part of the equation. Automation isn’t interested in getting rid of the person driving, it wants to get rid of the expense associated with the person. So humans need to look for other ways they can add value.

For those of us who sit in front of a computer all day it’s easy to see this with detachment. But regardless of whether your job is compiling reports,  writing copy or assembling components artificial intelligence is looking for ways to perform that function better and cheaper.

The good news is that it won’t happen overnight. There will be places where automation makes too much sense to ignore and it will be implemented quickly. A straight clearly marked super highway is easier for A.I. to navigate than narrow, twisty and unmarked back roads. Analysis based on intuition and off-line data sets will be far more difficult for A.I. than a forecast for standard, everyday necessities.

We have to stop thinking about getting paid to follow a process. What we all have to start thinking about is where we can add value. Asking what you can do that wouldn’t be the same if a computer did it will lead you to opportunities that may not have existed before. Future opportunities won’t come from doing things that a computer can’t do, they’ll come from doing things that humans would rather have done by another human.


Conversational tid-bits on artificial intelligence

With 62% of enterprises planning to deploy some type of Artificial intelligence by 2018, AI is a hot topic these days. It’s concepts and capabilities are invading our cars, kitchens, and jobs. But what is it? And what does it mean?

Image courtesy of Kevin Curtis via Unsplash

The other issue is how do you get involved. If you go out for drinks after work and people are talking about artificial intelligence how do you identify the pretenders? Or better still, how can you ask a decent question to show that you are interested?

This is not intended to be an engineering paper or give you the tools needed to present an AI recommendation to the executive staff at work. It’s not even AI 101 level stuff. I want to offer a few tid-bits so that when someone starts talking artificial intelligence you can exhibit a little human intelligence.

Right now it looks like the Finance and Healthcare  industries are deploying the most advanced AI instances. But if you’re in a manufacturing, transportation or marketing related field use of artificial intelligence is picking up speed.

When it comes to discussing Artificial Intelligence, it is obviously a broad topic. Don’t get drawn into an argument or conversation with anyone who paints with the “AI is about to take all of our jobs” brush. Those people are out there and they’ll cite examples from random websites but not have a lot of details to support the theory. Most likely they are afraid of change and don’t have the time to learn more.

Many of us, myself included, immediately think of what the experts call Artificial General intelligence (AGI) when we hear AI. AGI is the walking talking robot with a little bit of attitude and a whole lot of data. Or Jarvis from Iron Man. AGI is cool and flashy, but it’s not coming for your job in the next 5 years.

If you want to jump in on the conversation it’s good to ask what role the AI will be used for. There are all kinds if different AI’s coming to market. The AI used in Tesla’s autopilot is not going to also suggest a gluten free recipe you might enjoy based on the ingredients in your fridge. And the AI used to recommend a parts forecast for the manufacturing team isn’t going to start generating content headlines for marketing programs. Much like new head count, AI will have a specific job description.

Other terms you’ll hear related to artificial intelligence are machine learning and deep learning (Wikipedia links so you can read more). If you hear these mentioned it’s likely someone who is working with or growing familiar with actually using AI.

Machine Learning is a subset of artificial intelligence where an application can change based on receiving new data. It lets the computer learn without being programmed. Machine learning is the power behind many of the recommendation engines you see on your favorite content sites. It’s what allows the computer to suggest content based on what you have previously looked at.

Deep learning is closely related to machine learning. The biggest difference is that deep learning will take huge chunks of data and project a model of a predicted outcome. Deep learning is what would forecast your purchase of a new TV based on your last two months of researching them online. You probably wouldn’t know when your surfing is being fed into a deep learning algorithm, but it is.

Artificial intelligence is capable of many things. In most cases it will help people do their jobs better, not replace them. Getting involved in the conversation sooner will help you operate more comfortably as AI comes into your business.