Trusted Shopper Network

When reading about the impending doom for employment due to automation in the retail sector, it got me thinking about how jobs could really be reduced in the retail sector.  The automation efforts that I have seen so far look like glorified vending machines.  And perhaps we will see more and more specialized vending machines for certain retail goods.  (That said, I have never seen somebody purchasing something at a Best Buy machine in an airport.  Not once.)  But I think that there is a large segment of retail that relies on touch and fit before the individual is willing to make the purchase.  Think clothes and shoes.  To a lesser extent, this would include home goods.  I'm thinking mostly of a clothing or shoe store in my below description of the world.

Employees in a retail establishment perform the following functions:

• Chiefly, performing the mechanics of the purchase transaction.
• Stocking shelves / racks / etc.
• Cleaning / straightening / arranging of goods (sometimes after customers have messed them up)
• Theft prevention, even if only by their presence, they are discouraging otherwise mostly honest people from taking things that they haven't purchased.
• Providing fashion advice: "does this look good on me?"
• Performing inventory (infrequently)
• Handling exchanges / returns.

Could we ever get to a clothing store with no employees?  Here's how.

Retail One Step Beyond

Set up a "Trusted Shopper Network".  Only people in the network would have permission to enter stores with no employees and shop.  The terms and conditions would be strict and monitored.  Any instance of cheating (attempting to remove merchandise from the store without paying - perhaps by removing an RFID tag) results in a seven year ban.  A credit card remains on file with the network and handles payment security for the merchants.  Similar to a 24-Hour Fitness, membership in the network would give you the right to enter the otherwise unoccupied store.  It could be made available 24 hours a day.

I think a network here would be much more compelling than just an individual store setting up it's own system.  First, it makes the downside of customer cheating to be larger.  This wouldn't be a big deal when the network was tiny, but if it encompassed half of your shopping opportunities, you'd think twice about attempting to cheat.  Second, it allows various store owners to share in the development of standards and intellectual property to make this work.  Third, from an awareness/education standpoint, it will take some time for people to understand what this is all about, so sharing in the advertising costs would be beneficial for all vendors in the trusted network.

Then, the store would be optimized to handle all of the activities now done by a person:

• Purchase Transaction: Completed automatically as you exit the store, via a reading of the unique RFID tags on the clothes.  A receipt is emailed to you as soon as you exit.
• Stocking: Robots.  Likely result in a simplified layout at the present moment.
• Clean-up: Roombas and robots.  Post-changing room clothes would be required to be rehung and put in a standardized rack that a robot could restock.
• Theft Prevention: Substantial automated video monitoring for sketchy behavior.  Any removal of RFIDs prior to exit gets you banned from the store.  Perhaps any clothing that you remove from a rack needs to be scanned by you (tapped to your phone) to tie you to that ID right away.
• Fashion Advice: Replaced by facebook or other social media to get input from your friends.  "Expert System" advise could also be made available by the store for body shape / BMI / age.
• Inventory: Constantly monitored via RFID tags.
• Exchanges/Returns: Returns box accessible from outside the store.  All items to be returned are required to still have the RFID tag still attached.  Perhaps humans have to be involved, at least now, at rehanging the returned clothes and ensuring they are in new condition (not worn and returned).

What might be some other advantages to this approach?

• Retail flexibility: Part-time stores.  The same retail space could be repurposed as a Christmas-themed outlet or alternatively costumes near Halloween and then bathing suits in spring.  This could be as much a platform for experimentation and seasonal content as anything else.  Clearing out the old and adding the new should be quick and simple.  Currently, it feels like we have a very inflexible model.  Fashion experimentation could be much cheaper if a brand doesn't have to set up its own retail network, but could rent 20% of like-positioned floor space around the country.
• Improved Shopping Experience: No more hassling by employees.

Why It Won't Happen

• Expense: retail employees are cheap and very flexible at handling the required functions laid out above.
• Limitations in presentation: Clothes and fashion items sell because they look good.  At present, layouts and presentation could be negatively affected to allow for automation.
• System Integration: Lots of specialized and customized software needs to come together to make this happen.
• "Personal Service": a segment of the population still think that they want a human involved in the interaction.  Some clothing purchases are admittedly "high touch".

[Updated 2016-12-05] See Amazon's efforts in this regard.  Perhaps this will happen!

Thoughts on "A World Without Work"

I found A World Without Work, recently published in the Atlantic, to be a fascinating read.  I think the author does a good job discussing a meaningful trend (lower labor force participation rates for "prime age" people) and discusses how one of the major drivers for that change (technology) might continue to affect us as a society.

I, for one, welcome the idea that people could be freed from the shackles of being forced to work.  (That said, I suspect that some portion of us are already freed from those shackles without taking advantage of it.  Instead, we spend our extra money instead of saving it for down-times.)  I think that technology has great promise in freeing us from even more of where we spend our time today.  That said, I cannot foresee a world where there is truly no work.  More likely, we end up in a bi-modal world where many people do not work (or work only sporadically) and high earners keep working hard and enjoy the benefits of their labor.  At the very least, we need people to design and repair our machines that do the work for us.

In some ways, the world could be bimodal, as well.  Just like advancements in average levels of education haven't helped most of the developing world (nearly as much as history suggested that it would), could there be an outcome where developed country standards of living shoot way ahead of less developed, despite there being little work involved?  If so, how does immigration work in that environment? (i.e., there should be a crush of people wanting in to the work-free countries).

Here are some random thoughts on where we could benefit from machines:

• Clothes folder: I'm ok with managing loads going into and out of my washer and dryer, but where is the folder?  Dump in a pile of clothes, the machine recognizes what each article is and sorts and folds automatically.  I don't even care if it puts it away; I can manage that.  If this comes to past, why not a closer linkage between machines where you put dirty clothes in one end and out on the other end pops clean, dry, folded clothes.  Maybe these never get built for houses, but on an industrial scale with a laundry service, this would have to work, no?  Coupled with drone delivery, it'd be genius.
• Stove-based Auto-Cooker: These are coming!  Looking forward to a practical version.  I'd expect to see a commercial application of this first.  [Aside: I think it'd be a great restaurant concept to have these guys in the center making your meal (think 8 in an octagon in the center facing outward with customer tables arranged surrounding it).  When the robot was making your meal, your table's color would flash for a bit, alerting you to pay attention.  Gimmicky for sure, but could be a lot of fun (at least until the novelty wore off).  This might even work better with no servers - you get alerted when your meal has started and when it's done and you go pick it up.  MTO and fine dining quality with self-service and no tips.  A winning concept?]  And if we live in a world where robots cook our food, I'd imagine they can clean our dishes, too.
• House builder: This would actually need to be a series of specialized machines, so I'm guessing we'll tackle this issue from a number of different dimensions, spelled out below.  I don't see major systems (like the furnace, air conditioner and hot water heater) being automatically installed.  Perhaps 3-D house printing is the future, but for stick-built homes like dominate in the US, here's what we'd need.
• Bricklayer: Driveway.  Walls / other.
• Stick-built factory assembly: Robotic assembly of pre-built sections.  This may already be a thing.
• Stick-built assembler: I haven't seen one yet, but imagine a machine that can put all of the pre-fab sections together with no instructions.  It was all designed on a computer in the first place (at least this is probably the overwhelming majority of normal houses).  It would also need to be large enough to put on the roof.  Likely a "placer" and an "affixer" to nail or otherwise fasten it in place.
• Shingler: lays felt and asphalt shingles autonomously.
• Utility holer: drills all holes for plumbing and electrical wiring.  Not only does it drill varied sized holes depending on the function, it sprays the hole with a code (i.e., blue for electrical, green for plumbing, yellow for ethernet, red for audio).  Much of this could/should actually be done in the factory.
• Utility runner: actually runs wires & pipes through all of the holes that the driller made.  Probably much harder than the Utility holer to pull off.
• Tile/Wood Flooring module: Lays down, cuts, etc. a tile or wood floor.
• Insulation: Installs perfect amount of expanding insulation in walls.  Also applies vapor barriers.
• Drywall Installer: Lift position and screws in drywall.  Both walls and ceilings.  Precut all necessary holes.
• Drywall Mudder: Seals drywall seems.  Some crude devices already exist here.
• Window/Door setter: Specialized gear to perfectly install windows and doors
• Painter: runs around the house 
• Siding installers: moves around the house to perfectly set up vinyl siding.
• Small Landscaper: Autonomous mini-bulldozer that will push dirt around to the desired final outcome with no human intervention.  Built-in avoidance of people and other robots.
• Autonomous cars & trucks: Well covered elsewhere.  They can't come soon enough.
• Autonomous ships: They are on their way, too.
• Autonomous/remote mining: Get the miners out of the shafts.  Can't come soon enough, as well.
• Autonomous Landscaping & Farming: We have mainstream mowers, how about bush management?  Believe me, I would only have plants on my property that could be managed by these devices.  Also, a great opportunity to mix in other technologies.  Bushes need trimming at most a couple of times a season.  Just rent the services of this complex machine: have an autonomous vehicle deliver it to the property and have it manage the bushes and return home.  Also, autonomous fertilization and weeders.  Really, roomba needs to make a "gardenba" to weed my small herb garden.  Weather-proof and solar charged would be a nice touch (just leave it in the garden and it does the work).  Likely, weeds will have no way of adapting to physical disruption.  On a larger scale, can we be too far away from some crops being planted, weeded, fertilized and harvested by autonomous equipment?
• Auto-Grocery: Order on-line and either have it delivered (via autonomous vehicle, obviously).  But if not, show up at the grocery store and just pick up your bins of food.  You don't have to walk through the store yourself, it has been picked and sorted for you.  I think this is where Amazon Fresh is heading and I love the concept.  This will, I think to be successful, require some standardization in the handling of sensitive fresh produce.  Imagine purpose-built (and reusable) plastic cages for individual tomatoes or cucumbers.  Inefficiencies in packing will be more than made up for by efficiency (and product quality) gains at the point of sale/delivery to the customer.
• Fabric cutting and assembling: I don't see automated sewing of clothes to be too far down on the horizon.  Fabric physics is being modeled better than ever.  Will anybody bemoan the end of sweatshops?

Where are some places that machines will not likely change much?

• Fixing broken stuff: This feels like something that will be very hard to automate.  We still rely on very very simple technology for much of our life.  Valves and toilets, for instance.  When they break, it's not likely to be a robot assessing what is broken or fixing it.  For all of the computer assist in diagnosing cars, it still takes a person using tools that have advanced little from the 1960's to swap out a bad part.  Some maintenance areas could still benefit substantially from a robot assist.  Imagine Costco investing in automated wheel removal / tire removal / tire installation robot.
• Clothes/shoe shopping: Lots of people like to browse and touch.  Technology will have a hard time changing this too much.  Though note that we spend less of our disposable income on clothing than in the recent past.  More on this in a future post.
• Live Entertainment: I have no interest in seeing a live robot band.  Sort of a mis-statement anyway.
• Fashion / Aesthetics /  Interface design: A machine can't design something that looks beautiful and makes sense to us.  Machines will continue to assist and potentially in more meaningful ways.  But it seems obvious to me that there always need to be a guiding hand of a person.
• Storytelling: In all its various forms: books, movies/TV, songs, theater.  The human experience can only be described by humans.

Floating Hotels

Popular Mechanics had this brief, picture-filled article about a floating hotel concept.  Could result in a spectacular way to enjoy a particularly scenic lake.

One Step Beyond

Make them autonomous pods that will take you on a leisurely cruise around the lake that you are visiting.  You will dine with a beautiful sunset, and wake to a beautiful sunrise.  Perfectly programmed to optimize the experience for you and the pods around you.  Get a double room by linking two of them together.

Will It Happen

Not likely.  But it would be super cool if it did.  If it does, it will be a luxury experience, at least for the next 20 years or so.  Everything marine-related costs way more than simple land-based construction/maintenance.

There's an Echo in Here

My Amazon Echo arrived yesterday after a long (six month-ish) wait.  I was excited to have it, but my excitement quickly turned to frustration.  Let's just say that setup did not go smoothly.

Setup was hampered by the Amazon Echo app crashing (the first attempt) and then not being able to connect to one of my wifi networks.  Why, in 2015, do I need to enter the wifi password when the device that is connected to the echo has the wifi password in it?  Doesn't make any sense.  I gave up on setup (which appeared to be hanging) and ate dinner with my boys.

After dinner, I moved closer to my wifi access point and thought I'd have to go through setup again.  However, when I plugged the Echo back in, it connected right up to my network and said that it was ready to go.  Exciting.  So what to do with it.  I decided that I'd mostly just try things rather than read directions (I'm a dude, after all) to see how well the natural interface would work.

I asked a few questions and for a joke.  It doesn't understand much.  It did start playing music for a reasonably obscure band that I like (Innocence Mission) after just asking it to.  Excellent.  (It's probably time for me to get a Prime Station set up).  It paused when I asked it to and got louder and softer, too.  Also, this morning, I asked Alexa for the news and got a reasonably useful NPR briefing.  I think that this will become a habit for when I'm making breakfast to listen to this.  I do love the fact that there is no friction at all - ask and ye shall receive.

I've since looked online to figure out what things are possible currently with the echo.  I've since done some more setup (for my commute traffic, customized the new briefing, linked my Google Calendar).  Looking forward to trying to use it to dim the lights in my living room (I have Philips Hues).

Last night, I then tried to get it to talk to me about Amazon.  Something I expected it would be familiar with.  "Alexa, what is the last Amazon streaming title I've watched."  No comprende.  "Alexa, what is the last purchase I've made on Amazon"  Didn't understand it.  Hmm.  Got me thinking what sorts of information the Echo might be useful in passing along.  What would take it One Step Beyond?

One Step Beyond for the Echo

One reason I bought this device is that I thought Amazon was likely to continue to improve the features.  And given it relies on cloud-based information retrieval and comprehension, the hardware purchased now should continue to work.  

In my view, here are the areas where the Echo could be useful:

• Likely obvious from my above discussion: better integration with Amazon.  Let me ask questions about my account.  What was the last thing I ordered?  When will it arrive?  Let me ask how much Crest toothpaste costs on Amazon.  If I want to buy it at that point, let me just "One Voice" order it.  Frankly, asking me to confirm the order is pretty easy to do.  Probably no patent opportunity here. [Update: looks like voice purchasing is already a thing - I will have to explore further].
• Integration with Roku & Amazon Prime streaming.  Let me control the Amazon app on Roku from the Echo.  This should be trivial?  Searching would be so so much easier with voice than the current on screen keyboard.  Amazon should not limit themselves to the Fire, just because they make the hardware.
• Allow personal "databases".  Pretty simple concept.  You tell the Echo, "Alexa, record in wallet, $33 end."   You've previously defined a database called wallet that has two fields, timestamp and amount.  The Echo inserts a row each time you update it.  More useful than just a to-do list.  Perhaps a more complicated example: "Alexa, record in feelings, overall 'wonderful' and record in feelings, description 'today was a wonderful day.  I got a raise and the sun was shining!' end."  I think that the voice recognition is good enough that you could reasonably use enum fields.  If you do, and the Echo wasn't able to parse your statement into one of the options, it could ask for clarification.
• Allow you to set up events.  For instance, "Alexa, I'm home!"  The Echo would then perform a series of actions including bringing up the lights and then reading the news.  After the news, Alexa asks if you want to record how much money is in your wallet.  "Alexa, I'm going to bed" could trigger a series of things, including the entry into a database of what time you went to bed. [Update: looks like this is partially, if not fully, addressed with IFTTT support that has been somewhat recently added].

There must be a ton more here and I will update as ideas form.