Tills, Acknowledgements and Inner Voices
Who is on the other end of that chatbot? What do Colson Whitehead and Sam Anderson have in common?
I’m aware that I have weird interests.
Recently, I’ve become fascinated by restaurant tills. I love watching how the serving staff input the order. Is it just one quick tap and it’s done or are they tapping and scrolling and tabbing to find the right button? What options do they have? How flexible is the interface? When I ask if I can have chips instead of potatoes is that a couple of taps or does the interface bomb out to a notes field for them to manually type in the alteration? Sometimes I can’t believe they entered the intricacies of my order so quickly, other times I’m amazed they’re still fighting the system to input a cup of tea.
This is no reflection on the serving staff themselves. It’s the design of the interface. Or rather it’s the overall system design. Perhaps the restaurant bought an app that doesn’t fit with their processes. If it’s a chain, maybe it was built by the IT department, who may or may not have ever visited the shop floor. Sometimes the serving staff apologise and say it’s a new system and they’re not used to it yet. This feeds my fascination. If they’ll indulge me, I ask them how it all works, how it was rolled out, how they feel about it. My dining companions roll their eyes.
The best till system I’ve ever seen is in a market stall bakery. Although they sell dozens of different cakes, they categorize them into consistent prices. Cookies are one price, cakes another. On their payment system, which is an iPhone housed inside a contactless credit card case, they have large coloured buttons for each product group. When you order, they enter it faster than you can say it.
You might say comparing this to a phone-directory-sized menu isn’t fair. Of course a restaurant’s system will be more complex than a few Fischerprice coloured blocks. But this is why it’s system design, not IT design. The market stall’s whole operation works seamlessly and efficiently. The IT is just one part. They could have priced every cake individually, each requiring a separate entry. They could have added variants and options. Large or small. With or without cream. But they simplified and, consequently, their system is frictionless. It is the best I have ever seen. And I say this as someone who has spent a frankly unhealthy amount of time watching people use systems.
This market stall, of course, does not have an IT team. They purchased an app. They are probably using the default ones that comes with the card reader, SumUp or Square or Shopify. But the way they integrated their solution with their business process would be the envy of every corporate IT department I know.
When I see IT projects go badly, I see fault across the board. Yes, the software is buggy (developers). And yes, it has been rolled out prematurely (exec sponsors) and doesn’t look very nice (designers) and is hard to use (UX). But also the processes behind it aren’t well thought out (the business) and the software is being asked to do things it never should (stakeholders) some of the processes have been misunderstood (business analysts). People have not been steered to make better decisions (product management). And now they can’t migrate away because they signed a ten-year deal with a third party (commercial).
The victims, ultimately, are the people who use it. You can tell if the software is helping or hindering a restaurant when you make a change to an order, or sometimes even if you order a standard menu item. You see the serving staff scrolling and scrolling through a non-alphabetical, non-filterable list of every item on the menu to get to the specific one you asked for and then typing extra notes into a plain text field using one hand on an onscreen keyboard while balancing the tablet in mid-air. Is this, I wonder, better than a piece of paper? Or have we eaten up any benefits we got from automatically sending the order to the kitchen by making it so painful to use in the first place?
When the environment is right, I like to ask the serving staff about these systems. Often the answer is the same. “Oh it’s a nightmare,” they say, a pained look in their eyes, “But you should have seen the system we had before.”
Elsewhere
From Literary Hub, An Ode to Acknowledgements by Sarah Wheeler
In The Sentence, Louise Erdrich spends the greater part of a page thanking a dictionary she received as the prize for a high-school essay contest. In the acknowledgments to Harriet Lerner’s The Dance of Anger, we learn that the author is a member of a women’s group, which is not surprising, but nonetheless fun to visualize. Marlon James won’t let his mother read two of the pages in his book Black Leopard, Red Wolf. A pair of my favorite authors, Colson Whitehead and Sam Anderson, though quite different in style, both have children named Beckett.
It occurs to me, that if there is a theme to this edition of this newsletter, or, indeed, to every issue of it, it’s about looking behind the curtain at what is really going on. The reason I am fascinated by restaurant tills is because they are a window into how restaurants are run. They are almost a story in themselves. Acknowledgements, as Wheeler notes, are “where you find juicy, personal tidbits.” And, in the same way, they too can sometimes be a story in themselves.
If authors are rock stars, which in my mind they are, the acknowledgements are the part of the concert where they take a moment to shout out the names of the band members. “My tireless agent, Candice, on the drums! My husband of 20 years, Bill, on bass! The sweet barista at the coffee shop in Pittsburgh where most of this book was penned, whose name I can never remember, on keyboards!”
Jonathan Moses, writing in New Humanist: Talking to ourselves
In 2014, the Polish scholar Małgorzata Puchalska-Wasyl conducted a statistical analysis of common inner voices. Depending on need, context or personality, she discovered, participants might experience – or even strategically deploy – different styles of voice: from the “faithful friend” to the “ambivalent parent”, “proud rival” or “helpless child”. The personae of the voices can vary as much as their character: arriving in versions of our own voice or adopting the voice of another. They can be strict or gentle, punishing or kind. Yet for the most part, we are oblivious to their presence.
Another piece about going behind the scenes, perhaps. I have an inner voice that talks to me almost constantly, but there are no visuals. I’m aware other people are the other way around and have inner lives that are almost entirely visual.
Hazy images, some real, others half-real or pure fantasy, well up from their unconscious store, only to be swept aside at the command of an inner voice, which sounds like me but behaves very differently. Impatient, supercilious, prone to denunciation, the voice marshalls my thoughts, I quickly realise, towards the hidden standards I set for myself: deciding which memories are interesting enough, “me” enough, to share with the rest of the room. The voice reminds me that you don’t just do things, you do them well. And it does say “you” – the second person – even though it is me, just as the you is me too. It is the voice of my ego; the overlord of my imagination.
In The Guardian, by Laura Preston. Becoming a chatbot: my life as a real estate AI’s human backup
Meanwhile, we operators, with our advanced degrees in the humanities, had aptitudes Brenda lacked. We were intuitive, articulate and sensitive to the finer points of delivery. At $25 an hour we also cost almost nothing to employ, by corporate standards. Under the Brenda-operator alliance, everyone came out ahead: the operators got paid better than they would as adjunct professors, and Brenda became more likable, more convincing, more humane. Meanwhile, Brenda’s corporate clients were satisfied knowing they had not replaced their phone lines with a customer-service bot. What they were using, instead, was cutting-edge AI backed by PhDs in literature.
There was a story, a few years ago, about an AI transcription service that turned out to be low-paid offshore workers transcribing the audio. An advanced machine learning system provided to the NHS was revealed to be an Excel spreadsheet. This story from The Guardian is another “behind-the-scenes” of how humans plug the gaps when AI fails.
My recruiter had assured me that my sophisticated language skills qualified me for the position. In reality, the job was little more than a game of reflexes. The moment I logged on to the command station, messages stacked up in real time. Each message made a ping when it hit the inbox, a ping I soon learned was impossible to mute, and often the messages arrived in such quick succession that the pings stuttered and ricocheted off one another. Some timers were closer to zero than others, and I had to quickly assess which ones needed attention first.
The dream behind our powerful new AI solutions is that we will leave all the drudge work to computers, freeing us to pursue more leisurely, creative lives. But here we are, identifying pictures of bicycles while the computers generate beautiful dreamscapes in Dall-E. Creative writing graduates are rushing to look up mundane details of payment terms while timers tick down. This is the thing with looking behind the scenes: it’s not as nice back there as it is in front of the house.
That’s all for today,
Simon