top of page
  • Writer's pictureJason Millar

The Human Machine

I am old enough now to remember the great yawning emptiness of the early internet; what we then called the World Wide Web. I remember excitedly first connecting to it and then wondering what it did. No one was emailing me (it was 1997), there was no YouTube, no Facebook and no Twitter. AoL was king, but back then it was still America Online, installed by CD-ROM. I remember the whine and gurgle of 56k, the dropped connections when my mum lifted the phone, and, a little later, a website called Amazon that sold books.

At the time, this was a huge boon to me. Amazon was new and so much more exciting than charity shops or the local library, and its algorithm – Customers Who Bought This Also Bought – was world-enlarging. I lived in darkest countryside and the nearest proper bookshop was sixty miles away. I used to send relatives injunctions to stop in second-hand bookshops they might be passing to pick up books from certain authors, or find the second book of a trilogy I couldn't start until I had the set.

Amazon's formula has, perhaps reassuringly, remained quite basic as anyone who has bought anything on there will know. You buy a bathmat, and until you buy something else, are inundated with bathmat recommendations previously examined and rejected. Today as questions are raised about whether we have reached peak technology, the joy of algorithms is starting to lose its shine. Now I spend my time going to physical bookshops and taking increasing pleasure in defying the algorithms: messing with their little machine minds, smiling as Netflix attempts to triangulate the third point of Call Me By Your Name and Derry Girls.

I am sure that humans, despite their comfort zones and safety blankets, are not logical. Somewhere, deep down, they like to be surprised. That's a problem for algorithms; they are necessarily predictable. From Tinder to Twitter, they can never give you what you didn't know you wanted. For now, they're just machines processing data we give them.

What has any of this got to do with wine? Not much, but I feel as humans we are in danger ourselves of starting to think in algorithms rather than working creatively, surprisingly. We see a customer buy merlot, we recommend malbec. Like the horizon-shrinking calculations of Netflix, suddenly the huge, chaotic world of wine is reduced to tiny increments of risk away from our comfort zone.

Where is the thrill? Who is ever going to get to mature Barolo from Barefoot Moscato at this glacial pace? It could take them all their lives to take that trip.

Human intuition, not algorithmic sophistication, is the way out. We cannot compete with the cautiously formulaic brilliance of the machines by being more machine. We must be more human, utilise fickle psychology and have confidence in our human capacity for irrationality.

The great pleasure of being a person talking to another person is the potential to make jokes, be a bit outrageous, indulge your prejudices slightly (without being a prosecco snob) and give range and air to your passions. There's something refreshing about forgetting that you're in a transactional relationship on the shop floor, and instead handing out cocktail recipes and recommendations for bars in the neighbourhood.

Gossip. Sarcasm. Laughter. These are the mortal energies that we can deploy in the face of the algorithms. Nothing is more challenging to the machines than contradiction; a quality wine and humans have in common.

Originally published in Drinks Retailing, August 2019


bottom of page