AutoRiposte

It starts with the flip of a bit uploaded to servers around the world. Reply suggestions appear in your chat app. A friend texts you, and beneath their words, three Gumby-edged rectangles appear, presenting you reply options. Would you like to respond with “Hello”, “Hey!”, or “‘sup?”?

You’re turned off by this the first time you see the chirpy replies inviting you to take the easy path. You refuse to play their game and defiantly type, “Hey!”

When you press send, you realize your response was identical to one of the suggestions. It’s irritating that they got it so right.

The second time you see this, the stone in your aorta rolls aside. You were going to type “Hey!” Anyhow, you might as well save yourself some keystrokes. RSI can be a bitch.

The AI that predicts your responses gets better over time. The most capitalized multinationals to ever exist fund Research labs. (See how I capitalize the ‘R’? This work is looked on with religious fervor in tech circles). Pipelines devour, separate, and pour every word that appears on the Internet into two-hundred-million-per processors optimized for high-dimensional matrix multiplication. Then the engineers take what comes up and through digital alchemy make it small enough to run on the supercomputer in your pocket, where you train it every time you select one of its suggestions.

The suggestions become so good that you almost never write out a reply anymore. You suspect your mom, friends, and lovers are using the canned replies, too, though you can’t be sure.

Another flag flip presents you with a new option to enable something they call “AutoRiposte.” Unless sensitive topics are detected, your phone can automatically reply with your predicted response after a reasonable stochastic delay.

It’s great for mindfulness, the marketing blurb proclaims. Reduces the interrupts in your day so that you can focus on your real life in front of you.

Sometimes, when you feel lonely at the end of the day, you’ll open your messages app and read conversations between the simulated you and your simulated contacts. You’ll marvel at how well you explained yourself, describing your various adventures. Your phone knows where you go and can make a decent guess at what you get up to there. When you read these conversations back, you swell with pride. The real essence of yourself is there in these conversations—true and authentic, but better put than you would have dashing off a reply in the toilet between meetings.

Perhaps the truest version of your friends are there, too.

Comments

Post a Comment

Popular posts from this blog

Just what my TBR stack needs...

 I visited the Bay Area Book Festival on Sunday, attending two sessions devoted to speculative literature. The earlier session, " Cataclysm or Cure-all: Novelists Engage with the Promises and Perils of Our Tech Future ," featured four writers whose work straddles the line between literary and speculative fiction. Vauhini Vara is a tech reporter for the Wall Street Journal, but in parallel has worked on her debut novel, The Immortal King Rao , for thirteen years. It's about a Dalit CEO in a dystopian state where citizens are shareholders. She also co-wrote an essay with GPT-3 on grief. Claire Stanford has somewhat of an outsider's perspective on the tech world, and her novel, Happy For You , explores why so many talented people expend their efforts in the tech industry. Her main character is a philosopher-in-residence at a tech company, creating a quantification-of-happiness app called "Joyful." Kate Folk published a short story collection, Out There , f
Read more

Reading Guilty Pleasures

My reading taste tends toward literary most of the time, and I gravitate towards difficult reads. I like a challenge and learning new words. I like there to be a deeper layer of meaning beneath the words on the page, and experience a thrill of discovery about the unreliability of a narrator. I love connecting various unrelated threads together in a way that's not explicitly spelled out by the author. Vagueness, ambiguity, and conflicting points of view are my jam. But right now, friends, I am exhausted. September/October 2020 has hit like so much of the rest of the year--unrelenting in its awfulness while at the same time passing miserably as more of the same. Every day requires a new evaluation of statistical tradeoffs, the type of thing that human brains excel at. I'm a manager, which means supporting my reports' existential dread while managing my own simultaneously. A Republican congress has defied the dying wish of a feminist icon and Supreme Court justice by shoving t
Read more

A Call for Milquetoastification

Image
There's a sentence in System Error  that contained so many disparate elements, I needed to unpack it: We must confront the algorithmic sorting of users into filter bubbles that contribute to growing polarization, extremism, and decreasing social trust, all of which threaten the health of democracy. This is stated baldly, and with no explanation, so I'd like to examine two questions: First, are these things related in a self-evident way? and second, what are the causal chains at play? Photo by  Alexander Dummer  from  Pexels First, let's consider filter bubbles. Is this a concept that the general public understands well? Perhaps, but let me explain it just to be clear. In the early days of mass media, there was no way to individualize news and entertainment content to tailor it to the individual interests and world views of each consumer. Outlets had their own ideological bent, but given the limitations of the platforms and market pressure to be as broadly applicable as poss
Read more