NEW BOOK!
Explore a better way to work – one that promises more calm, clarity, and creativity.

On Ultra-Processed Content

When I visited London last month, a large marketing push was underway for the paperback edition of Chris van Tulleken’s UK bestseller, Ultra-Processed People: Why Do We All Eat Stuff That Isn’t Food…and Why Can’t We Stop? It seemed to be prominently displayed in every bookstore I visited, and, as you might imagine, I visited a lot of bookstores.

Unable to ignore it, I eventually took a closer look and learned more about the central villain of van Tulleken’s treatise: ultra-processed food, a term coined in 2009 as part of a new food classification system, and inspired by Michael Pollan’s concept of “edible food-like substances.”

Ultra-processed foods, at their most damaging extreme, are made by breaking down core stock ingredients such as corn or soy into their basic organic building blocks, then recombining these elements into hyper-palatable combinations, rich in salt, sugar, and fat, soaked with unpronounceable chemical emulsifiers and preservatives.

As Chris van Tulleken points out, the problem with ultra-processed foods is that they’re engineered to hijack our desire mechanisms, making them literally irresistible. The result is that we consume way more calories than we need in arguably the least healthy form possible. Give me a bag of Doritos (a classic ultra-processed food) and I’ll have a hard time stopping until it’s empty. I’m much less likely to similarly gorge myself on, say, a salad or baked chicken.

I was thinking about this book recently as Scott Young and I were prepared to re-open our course, Life of Focus, for new registrations next week. One of the three month-long modules of this course focuses on implementing ideas from my book Digital Minimalism to help you regain control of your attention from the insistent attraction of screens.

It occurred to me that in this concept of ultra-processed food we can find a useful analogy for understanding both our struggles to disconnect, and for how we might succeed in this aspiration going forward.

Read more

Manchester United Embraces Pseudo-Productivity

Earlier this month, Jim Ratcliffe, part owner and operations head for the storied English football club Manchester United, announced an end to the flexible work-from-home policy that the club’s approximately 1,000 employees had enjoyed since the beginning of the coronavirus pandemic. “If you don’t like it,” he said in a recent all staff meeting, “please seek alternative employment.”

Ratcliffe is not necessarily wrong to view remote work with skepticism. Having covered this topic extensively for The New Yorker, I don’t align myself with the crowd that automatically associates telecommuting with a self-evident pro-labor progressivism. Though I agree that flexible work arrangements will play an important role in the future of the knowledge sector, I also think that they’re hard to get right, and that we’re still in the early stages of figuring out how to implement them well — so for the moment, wariness is justified.

My problem with Ratcliffe’s return to office plan is instead the evidence he used to justify it. As reported by The Guardian, Ratcliffe supported his new policy by noting that when he experimented with a work-from-home Fridays program with another one of his companies, they measured a 20% drop in email traffic.

Read more

Do We Need AI to Revolutionize Work?

In recent months, I’ve been doing a fair number of interviews about my new book, Slow Productivity. I’m often asked during these conversations about the potential impact of artificial intelligence on the world of knowledge work.

I don’t talk much about AI in my book, as it focuses more on advice that individuals can put into place right now to escape busyness and find a more sustainable path toward meaningful accomplishment. But it’s a topic I do think a lot about in my role as a computer scientist and digital theorist, as well as in my recent journalism for The New Yorker (see, for example, this and this).

With this in mind, I thought I would share three current thoughts about the intersections of AI and office productivity…

Read more

Two Chances to See Me Next Week

I really enjoyed meeting so many of you at my Politics and Prose event a couple weeks ago. It was meaningful for me to be talking about my books in person again after having to launch my last title during a pandemic.

It was also a great opportunity to thank people for their support of Slow Productivity, which just yesterday landed #4 on the NYT’s monthly business bestseller list. (If you haven’t bought Slow Productivity yet, you should! If you read this newsletter, you’ll love it…)

It’s with all this in mind that I wanted to briefly share two upcoming opportunities to come meet me and hear me talk about Slow Productivity in the DC area next week!

Read more

Can You Tweet Your Way to Impact?

Earlier this month, a group of scientists from universities around the world published the results of an ingeniously simple experiment in the journal PLoS ONE. Every month, for ten months, they randomly selected an article from a journal in their field to promote on their Twitter accounts, which, collectively, added up to around 230,000 followers. They then later compared the success of these tweeted articles with control articles randomly selected from the same issues.

The result? No statistically significant increase in citations in the promoted articles versus the controls. There was a difference, however, in the download numbers: more people took a look at the tweeted citations.

Read more

ChatGPT Can’t Plan. This Matters.

A brief book update: I wanted to share that Slow Productivity debuted at #2 on the New York Times bestseller list last week! Which is all to say: thank you for helping this book make such a splash.

If you still haven’t purchased a copy, here are two nudges to consider: (1) due to the rush of initial sales, Amazon has temporarily dropped the hardcover price significantly, making it the cheapest it will likely ever be (US | UK); and (2) if you prefer audio, maybe it will help to learn that I recorded the audiobook myself. I uploaded a clip so you can check it out (US | UK).

Last March, Sebastien Bubeck, a computer scientist from Microsoft Research, delivered a talk at MIT titled “Sparks of AGI.” He was reporting on a study in which he and his team ran OpenAI’s impressive new large language model, GPT-4, through a series of rigorous intelligence tests.

“If your perspective is, ‘What I care about is to solve problems, to think abstractly, to comprehend complex ideas, to reason on new elements that arrive at me,'” he said, “then I think you have to call GPT-4 intelligent.”

But as he then elaborated, GPT-4 wasn’t always intelligent. During their testing, Bubeck’s team had given the model a simple math equation: 7*4 + 8*8 = 92. They then asked the model to modify a single number on the lefthand side so that the equation now equaled 106. This is easy for a human to figure out: simply replace the 7*4 with a 7*6.

GPT-4 confidently gave the wrong answer. “The arithmetic is shaky,” Bubeck explained.

This wasn’t the only seemingly simple problem that stumped the model. The team later asked it to write a poem that made sense in terms of its content, but also had a last line that was an exact reverse of the first. GPT-4 wrote a poem that started with “I heard his voice across the crowd,” forcing it to end with the nonsensical conclusion: “Crowd the across voice his heard I.”

Other researchers soon found that the model also struggled with simple block stacking tasks, a puzzle game called Towers of Hanoi, and questions about scheduling shipments.

What about these problems stumped GPT-4? They all require you to simulate the future. We recognize that the 7*4 term is the right one to modify in the arithmetic task because we implicitly simulate the impact on the sum of increasing the number of 7’s. Similarly, when we solve the poem challenge, we think ahead to writing the last line while working on the first.

As I argue in my latest article for The New Yorker, titled “Can an A.I. Make Plans?,” this inability for language models to simulate the future is important. Humans run these types of simulations all the time as we go through our day.

As I write:

Read more

Come See Me Saturday in DC + TikTok Falters

I know it’s been a minute since I’ve published one of my normal essays. I’ll be returning to these soon as the chaos of the Slow Productivity launch dissipates.

In the meantime, I wanted to share two quick notes: one about the book, and one about something interesting (but completely unrelated) that several of you have sent in my direction recently…

A note about the book

On Saturday, March 16th at 3:00pm, I’ll be appearing at Politics and Prose on Connecticut Avenue in Washington, DC. I’ll be joined in conversation with David Epstein, the New York Times bestselling author of Range. We’ll talk Slow Productivity and take questions from the audience. (For a preview, see my recent interview in Dave’s excellent newsletter.)

This is my first live event of the book tour, so if you’re in the DC area, I’d love to see you there! (More details.)

(You might also be interested in my most recent essay for The New Yorker, titled “How I Learned to Concentrate,” which discusses how my early years at MIT shaped almost everything I’ve written about ever since. I had fun writing this one: lots of Stata Center nostalgia!)

Read more

How the Acquired Podcast Became a Sensation

My podcast producer recently turned me onto a show called Acquired, which features its co-hosts, Ben Gilbert and David Rosenthal, diving deep into the backstories of well-known brands and companies, from Porsche and Nike, to Amazon and Nintendo.

It turns out I was late to this party. In the eight years since Acquired was originally launched, it has grown into a huge hit. The show now serves more than 200,000 downloads per episode. As Rosenthal revealed in a Fast Company profile last summer, they now face the problem of their audience becoming too large for their advertisers to afford paying the full fair market price for their spots.

What interests me about Acquired, however, is less what they’ve accomplished than how they did it. The conventional wisdom surrounding new media ventures is that success requires frenetic busyness. You need to produce content perfectly-tailored to your audiences’ attention spans, master The Algorithm, exist on multiple platforms, and above all else, churn out content quickly to maximize your chances of stumbling into vibe-powered virality.

Acquired did none of this. Gilbert and Rosenthal’s podcasts are very long; the two-part treatment of Nintendo I just finished clocked in at a little under seven hours. They also publish on an irregular schedule, often waiting a month or more between episodes. Combine this with the reality that they largely ignore YouTube and have no discernible social media strategy, this venture should have long ago crashed and burned. But it instead keeps growing.

Read more