Reading List

Scanlon's argument: today's graduates are booing speakers because they trained their entire lives for a credentialing system that's breaking down underneath them. The on-ramp jobs that used to teach junior employees are being automated away. Her framework is "dimensionality" — high-dimensional jobs (many tasks, judgment, accountability, relationships) stay safer than narrow ones.

Many college commencement speakers are getting booed off stage for delighting in the productivity of AI in front of the audience it threatens to displace. The asymmetry is the thing being booed. These students have been training their entire lives for a credentialing system that is now being dismantled in the name of efficiency and profit. They aren't anti-technology, in the same way the luddites weren't anti-technology. They are anti-technology-without-solution and anti-future-without-hope.

What is a Job in the Age of AI?

In 2019, a human read my resume. In 2026, that doesn't always happen. Humans have more or less exited the job process. The hiring funnel has been almost entirely intermediated by AI on both sides — a "plexiglass wall" with both applicants and companies throwing technology at the process, neither side really seeing each other.

In recent years, entry-level hiring has been depressed. Unemployment for college graduates aged 22–27 is 5.6%, up sharply from 3.6% before the pandemic. Part of it is this AI confusion/problem/opportunity — an April 2026 paper finds that when AI absorbs the work juniors used to do, firms cut junior hiring, which weakens the pipeline that creates future seniors. The result is "lost cohorts."

Alex Imas says the dimensionality of a job really matters — the number of distinct tasks it involves. A job with two tasks, one of which AI can do, is high-risk. A job with seven tasks, one of which AI can do, gets more productive — the worker has more time for the six tasks that can't be automated. Imas calls this the focus effect: productivity gains that raise wages instead of eliminating the role.

He compares a long-haul truck driver to a consultant. A truck driver could be automated, because it's theoretically one task — going from A to B. A management consultant has a bundle of tasks including being a scapegoat for companies — some can be automated, but political work, client communication, and presentation of results cannot.

Ernie Tedeschi uses travel agents as the example. In 2000, the job was exposed to automation — booking, fare aggregation, ticketing — and no one was safe. Travel agent headcount fell by more than 60% from the dotcom peak. But some survived by moving up market. They now earn close to 99% of the private-sector average wage, up from 87% in 2000. Those that survived leaned on judgement, accountability, relationships, and the ability to swoop in when things go wrong at 2am.

The pillars of strength in the AI age: being able to evaluate, having some element of taste, holding accountability, building out networks, and the capacity to deliver a human touch in an increasingly automated world.

My take: AI is sort of like freeze-dried camp food. It does a good enough job, is filling enough, but it isn't something you want to eat every day. People crave inefficient, home-cooked meals for a reason — not because of macros, but because it's a nice thing.

Is AI Even Cheaper than Labor?

The technology is very expensive. Much of our assumptions on AI displacement seem to be based on the idea that it is cheaper than labor. Right now, that assumption is just not true. Agentic workflow is more compute intensive than a chatbot. In April, Uber's CTO said he had to go "back to the drawing board because the budget I thought I would need is blown away already." Over 70% of companies exceeded their AI budgets in 2025.

Companies have fired people in the name of AI efficiency, but in many cases, it hasn't really made anything more efficient. Uber's COO recently said that "tokenmaxxing" was making it "harder to justify AI costs within the company" because more tokens spent by engineers wasn't leading to a measurable increase in "useful consumer features."

Are Firms Actually Using AI?

Some firms are. Six of the largest US banks posted $47 billion in profits in Q1 2026 while shedding 15,000 jobs, with CEOs openly crediting AI on earnings calls. Some firms are using humans to train the AI to eventually fire themselves — Meta laid off 8,000 employees in an AI-push while assigning 7,000 to "AI-initiative focused teams." Kirkland & Ellis announced a $500 million investment to build its own AI platform from interviews with 250 of its lawyers — the senior generation pouring its knowledge into a system that will eventually do the work that juniors used to do to learn the job.

But only 1 in 5 firms across the board are actually using AI. The economy can absorb AI the way it absorbed previous technologies — slowly, unevenly, with new sectors appearing at the same time that the rungs of the economic ladder get squeezed.

Who Gets to Be Wealthy in the Age of AI?

About 10,000 people working at Anthropic, OpenAI, Nvidia and Meta have hit retirement-level wealth — more than $20 million dollars — within the past five years. OpenAI and Anthropic are sniffing around IPOs, but both are near $1T private market valuations, leaving little room for upside. SpaceX just filed to go public after 24 years — unprofitable on $18.7 billion of revenue, losing almost $5 billion last year, IPO-ing at a $1.5 trillion valuation with Musk holding 85% of the voting power.

If you don't own equity here, your participation in the AI boom is entirely through the lens of: is this thing going to destroy my life?

What Do You Do About It?

1. Pick the manager, not the company. The training pipelines that used to build up juniors have been gutted. The single biggest variable in whether your first job teaches you anything is the person who manages you.
2. The search is the job. One warm introduction is worth fifty cold applications. If you don't have a network, the work of building one is the work.
3. Lean into creativity. Firms want unique thinkers who can look at hard problems and come up with novel solutions. The goal isn't to compete with the AI, but to be the person who knows what to do when it all stops working.
4. Build something. Develop evidence that you can think in public.
5. Be AI native, not AI proof. The jobs that pay the most tend to be the most AI-exposed.

The despair and the doom and the booing and the financial nihilism are all rational responses to being told that you are behind on a schedule that doesn't really apply anymore. Bounded downside, a predictable floor, reward for work, and a believable path to the next thing are the conditions under which a future feels possible. For a lot of people, none of that holds.

Condensed from the original ~8,000-word essay.

The Wired co-founder on founding the magazine, the origin of "1,000 True Fans," and his core thesis that the only way to steer technology is to actively use it. Plus his concept of "protopia" — incremental 1%-better progress instead of utopia.

On building Wired in the early days of the web

The thing that a lot of people did not believe was that people would read online. And secondly, people don't remember, but in the '80s, there were a lot of people saying that writing was over. Nobody was going to write. What we learned from the web was: no. People are going to write. Almost any of us writes far, far more words per day than our grandparents ever did. We're all writers, really.

On the origin of 1,000 True Fans (2008)

The premise is that if you have direct contact with your audience or your customers, you don't need millions of them to make a living. If you have a label or a publisher or a studio in between you, then you might need millions. But if you take those away and go directly to them, you have a much more feasible number. $100 per year from 1,000 true fans, you can make it. I defined a true fan as somebody who would purchase anything you made.

Even the weirdest, wackiest, esoteric idea that only appeals to one in a million people, with 8 billion people in the world, there's still a thousand people who are going to be into your weird thing. The challenge is finding them and making that connection.

On the inevitability of AI — but not its specifics

AI is inevitable. Any civilization anywhere in the galaxy that invented electricity and steam engines and motors is going to make AI. But the character of the AI is not inevitable. Quadrupeds are inevitable on any planet with our kind of gravity, but a zebra is not. The specificity is not inevitable. How it shows up, the particulars, are completely up to us.

Who owns AI? Is it public? Is it international? Is it closed? Is it open? How is it financed? All those choices are choices we have. But AI itself is inevitable.

On steering technology by using it

The way to steer technology is through use. You want to embrace technologies and use them, because that's the only way you get to steer them. If you prohibit them or ban them or refuse them, you don't get to steer.

Kelly calls the opposite view thinkism — "this idea that just thinking about things will solve stuff. We can't. We have to actually use it."

On protopia

We aren't headed to utopia, where everything's fine. We're going to come to a world where things are a little tiny bit better, and we're just incrementally prototyping our way forward. No massive jumps. Just 1% better, 1% better. That means there could be 49% crap in the world, 49% harm, and everybody knows it. But there could be 51% good. That 2% difference is hardly even visible.

On prototyping your life

"The idea of prototyping your life rather than deciding what you're going to do for the rest of your life. You try it for a couple of years and you go on. You prototype everything out of cardboard first. You make one to throw away, as the makers say."

Postmortem on Everlane and the D2C "good brand" promise. The real find inside the column: lo-fi / anti-AI aesthetics — artists and designers whose livelihoods are threatened by AI are aggressively embracing the handmade, messy, and accidental as a visible signal of anti-AI work. Also: a booming shadow market of sketchy Anthropic-equity vehicles, and why local newsletters get 2.5x the click-through of normal ones.

Everlane and the Death of the "Good" Millennial Lifestyle Brand

"Slop" is fast becoming the central metaphor for the cultural output of the late twenty-twenties. A.I. produces slop, but so do vampiric private-equity investors and rapacious billionaires, through corporate consolidation and hollow brand-licensing deals. The central promise of the millennial life-style brand was that, with the help of digital technology, new businesses could produce higher quality by cutting out the layers that had accumulated between producers and consumers. In the Shein era, the offering has declined to unapologetic slop, an ephemeral transaction that loses its appeal once a cycle of online attention has moved on.

A Booming Shadow Market of Sketchy A.I. Investments

Want a tiny piece of Anthropic equity? Some people will sell their houses for it. If you don't have a major hookup, the only way to go may be backroom deals — investments in investments in investments in Anthropic, three layers removed from the engineer who actually owns the equity.

If A.I. is going to create a "permanent underclass," as many prognosticators fear, then what choice is there but to try to cash in somehow while you can? One reason for the secondary-market fever is that OpenAI and Anthropic have stayed private for too long. "If and when these large companies go public at one-trillion-dollar valuations, retail is going to get hosed."

Your Friendly Local Neighborhood Newsletter

The e-mail newsletter platform Beehiiv has found that the click-through rate for neighborhood newsletters is two and a half times higher than that of the average newsletter across the platform. The more hyper-local and hyper-niche you make your content, the more engaged your audience might be.

A Lo-Fi Rebellion Against A.I.

This instantaneous slickness — which the trend forecaster Emily Segal has labelled "tasteslop" — is easy and abundant, and the appearance of "good taste" is robotically ubiquitous. So it's no wonder that the artists and designers whose livelihoods are threatened by it have begun moving aggressively in the opposite direction, toward the scrawled, the sloppy, the seemingly mistake-riddled.

The size of the deal alone is the story — private credit financing AI compute at a scale that's starting to look like its own asset class. Pair it with Dell raising its AI server sales outlook to $60B for the year, and you have the clearest picture this week of how much capital is being routed through hardware to chase model performance.

CLO managers like Blackstone and Guggenheim are now marketing software-light loan portfolios because investors fear AI will obsolete software makers. This is the first time the AI-disruption thesis has shown up structurally in fixed-income deal design, not just in equities. Salesforce's lukewarm outlook the same day fed the narrative.

From this week's roundup, the Deakin piece is the one to read first. Also strong: Zadie Smith on the courage to be more than yourself, and Virginia Woolf on truth, fact, and how we come to know reality.

One of my earliest and most vivid childhood memories is of swimming in a cool pool bounded by boulders in the middle of a river in the mountains of Bulgaria, the late-afternoon sun casting komorebi on the water through the rustling leaves. I can still hear the feeling-tone in my body, the strange and lovely simultaneity of absolute presence and absolute peace. I didn't yet know the word for transcendence.

Not long after that, I began swimming competitively in a chlorinated Olympic pool, investing long hours in perfecting my stroke and bettering my lap times. Those four years became a hard initiation into a culture that prizes productivity above presence. At eleven, I was beginning to see how the moment we incline action toward achievement, we drain the activity of joy; how anything we approach transactionally will never yield transcendence.

This spiritual dimension of swimming in wild nature comes vividly alive in Roger Deakin's delicious book Waterlog: A Swimmer's Journey Through Britain.

Such indelible swims are like dreams, and have the same profound effect on the mind and spirit. All water, river, sea, pond, lake, holds memory and the space to think.

Our profound response to water appears to be our evolutionary inheritance — we came out of the ocean, of course, but never fully. Drawing on Sir Alister Hardy's aquatic theory of human evolution, Deakin writes:

We spent ten million years of the Pliocene era of world drought evolving into uprightness as semi-aquatic waders and swimmers in the sea shallows and on the beaches of Africa. We went through a sea change to become what we are, and our subsequent life on dry land is a relatively recent, short-lived affair. Apart from the proboscis monkey of Borneo, we are the only primate that regularly takes to the water for the sheer joy of it.

When you swim, you feel your body for what it mostly is — water — and it begins to move with the water around it. The swimmer experiences the terror and the bliss of being born. So swimming is a rite of passage, a crossing of boundaries: the line of the shore, the bank of the river, the edge of the pool, the surface itself. When you enter the water, something like metamorphosis happens. Leaving behind the land, you go through the looking-glass surface and enter a new world, in which survival, not ambition or desire, is the dominant aim.

Most of us live in a world where more and more places and things are signposted, labelled, and officially "interpreted." There is something about all this that is turning the reality of things into virtual reality. It is the reason why walking, cycling and swimming will always be subversive activities. They allow us to regain a sense of what is old and wild — by getting off the beaten track and breaking free of the official version of things — to access that part of our world which, like darkness, mist, woods or high mountains, still retains most mystery.

Substack interviews Kate Lindsay, who named the "millennial pause" and has built her career describing what the internet is currently doing. Paywall strategy, Stuart Little memes, and how moving to Substack let her write posts with "no rules."

On her best-performing post

It's never the ones you expect. As for why it hit — I think you can't overestimate the value of articulating a simple truth. I can sometimes hold myself back by not writing something purely because I assume it's too simple or something people already know. But even if people already know something, it's validating to see it in writing.

On what's kept Embedded going

When Nick and I got laid off, we decided to start a Substack to keep ourselves sane during another pandemic summer. We treated Embedded like our full-time job while still receiving severance. It was exciting to finally have no rules and explore all types of content, and I think this freedom really fueled us those first few months. I had spent much of my early career fighting for success within traditional media without much growth, and ironically, once we went independent, that's when those editors and publications found my work.

On the newsletter feeding her other work

I see Embedded and ICYMI as practically sisters at this point. Often, I'll write something for Embedded and not even realize it helped me work out my thoughts for an adjacent episode. While there's overlap in readers and listeners, Embedded will always be my rock. It's the reason I feel even a marginal sense of safety in an unpredictable media landscape.

On her advice for building on Substack

Every day I am grateful to not be beholden to an algorithm that no one can explain. My advice isn't sexy, but it's real: Write for your readers, and if you don't have any yet, write the newsletter you want to read. Write it consistently even on the weeks you don't feel like it. And don't use AI. If you're using AI to write, I don't understand why you're here.

On voice notes

I am pro voice note, but I am glad to finally have a space to complain about how maddening it is that listening to a voice note doesn't register as activity on an iPhone.