MACHINE EARNINGSFEB 3, 2026

Rage Against The Slop

Notes on information diets

◆ ◆ ◆

I read two things that completely changed how I think about internet content: a post on the Dead Internet Theory, and Anu's piece on performative praise. Anu's post, in particular, felt like it was subtweeting my entire online life, so I won't even try to summarize it here—go read it first, then come back.

OK and we're back.

Anu's point, at least the way it hit me, is that once you factor in performative praise, we're already somewhere inside the Dead Internet Theory. Between "LFG!", "congrats goat," or pretending like we had anything to do with a funding round/exit/insert-generic-corporate-event here, we're just perpetuating the problem.

"There was no way I, Ian Kar, was just mainlining AI slop and engagement-optimized press releases like everyone else. I'm a writer, a reader, I told myself, I'm the guy with Opinions and Taste."

This fact honestly broke my brain. Part of my identity has always been around being an intellectual (I'm not, but I wish I was.) And I was letting that part of me down.

The truth is I get all my news from X and social media. I don't really subscribe to any magazines or longform writing anymore. All the books I read are work-ish related, and all the newsletters I read now are for stock market investors or AI product stuff. But all that is extremely promotional content. None of that is neutral; if anything I'm basically just reading glorified marketing slop.

It's a bit embarrassing when you realized you've outsourced your brain to a bunch of founders, VCs, and hype men. In a way it makes perverse sense; there's too much noise in the world today and the only way to stand out is to be loud and frequent. Short form content that's been shoved down our throats has fried our brains.

I ask AI to summarize everything. If it's not in bullet form, I don't want it. The longest thing I read this week were the Epstein emails, which made me feel even worse that I was spending so much time voyeuristically reading these painful and horrific stories.

◆ ◆ ◆

My boomer fix was to subscribe to the Financial Times, the Serious Paper for Serious People, at an extremely stupid price point. It was…fine? Very European, obsessed with Trump wanting Greenland and open letters from Nordic royalty. Which, whatever. I do want to be well read about the world news, but I'm also interested in hyper niche topics too, like AI and banking, prediction markets, weird market microstructure.

I still begrudgingly keep the subscription, but it became painfully clear that neither the FT, nor the Journal, nor Perplexity, nor anyone else was going to hand me the specific, extremely personalized info feed I wanted.

At a certain point you either keep complaining about the sludge or you start cooking your own.

So I decided to make it myself. Welcome to Machine Earnings, my attempt to build a small, human-curated, slightly unhinged signal in a very dead internet. It's dedicated to keeping you (and me) up to date on the Machine Economy—because if the machines are going to run capitalism, we should at least understand the operating manual.

◆ ◆ ◆

Machines already run the economy. Humans are still very much important—we still coordinate, orchestrate, and supervise these systems. But anyone on the ground knows the truth: if a line of work can be handed to software or an agent, it eventually will be.

A lot of the technorati tell you to work insane hours so you don't end up in a "permanent underclass," which is a convenient story when you're the one with the equity. The more uncomfortable truth is that AI is going to create entire categories of work that don't exist yet—training agents, auditing models, architecting machine workflows—and millions of people are going to end up working there, whether they meant to or not.

If we sleepwalk through it, our descendants will wake up in a neofeudal society working for Lord Altman of California or King Musk the 5th, clocking in to earn OpenAI or X credits to pay for their glitchy, government-subsidized AI services, extremely pissed as to why their great-great-grandfather spent the equivalent of the American Gold Rush scrolling through TikTok.

We owe it to ourselves to get smarter and more self-reliant. Luckily with AI, all that becomes a lot easier. Using agents to sort through the never ending firehose of information between traditional news, social media, blogs, newsletters, YouTubes, and all the private communities and chats I'm in makes information curation a lot easier. It also becomes something a lot more sharable and engaging too.

WHAT TO EXPECT

Two briefings a week — Tuesdays and Thursdays on what's happening in the machine economy

5x-a-week firehose — For those who want the full signal feed

Let's get to work.

— Ian

MACHINE EARNINGS

SIGNAL IN A DEAD INTERNET

© 2026 Machine Earnings