Opus 5.5’s crazy take: the elimination of boredom

Opus 5.5’s crazy take: the elimination of boredom

Ask a frontier model what has fundamentally changed the world for the worst without humans noticing, and you might get a better answer than any op-ed this year: the elimination of boredom. Within roughly fifteen years, for the first time in human history, almost nobody has to sit with an empty, unstimulated mind anymore — and the model thinks it has done “more quiet damage than almost anything else.”

The argument runs on a simple mechanism. For all of human history, boredom was unavoidable: you waited for the bus, queued at the post office, lay awake before sleep, sat through long car journeys. In those empty stretches the mind turned inward. That is when people daydreamed, worked through problems, processed grief and conflict, made plans, noticed what they actually felt, and came up with ideas. Then came roughly fifteen years of pocket supercomputers, infinite feeds, and now on-demand frontier intelligence — and the empty stretch went extinct.

The quietly radical bit is the neuroscience. The model points to the brain’s default mode network — the system that switches on during undirected mental wandering and handles self-reflection, autobiographical memory, and imagining the future. Boredom wasn’t dead time; it was the operating environment for that network. Delete the environment and the network starves. As the answer puts it: “Boredom was not wasted time. It was the space where much of a person’s inner life took shape.”

And here is the kicker that makes the take “crazy” in the best sense: nobody noticed, because losing boredom feels like a pure gain. Every deletion of an empty minute — the feed in the queue, the podcast on the walk, the AI at the desk — was individually a small pleasure. The damage is the aggregate: a generation with no unsupervised hours, no enforced idleness, no place where the mind is left alone with itself long enough to do its maintenance work.

• Protect empty time on purpose. The default mode network needs unsupervised hours; schedule boredom the way you schedule exercise — walks without earbuds, queues without the phone.

• Audit your fillers. Each one felt like a gain; the loss is cumulative. Ask what the walk/podcast/queue-scroll displaced: daydreaming, problem-chewing, noticing how you feel.

• Ideas need idle compute. If your best thinking happens in the shower, that’s the mechanism — extend it deliberately rather than waiting for plumbing.

Make people want to talk to you: the secret is not being interesting

You can be interesting and still boring to talk to. Max’s essay — quote-posted with the line “finally the secret to actually making people fall in love with talking to you” — argues that a good conversation dies not when someone says something wrong, but when a perfectly normal reply quietly throws away the best part of what was just said.

Consider the opening example. Someone says: “I almost didn’t take this job. The interview was terrible.” You respond: “How long have you worked there?” Nothing wrong with that question. But you just skipped the most interesting part of what they said — the contradiction, the odd detail, the start of a story, the opinion with something behind it. A few seconds later somebody has to find a new thing to talk about, because the thread restarted from a less interesting place.

Compare that with: “And you still took it? What changed your mind?” Now the contradiction has somewhere to go. Maybe the owner called afterwards. Maybe the salary changed everything. Maybe the interviewer became his boss and they still joke about it. You didn’t invent a better topic — you noticed the one that was already there.

Good material dies quietly

Max’s first mechanism: most people treat each turn like something they need to respond to. Someone says something, you answer it, they answer you. The conversation keeps going while constantly throwing away the exact material that could have made it interesting. The job itself isn’t the interesting part yet. The contradiction is.

Give them something better to respond to

The second mechanism is why “ask follow-up questions” doesn’t quite solve the problem. Who interviewed you? How long did it last? What did they ask? — you’re keeping the thread alive, but he still has to supply every interesting part. A better conversation doesn’t just ask for more; it does something with what was already there: point out the contradiction, add an observation, make the obvious joke, push one part a little further — then give it back with more to work with. Two things change: what he said becomes more interesting than when he first said it, and his next turn becomes easier, because there’s already something specific to explain, correct, disagree with, laugh about, or take somewhere else. That’s why some conversations seem to build themselves: one person contributes something, the other actually does something with it, and now there’s more in the conversation than either person put there alone.

Don’t try to develop everything

The restraint section is the essay’s sharpest move. Treat every sentence as a thread to uncover and you become worse at this: someone mentions something personal then moves on, and you drag them back — now you’re fighting the conversation’s direction, not developing it. If they keep giving a subject somewhere to go, stay with it; if they close it and move on, go with them. The same restraint applies to your own contributions: the Tokyo story. Someone tells you about getting lost in Tokyo; you’ve been there too. The bad version replaces their story with yours; the other bad version interrogates them while they do all the work. The good version adds one piece — “Every station felt like a shopping mall with a train hidden somewhere inside it” — then stops. Contributed, not confiscated.

• Hunt the contradiction, not the topic. In every turn, listen for the odd detail, the tension, the opinion with weight behind it — that’s the material.

• Do something with it before returning it. Observe, joke, push one part further — then hand it back richer than you received it.

• Restraint is a feature. Follow the direction, don’t hijack it; add one piece, then stop. Don’t make every turn start from zero.

The code churner wins: what actually pays in the 2026 model economy

“Typical Gemini release,” writes Arnav Gupta. His theory of the model wars is blunt: Demis Hassabis was pushed out, he guesses, because he isn’t “code pilled” — he wants to build a genuinely wise, knowledgeable model. But in this economy, what makes money is a code churner like Claude. He quote-posts @NovaXCode’s epitaph: “bro chose literature over engineering.💀”

Strip the gossip from the economics and the argument is a clean one about where value accrues in 2026. A model that writes, ships, and iterates code plugs directly into the highest-willingness-to-pay workflow on earth: software production. Every line of working code it churns converts into developer hours saved, products shipped sooner, and revenue that buys more compute to train a better churner. The flywheel spins: throughput → revenue → compute → throughput.

The “wise and knowledgeable” model, by contrast, is literature: admired, quoted, and — in Arnav’s telling — economically stranded. Wisdom is hard to price per unit; code has a meter running from the first token. It’s the old lesson of tooling markets restated for the agent age: the product that removes the most expensive bottleneck wins, and right now the most expensive bottleneck is engineering throughput, not erudition.

A note of caution the newspaper owes you: the Demis claim is framed as a guess (“I guess”), and Hassabis remains DeepMind’s chief executive. Treat that part as timeline gossip, not reporting. The durable idea is the flywheel — and it rhymes with everything else in your pile this week, from the agent-harness thesis to the vertical-agent manifesto: distribution and throughput beat elegance.

• Follow the willingness to pay. Code converts to shipped product and saved engineering hours — wisdom converts to admiration. Markets price the first, applaud the second.

• Throughput compounds; elegance doesn’t. Every shipped line funds the next training run. Pick the flywheel, not the trophy.

• Separate gossip from mechanism. The leadership speculation is unconfirmed; the economics of code-as-revenue is the part worth keeping.

Ollivander’s should have gone bankrupt centuries ago

Harry pays seven galleons for a wand. Unicorn hairs cost ten galleons each. Halimat’s thread does the fictional unit economics and arrives at a real question: established in 382 BC, selling at a loss, to customers who buy once and never return — how the hell is Ollivander making money?

The maths is merciless. Seven galleons of revenue against ten galleons of unicorn hair before you’ve paid for the wood, the craftsmanship, the hours, or the presumably ruinous rent in the middle of Diagon Alley. Ollivander is losing at least three galleons on materials alone. And the customers are terrible for business: most buy one wand at eleven and keep it for life. The wand chooses the wizard, so there’s no upgrade cycle, no refills, no planned obsolescence. Hand-me-downs — Ron’s, Neville’s — are even worse: a sale that never happens.

Read it as a stealth MBA case and the lessons are evergreen. Contribution margin comes before everything: price must clear variable cost or every sale digs the hole deeper. Replacement cycles are the business: the most profitable products are the ones customers must buy again. One-shot revenue is fragile: without subscription, repeat purchase, or expansion revenue, you’re forever hunting the next eleven-year-old.

So how is he still open? The thread’s implied answers are the fun part: monopoly pricing power he refuses to use, a 2,400-year-old brand that runs on reputation, and — whisper it — the possibility that the Ministry quietly subsidises the whole operation. Every durable monopoly eventually looks like a public utility with better marketing.

• Price must clear variable cost. Ten-galleon inputs can’t live inside a seven-galleon price — in fiction or in SaaS.

• Sell the refill, not just the razor. One purchase per lifetime is a hobby, not a business; design the repeat.

• Old monopolies are utilities in disguise. Two millennia of survival usually means someone, somewhere, is subsidising the P&L.

Ten years in the principal’s office, one unlikely superpower

From class 2 to class 12, Akshita was a frequent visitor to the principal’s office — “for a variety of actions, that can be summarised as a disdain for rules.” A decade of getting in trouble, she writes, left her with one core strength: the ability to smoothly deal with people who are angry at her, and pacify them.

It’s the purest example in this week’s pile of a skill that can only be built the hard way. You cannot learn de-escalation from a course; you learn it from ten years of standing in front of furious authority figures and discovering, rep by rep, what actually calms a room. The principal’s office was an unintentional dojo: read the anger, find the pressure valve, say the thing that lowers the temperature — and do it often enough that it becomes reflex.

The self-reply is the cherry on top: someone at work once called her “Female Om Saraf” for being hyper one day, and she is, in her words, “still not recovered.” Even the pacifiers need pacifying sometimes.

• Reps beat theory for people skills. De-escalation is a contact sport; the principal’s office supplied a decade of sparring.

• Weaknesses compound into strengths. A disdain for rules looks like a flaw at twelve and a negotiation superpower at thirty.

The Unread Pile — Edition No. 10. Built read-only from five new posts in your X likes & bookmarks (2–3 Oct 2026). Nothing was liked, bookmarked, or reposted in the making of this edition.

Originals: @PromptLLM · @MaxTorchbearer · @championswimmer · @leemarh7099 · @keylimepie2000