Essay
The right amount of attention
You can lose an hour and have nothing to show for it. A clip. A quote-post. A chart with the axis cut off. A row about a stranger. None of it asked you to change your mind. It only asked you to stay.
Herbert Simon named the trap in 1971, before a feed existed. "What information consumes is rather obvious: it consumes the attention of its recipients. Hence a wealth of information creates a poverty of attention and a need to allocate that attention efficiently among the overabundance of information sources that might consume it."[1] People, he wrote, are like the computers of his day: serial devices. They attend to one thing at a time.[1:1] And in an information-rich world, "most of the cost of information is the cost incurred by the recipient."[1:2]
That cost is the whole business of this publication. Attention is scarce. The timeline does not price it. We will.
What "right" means
Right attention is proportional attention.
A topic deserves your time in proportion to three plain questions. How much can it change a decision or an understanding that matters: health, money, a skill, a relationship, or the way the world actually works? How costly is it to ignore, including whether the window closes? And is the claim true enough that acting on it is rational?
Virality measures a different quantity. It measures how much attention a thing has already collected. Chamath Palihapitiya, on The Joe Rogan Experience #2494, put the split in ordinary speech. A third-party transcript, not official captions, records him saying that "sometimes what is true and what people pay attention to are not the same thing," and that the thing you should be watching "gets lost because the thing that you are paying attention to gets more attention because it's more interesting and useful."[2] Later in the same conversation he says society tells you, in subtle ways and in ways that bash you over the head, that "attention is just the most precious asset."[2:1] He does not, in the lines this essay relies on, call it a commodity. The precious-asset line is enough, and it is his.
The same episode walks thirty years of technology with one word in the middle. Search engines learned what people clicked. Social feeds counted who had more likes than whom. Then a 2017 paper, "Attention Is All You Need," proposed a network built solely on attention mechanisms, "dispensing with recurrence and convolutions entirely."[3] On the WMT 2014 English-to-German task the model reached 28.4 BLEU, more than 2 BLEU over the existing best results, including ensembles. On English-to-French it set a single-model mark of 41.8 BLEU after training for 3.5 days on eight GPUs.[3:1] The paper is about machine translation. The rhyme with the feed is the point Palihapitiya was making. A rhyme is not a proof that your evening belongs inside a model.
He also said some people treat attention as an absolute value. Positive or negative, the number counts.[2:2] That is how a nonsense story outruns a useful one. It is also why a serious publication still covers nonsense. Silence does not shrink a crowd. Naming the trick, briefly, and pointing at what the trick crowded out, is the useful move.
A score, not a mood
Taste is not a method. Here is a method. It is a rubric for editors, not a law of nature, and it will be revised when the posts teach us we were wrong.
Score a topic on three questions. Each runs from 0 to 5.
Consequence. If the claim is true, how far does a real decision move? A central-bank move that changes a mortgage is a 5. A stranger's airport outfit is a 0 or a 1. A new training result that would change how a working engineer spends next quarter sits in the middle, and only climbs if the result is the sort of thing a careful person would bet a week on.
Confidence. How sure should a careful reader be, this week, that the claim is true? A paper with a table you can open is high. A screenshot of a screenshot is low. Naval Ravikant's rule belongs in this slot. "You have to reject most advice. But you have to listen to enough of it, and read enough of it, to know what to reject and what to accept."[4] He adds the constraint Simon would have recognised: "Your brain space is finite. You have finite neurons."[4:1] Confidence is how those neurons stay off the books of every confident stranger.
Use. Can a specific person do something different on Monday? Change a savings rate, a training plan, a hiring bet, a reading list. If the honest answer is "be entertained," Use stays low. Entertainment is allowed. It does not get to pretend it was a briefing.
Deserved attention is the average of those three, rounded to the nearest whole number from 0 to 5. Received attention is how much of the public conversation the topic already occupies, also 0 to 5. A topic leading the trend list is a 5. A careful paper three specialists have read is a 1. Do not fake extra precision. A 4 is a 4. The argument that justifies it has to fit in a sentence.
The number that matters is the gap. Deserved minus received.
A positive gap is under-attended signal. That is the main job.
A negative gap is attention that outran the facts or the stakes. Cover it short. Name the mechanism. Redirect.
Both high: the town hall is already underway. Explain the thing cleanly. Do not add heat.
Both low: skip it. Skipping is the product. Simon again: "Information does not have to be attended to (now) just because it exists in the environment."[1:3] The design principle he wanted was different from "the more information the better."[1:4]
Worked example, labelled as an illustration of the rubric, not as a study. Take a widely discussed celebrity dispute and a dull statistical release that revises how households actually hold wealth.
The dispute might score Consequence 1, Confidence 2, Use 0. Average about 1. Received attention 5. Gap: minus 4. One post. What the clip is doing (status, absolute-value attention, a sugar high). What a reader should look at instead.
The wealth release might score Consequence 4, Confidence 4, Use 3. Average about 4. Received attention 1, because it has no face. Gap: plus 3. That is a main-slot piece. The Federal Reserve's own table, discussed below, is the kind of document that lives in this slot while a louder story eats the afternoon.
The mix
A feed made only of signal never meets people where they already are. A feed that is half spectacle becomes the thing it set out to price.
The target mix is 60 percent under-attended signal, 25 percent stories that deserve the noise, and 15 percent spectacle used as a teaching case.
In a week of seven posts that is four, two, and one. The percentages land near 57, 29, and 14. The round numbers are the aim. The week is the practice.
Why keep a quarter of the week on stories that are already loud? Because the interesting thing crowds out the true thing, which is Palihapitiya's observation, and because a publication that never touches the timeline people actually inhabit will not be read by them.[2:3] Those posts are the town hall. They take a subject that is already noisy, strip the herd's frame, and say what the mechanism or the table supports. Joe Rogan's show is one version of that room: long, curious, uneven, watched by the same ambitious adults this brand is for. The job here is the shorter cut. Same instinct to stay in the public argument. A tighter standard for what earns the minutes.
Why cap spectacle at fifteen percent? Because attention is an absolute value for the platforms, and it will become one for us if the scoreboard is allowed to vote. A single joke about a viral nothing can teach the gap. A week of jokes teaches the audience to come back for the nothing. Fifteen percent is enough to stay honest about the feed.
Palihapitiya called the chase a sugar high. What people want, "even if they know they don't want it," he said, is purpose and meaning.[2:4] That is a claim about human beings, not a metric we have measured. The spectacle slot still has to point somewhere. If it cannot point, it does not ship.
Time is the budget attention spends
Bill Perkins built Die With Zero around a reallocation. The book's site states it this way: "Most people optimize the wrong number. It isn't net worth. It's net fulfillment."[5] Experiences, the site argues, pay memory dividends, "again and again, in the recall of it. The earlier you spend on memorable experiences, the more dividends compound."[5:1] Time buckets make the constraint physical. "You can't ski at 80. You can't backpack with toddlers at 60."[5:2] "Your peak life-experience years aren't the same as your peak earning years."[5:3]
A line on the same site, presented as Perkins's, says: "My money philosophy is that money is a tool to drive your fulfillment, and that's it."[5:4] Another, attributed there to Perkins speaking to Iowa Magazine: "Going through life like it lasts forever is suboptimal."[5:5] On Peter Attia's podcast the show notes quote him directly: "In the game of life, what I'm solving for is regret minimization. . .I want the highest score in net fulfillment."[6] The notes also say that most people are over-saving and under-living.[6:1]
This essay is not a personal-finance plan, and it does not borrow Perkins to smuggle in a claim he did not make. The bridge is narrower. Money is one tool. Attention is how the hours get spent. You can finish with a large portfolio and a thin stock of memory. You can also finish having watched thousands of clips that never entered a decision. The site calls dying with unspent savings "a measurable life loss," and in one section illustrates that with "$200K of unspent savings."[5:6] That figure is the site's illustration. It is not a finding from the Survey of Consumer Finances, and it should not be repeated as one.
The site also prints "Median U.S. net worth at age 70–74: $225,390," with no table under it.[5:7] Do not use that number. The Federal Reserve's published band is different. In the 2022 Survey of Consumer Finances, Table 2 reports family median and mean net worth in thousands of 2022 dollars. For a reference person aged 65 to 74, the 2022 median is 409.9. That is $409,900. The 2019 median for the same band was 308.8, or $308,800. The mean in 2022 was 1,794.6 thousand dollars.[7] There is no separate 70-to-74 cell on that table. Net worth, the Fed notes, is assets minus liabilities.[7:1] A household can sit on that median and still have spent its sixties on a feed. The stock of money and the stock of attention are different accounts. Proportional attention is the second account, kept with the same seriousness Perkins asks people to bring to the first.
His own early life makes the time point without a sermon. Attia's show notes are the host's summary, not a transcript of Perkins speaking. They say he played football at the University of Iowa as a walk-on defensive back, mostly on the bench, with a partial scholarship, until he broke his leg at the growth plate and the football career ended.[6:2] The window was short. That is one person's version of a pattern the labour data also shows.
The group version is a National Bureau of Economic Research working paper, not a peer-reviewed journal article: Carlson, Kim, Lusardi, and Camerer, on NFL players drafted from 1996 to 2003, people whose income spikes for a few years and then stops.[8] "A career lasting 6 years (the median length) will provide an NFL player with more earnings than an average college graduate will get in an entire lifetime, plus a modest pension."[9] Even so, the abstract's conclusion is stark. "Bankruptcy rates are not affected by a player's total earnings or career length. Having played for a long time and been well-paid does not provide much protection against the risk of going bankrupt."[8:1] Figure 3 in the paper puts a number on the clock: "1.9% have filed for bankruptcy by year 2" and "15.7% have filed for bankruptcy by year 12."[9:1]
Read that against the legend. The paper is a warning about a short, loud earning life. It is not a licence to repeat the claim that most players are broke within two years. The two-year filing rate in this sample is 1.9 percent.[9:2] The useful attention is the twelve-year curve, and the finding that pay and tenure did not buy safety. The wasted attention is the round number that flatters a morality tale. An earlier private memo on athletes, this book, and the Rogan conversation can become a series later. One sharp case is enough for a founding argument.
Athletes are a vivid instance of a general fact. A season, a funding round, a health scare, a bull market: each arrives with a spotlight and a clock. The spotlight is received attention. The clock is why deserved attention is sometimes high while the crowd is looking elsewhere. You do not get the season back. Perkins's line about skiing at 80 is the same sentence in civilian clothes.[5:8]
Simon, writing for managers rather than for fans, wanted the information system to cut time, not to pile up paper. "The proper aim of a management information system is not to bring the manager all the information he needs, but to reorganize the manager's environment of information so as to reduce the amount of time he must devote to receiving it."[1:5] A publication that lives by this rule will disappoint people who want everything. It will be useful to people who want their Mondays back.
What this will not spend the hour on
Party politics, and the culture-war bait that travels like politics, stay out of the mix. Speech stays free. The publication simply refuses to spend its scarce slot on a team jersey and then call the spending insight. The audience already has plenty of places to do that. They do not have many places that will tell them a loud story was the wrong size.
Every factual claim needs a link that was actually fetched. If the fetch fails, the claim does not ship. That rule will make the account slower than the trend list. Slower is the point. Simon's cost sits on the reader. Shipping an unsourced number transfers that cost, plus the cost of being wrong, onto someone who trusted the tone.
The rubric is not wisdom. It is a way to make a choice in public so a reader can argue with the score. If spectacle posts pull about three times the engagement of signal posts for three straight weeks, the rubric is being trained by the feed, and the feed is a bad teacher. That result is an escalation. Auto-publishing pauses. The mix gets rewritten. The voice gets rewritten. Measuring is for changing your mind, not for framing a dashboard.
Hold the slogan next to the table
Palihapitiya described a society that hands people one incentive and repeats it until it sounds like physics. Attention, then attention again.[2:5] The 2017 paper's title left machine translation and became a slogan. Slogans are received attention. The paper's actual result is a translation score and a training-time figure.[3:2] Hold both. The slogan tells you what the culture is rewarding. The BLEU score tells you what the authors measured. Right attention is the habit of asking which of those you are looking at.
You have a finite life and, in Ravikant's phrase, finite neurons.[4:2] You also have a timeline that will spend both if you leave the allocation on default. Proportional attention is the refusal, practised in public. Give the larger share of the hour to the thing that can change a Monday. Spend a thin slice on the noise, so the trick stays visible. Leave the rest unspent on purpose.
Then do it again tomorrow, and correct the score when you were wrong.
Notes
Herbert A. Simon, "Designing Organizations for an Information-Rich World," in Martin Greenberger (ed.), Computers, Communications, and the Public Interest (1971). Quotations from the Gwern PDF of the chapter, book pp. 40–41 and p. 44: https://gwern.net/doc/design/1971-simon.pdf ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
The Joe Rogan Experience #2494, Chamath Palihapitiya. YouTube title confirmed by oEmbed on 8 October 2026: https://www.youtube.com/watch?v=LSihotD-PQA. Quotations are from the Singju Post transcript, a third-party transcript, not official captions: https://singjupost.com/transcript-chamath-palihapitiyas-interview-on-joe-rogan-experience-2494/ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin, "Attention Is All You Need" (2017). Abstract: https://arxiv.org/abs/1706.03762 ↩︎ ↩︎ ↩︎
Naval Ravikant, in conversation with Nivi, "Reject Most Advice," 20 May 2019: https://nav.al/reject-advice ↩︎ ↩︎ ↩︎
Die With Zero official site, fetched 8 October 2026: https://diewithzerobook.com/. The site's "$225,390" median for ages 70–74 is quoted here only to reject it. The "$200K of unspent savings" line is the site's illustration of a "measurable life loss," not a Federal Reserve figure. ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎
Peter Attia, MD, episode #237 with Bill Perkins, 9 January 2023. Show notes: https://peterattiamd.com/billperkins/. Biographical football details are the host's notes. ↩︎ ↩︎ ↩︎
Board of Governors of the Federal Reserve System, "Changes in U.S. Family Finances from 2019 to 2022," October 2023, Table 2. Age of reference person 65–74: 2019 median 308.8, 2022 median 409.9, 2022 mean 1,794.6, in thousands of 2022 dollars. https://www.federalreserve.gov/publications/october-2023-changes-in-us-family-finances-from-2019-to-2022.htm ↩︎ ↩︎
Kyle Carlson, Joshua Kim, Annamaria Lusardi, and Colin F. Camerer, "Bankruptcy Rates among NFL Players with Short-Lived Income Spikes," NBER Working Paper 21085, April 2015. Abstract page: https://www.nber.org/papers/w21085 ↩︎ ↩︎
Same paper, PDF. Median career length and the lifetime-earnings comparison are in the introduction. The 1.9 percent and 15.7 percent figures are printed on Figure 3. https://www.nber.org/system/files/working_papers/w21085/w21085.pdf ↩︎ ↩︎ ↩︎