davideisinger.com

My personal website
Log | Files | Refs | README

www-joanwestenberg-com-3uezea.txt (24522B)


      1 [1] [header-mar] > Westenberg. [MENU] [3] 1. Home [4] 2. About [5] 3. RSS [6] 
      2 4. Tools [7] 5. YouTube [8] 6. Forum | [9] / Search | [10] → Sign in [11] +
      3 Subscribe
      4 STATUS // operational
      5 Westenberg. | v1.0 | 2026
      6 2026-02-25 // 13 min read
      7 
      8 Everything is awesome (why I'm an optimist)
      9 
     10 AUTHOR // [12]JA Westenberg ACCESS // true
     11 Everything is awesome (why I'm an optimist)
     12 
     13 February is the month the internet decided we're all going to die.
     14 
     15 In the span of about two weeks, Matt Shumer's [13]Something Big is Happening
     16 racked up over 80 million views on X with its breathless comparison of AI to
     17 the early days of COVID, telling his non-tech friends and family that we're in
     18 the "this seems overblown" phase of something much, much bigger than a
     19 pandemic. Before anyone had finished arguing about that, Citrini Research
     20 published [14]THE 2028 GLOBAL INTELLIGENCE CRISIS (all caps) a fictional
     21 dispatch from June 2028 in which unemployment has hit 10.2%, the S&P 500 has
     22 crashed 38% from its highs, and the consumer economy has been hollowed out by
     23 what they coined "Ghost GDP": output that shows up in the national accounts but
     24 never circulates through the real economy, because, as Citrini helpfully
     25 observed, machines spend zero dollars on discretionary goods. Michael Burry
     26 signal-boosted it. [15]Bloomberg covered it. IBM fell 13%. Software and
     27 payments stocks shed over $200 billion in market cap in a single day,
     28 apparently because a Substack post called upon them by name and investors
     29 decided that constituted news.
     30 
     31 The doom loop Citrini described is simple: AI capabilities improve, companies
     32 need fewer workers, white-collar layoffs increase, displaced workers spend
     33 less, margin pressure pushes firms to invest more in AI, AI capabilities
     34 improve. Repeat until civilization unravels. Shumer, meanwhile, told people to
     35 get their financial houses in order because the permanent underclass is
     36 imminent.
     37 
     38 Both pieces went stratospherically viral, and both, I believe, are entirely
     39 wrong about where this is heading.
     40 
     41 I want to make a case for optimism.
     42 
     43 For anyone who read those pieces and felt the dread, whether you're building AI
     44 and worrying about what it means, or you've absorbed the pessimist consensus
     45 and started treating decline as a foregone conclusion, or you’re in the bucket
     46 of people Shumer insists are fucked; I'm going to argue that the pessimists
     47 have the best narratives and the worst track record. The doom scenarios require
     48 assumptions that don't survive contact with economic history, and the
     49 psychological posture you bring to this moment actually matters for how it
     50 turns out.
     51 
     52 Why the doom loop feels so right
     53 
     54 The central mechanism of the Citrini thesis: when you make intelligence
     55 abundant and cheap, you destroy the income that 70% of GDP depends on. A single
     56 GPU cluster in North Dakota generating the output previously attributed to
     57 10,000 white-collar workers in midtown Manhattan is, in their framing, "more
     58 economic pandemic than economic panacea." The velocity of money flatlines. The
     59 consumer economy withers. Ghost GDP accumulates in the national accounts while
     60 real humans stop being able to pay their mortgages.
     61 
     62 Noah Smith, writing on [16]Noahpinion the day after the selloff, called it "a
     63 scary bedtime story" and pointed out that Citrini doesn't use an explicit
     64 macroeconomic model, so you can't actually see what assumptions are driving the
     65 doom spiral. Smith noted that none of the analysts whose job it is to track
     66 Visa and Mastercard stock had apparently thought about AI disruption until a
     67 blogger spelled it out for them, which tells you more about sentiment-driven
     68 trading than it does about macroeconomics. The economist Gerard MacDonell
     69 described the entire piece as "allegorical" but pointed out that it ignores a
     70 basic economic principle: production generates income.
     71 
     72 Ben Thompson, on Stratechery, has been making a version of this counterargument
     73 for months, most forcefully in his January piece [17]AI and the Human Condition
     74 , where he argued that even if AI does all of the jobs, humans will still want
     75 humans, creating an economy for labor precisely because it is labor. Thompson's
     76 framing cuts to something the doom narratives consistently miss. They model AI
     77 exclusively as labor substitution: the same economy, minus humans. Every
     78 section of the Citrini piece is about replacing workers and squeezing margins
     79 on existing activity. What they don't model is what the freed-up surplus
     80 creates. As Thompson put it in [18]his analysis of the Citrini selloff, this is
     81 the real error: a refusal to believe in human choice and markets.
     82 
     83 It's an error that has been made, in nearly identical form, about every major
     84 technological transformation in modern history. Every single time, the
     85 pessimists looked at what was being destroyed and extrapolated catastrophe,
     86 while failing to imagine what would be created, because the thing that would be
     87 created hadn't been invented yet.
     88 
     89 Catastrophists keep being wrong
     90 
     91 In 1810, 81% of the American workforce was employed in agriculture. Two hundred
     92 years later, it's about 1%. If you had shown someone in 1810 a chart of
     93 agricultural employment decline and asked them to model the economic
     94 consequences, the only rational projection would have been apocalypse. Where
     95 would 80% of the population find work? What would they do? How would anyone eat
     96 if the farmers were all displaced by machines?
     97 
     98 The answer, of course, is that entirely new categories of work were created
     99 that no one in 1810 could have conceived of, and these new jobs paid
    100 dramaticaly more than subsistance farming. Factory work, office work, services,
    101 knowledge work, the entire apparatus of modernity: none of it was visible from
    102 the vantage point of the pre-industrial economy. The transition was brutal and
    103 uneven. The handloom weavers of England suffered. Dickens documented the
    104 squalor of early industrialization in prose that still makes you flinch. But
    105 the trajectory was real, and the people projecting permanent immiseration from
    106 the displacement of agricultural labor were, in the fullest sense,
    107 catastrophically wrong.
    108 
    109 Tom Lee of Fundstrat made this point with a specific example that I find
    110 clarifying. The invention of flash-frozen food in the early 1900s disrupted
    111 farming, taking agriculture from 30-40% of employment down to its current
    112 sliver. The economy didn't collapse. It reallocated value elsewhere, into
    113 industries and occupations that the frozen food pioneers couldn't have
    114 imagined. And today, I can't name a single family that subsists on frozen TV
    115 dinners.
    116 
    117 The Citrini scenario expects you to believe that AI will be the first major
    118 technological revolution in which this reallocation mechanism fails entirely.
    119 Where every previous wave of automation freed up human labor and capital to
    120 flow into new, higher-value activities, this time the loop... stops. The
    121 surplus accrues to the owners of compute, consumers lose purchasing power, and
    122 the negative feedback loop has no natural brake. It's worth sitting with how
    123 strong a claim that is. It requires every previous pattern of technological
    124 adaptation to be wrong, or at least irrelevant. And when you look at the actual
    125 data, there are signs that white-collar job postings have stabilized, layoff
    126 mentions on earnings calls remain well below early 2023 peaks, and
    127 forward-looking labor indicators show no sign of the displacement spiral that
    128 the doom thesis predicts.
    129 
    130 Does that mean AI won't disrupt specific industries and jobs? Obviously it
    131 will. Some of those disruptions will be painful and dislocating for the people
    132 caught in them. But there's an enormous gap between "this technology will cause
    133 serious labor market disruption that we need to manage" and "this technology
    134 will cause a self-reinforcing economic death spiral from which there is no
    135 recovery." Citrini is arguing the latter, while the evidence supports the
    136 former.
    137 
    138 Why vivid scenarios beat boring probabilities
    139 
    140 There's a reason the doom narratives go viral while the measured
    141 counterarguments get a polite nod // a fraction of the engagement. It has
    142 nothing to do with the quality of the underlying analysis. It has everything to
    143 do with how human brains process information.
    144 
    145 Daniel Kahneman's work on the availability heuristic showed that we judge the
    146 probability of events by how easily we can imagine them. Dystopia is easy to
    147 imagine. We have an extraordinarily rich cultural tradition of imagining
    148 technological nightmare scenarios in exquisite detail. Orwell did it
    149 brilliantly. Every season of Black Mirror does it competently. The Terminator
    150 gave us the visual grammar for AI catastrophe decades before anyone had a
    151 working language model. When Citrini describes a world where the unemployment
    152 rate hits 10.2% and the S&P crashes 38%, you can picture it. You can feel the
    153 dread. Hollywood has been training you to feel exactly this dread for your
    154 entire life.
    155 
    156 Now try to imagine the positive scenarios. Try to picture, in concrete sensory
    157 detail, a world where AI helps us solve protein folding problems across
    158 thousands of neglected tropical diseases, where it accelerates materials
    159 science research by orders of magnitude, where it makes high-quality legal and
    160 medical advice accessible to people who currently can't afford it, where it
    161 enables forms of creative expression and economic activity that we can't yet
    162 name because they don't exist yet. It's fuzzy and abstract. You can state it
    163 intellectually, but you can't feel it the way you can feel the unemployment
    164 spiral.
    165 
    166 This asymmetry isn't trivial. The [19]Ifo Institute has published research
    167 showing that investors are willing to pay more for economic narratives than for
    168 raw forecasts, and that pessimistic narratives command higher prices among
    169 certain investor types. As [20]Joachim Klement put it in his response to the
    170 Citrini selloff: investors value narratives more than actual recession
    171 forecasts. Stories travel faster than spreadsheets.
    172 
    173 Shumer's piece is a narrative construction, and a questionable piece of
    174 analysis. He opens with the COVID comparison: remember February 2020, when a
    175 few people were talking about a virus and everyone thought it was overblown? He
    176 positions himself as the insider who sees what's coming, who's been "giving the
    177 polite, cocktail-party version" but can't hold back the truth any longer. [21]
    178 Paulo Carvao, writing in Forbes, noted that it reads at times like a sales
    179 pitch. It’s a used-car pitch at that. The Guardian pointed out that Shumer
    180 "previously excited the internet by announcing the release of the world's 'top
    181 open-source model,' which it was not." (To be clear: this is a kinder way of
    182 saying [22]it was fraud.)
    183 
    184 But criticism doesn't travel like fear does. Fear is a better story. And so the
    185 doom narratives accumulate cultural mass while the boring, incremental,
    186 statistically-grounded counterarguments remain niche reading for economists and
    187 strategists.
    188 
    189 We remember disasters, not the ones we dodged
    190 
    191 Humans are spectacular at remembering disasters, passed down in every format
    192 from the written word to the oral tradition. We are (for obvious reasons)
    193 terrible at remembering the disasters that didn't happen. In 1962, during the
    194 Cuban Missile Crisis, a Soviet submarine officer named Vasili Arkhipov refused
    195 to authorize the launch of a nuclear torpedo, overriding two other officers who
    196 wanted to fire. The world didn't end. Most people today have never heard of
    197 Arkhipov. Everyone knows about Hiroshima and Nagasaki. The bomb that fell is
    198 seared into collective memory. The bomb that didn't fall is a footnote.
    199 
    200 The Y2K bug was going to crash civilization; then billions of dollars of
    201 engineering work fixed it, and everyone retroactively decided it was never a
    202 real threat. The ozone layer was going to disintegrate; then the Montreal
    203 Protocol worked better than almost anyone predicted, and ozone depletion feels
    204 like a quaint 1990s worry. Acid rain was dissolving the forests of North
    205 America; then sulfur dioxide regulations cut emissions drastically, and the
    206 whole issue evaporated from public consciousness. Every one of these was a
    207 genuine threat. Every one was met by human ingenuity and institutional
    208 coordination. Every one was subsequently memory-holed, because success is
    209 boring and failure is vivid.
    210 
    211 We're running our forecasting models on a dataset that systematically excludes
    212 our wins. It should be entirely unsurprising that the forecasts come out
    213 somewhat bearish.
    214 
    215 Ben Thompson (as usual) gets it right
    216 
    217 Thompson's core insight is that humans want humans. He points to the
    218 agricultural revolutions: in the pre-Neolithic era, zero percent of humans
    219 worked in agriculture. By 1810, 81%. By today, 1%. Machines replaced human
    220 agricultural labor entirely, and rather than the economy collapsing, entirely
    221 new categories of work were created that paid dramatically more. This cycle
    222 played out again with industrialization, with computing, with the internet.
    223 Every time, the displacement was real, and every time, new forms of
    224 human-valued work emerged that couldn't have been predicted.
    225 
    226 Citrini called DoorDash "the poster child" for AI disruption, imagining
    227 vibe-coded competitors fragmenting the market overnight. Thompson flips it:
    228 DoorDash is the poster child for why the article is absurd. DoorDash didn't
    229 always exist. It was built, and it wins through the active choice of customers,
    230 restaurants, and drivers. The doom thesis treats it as a static rent-extraction
    231 layer sitting on top of human laziness, but DoorDash created its market from
    232 scratch and generated new jobs for millions of drivers along the way. What the
    233 Citrini analysis lacks, Thompson argued, is any belief in human choice or
    234 markets. If your starting assumption is that things are as they are, you can
    235 only envision breaking them.
    236 
    237 Citrini predicted AI would collapse real estate commissions by eliminating
    238 information asymmetry. But the internet already did that. You can look up every
    239 house for sale right now, with full history and photos. Real estate agents
    240 still exist, which is one of the better arguments that humans are resourceful
    241 at giving themselves work to do even in fields where they arguably shouldn't
    242 need to.
    243 
    244 In a world of AI abundance, the things humans create will become more valuable
    245 precisely because they're human. AI art will make human art more desirable, not
    246 less, because provenance matters. AI-generated content will make
    247 human-generated content worth more, because the imperfections and
    248 idiosyncrasies are features.
    249 
    250 Is this optimistic? Yes. Could it be wrong? Sure it could. But it's grounded in
    251 a real observation about human psychology that the doom models don't account
    252 for. Citrini's Ghost GDP thesis assumes that when AI replaces human labor, the
    253 value simply evaporates from the consumer economy. Thompson's counterargument
    254 is that humans will create new forms of value that are specifically human, and
    255 that demand for those forms of value will intensify as machine-generated
    256 alternatives become ubiquitous. The history of technological disruption
    257 suggests Thompson has the stronger case.
    258 
    259 Pessimism as a self-fulfilling prophecy
    260 
    261 What actually worries me is the second-order effects of the doom narrative
    262 itself.
    263 
    264 When the smartest, most technically capable people in a field become convinced
    265 that the field is heading toward catastrophe, several things happen. Some leave
    266 the field entirely, removing exactly the talent you'd want steering the ship.
    267 Some stay but adopt a posture of resigned inevitability, which is functionally
    268 identical to apathy. Some decide that since disaster is coming, they might as
    269 well accelerate and cash out. And a vocal minority become so consumed by
    270 existential risk that they advocate for extreme countermeasures that would
    271 concentrate power in ways that create entirely new categories of danger.
    272 
    273 Robert Oppenheimer (in the wake of his famous invocation of the Bhagavad Gita)
    274 spent the years after the Manhattan Project arguing passionately for
    275 international cooperation on nuclear governance. He didn't say "we should never
    276 have done this." He said, essentially, "this is incredibly powerful, and we
    277 need to build institutions that can handle it." He was an optimist in the
    278 meaningful sense: he believed better outcomes were achievable if people worked
    279 to achieve them. He was right about that, because we're still here.
    280 
    281 The most effective people working on AI safety and governance right now are,
    282 almost without exception, optimists. They work on alignment because they
    283 believe alignment is solvable. They push for better governance becuase they
    284 believe governance can work. The ones who've concluded that the problem is
    285 unsolvable tend to stop doing useful work, for obvious reasons.
    286 
    287 Gramsci wrote about "pessimism of the intellect, optimism of the will." You
    288 look at the world clearly. You see the problems. And then you choose to act as
    289 if better outcomes are possible, because that choice is the precondition for
    290 achieving them.
    291 
    292 Nobody can see the next economy
    293 
    294 What both Shumer and Citrini miss is that they're modeling a future economy
    295 using the structure of the present economy. They see AI replacing white-collar
    296 workers within the existing economic framework and project the consequences of
    297 that replacement within that same framework. But every major technological
    298 transformation has changed the framework itself, creating entirely new economic
    299 structures that were invisible from the vantage point of the old ones.
    300 
    301 In 1995, if you told someone that one of the largest employers in America would
    302 be a company that let strangers sleep in each other's homes, they would have
    303 thought you were insane. If you told them that millions of people would make a
    304 living by talking into microphones about their opinions, or recording
    305 themselves playing video games, or writing newsletters on the internet, they'd
    306 have had you committed. The entire creator economy, the gig economy, the app
    307 economy, the SaaS economy that Citrini is now eulogizing: none of it was
    308 predictable from the vantage point of 1995. And that's a 30-year window. The
    309 agricultural revolutions played out over centuries.
    310 
    311 What will people do when AI can handle most current white-collar tasks?
    312 
    313 I don't know.
    314 
    315 And that's the whole point.
    316 
    317 Nobody knew what displaced agricultural workers would do, either, until they
    318 did it. The absence of a visible next chapter isn't evidence that there won't
    319 be one. It's evidence that we're bad at predicting what humans will invent when
    320 constraints shift.
    321 
    322 Choosing optimism with open eyes
    323 
    324 I'm not saying everything will be fine. I'm not saying the transition will be
    325 smooth. I'm not saying that the people displaced by AI won't suffer, or that we
    326 don't need better policy frameworks to handle the disruption. The
    327 distributional concerns at the heart of the Citrini piece are legitimate. If
    328 productivity gains accrue primarily to the owners of compute and capital while
    329 labor income stagnates, that's a genuine problem. Labour's share of GDP has
    330 been declining for decades. These are real numbers pointing to real challenges.
    331 
    332 What I am saying is that the leap from "this will be disruptive and we need to
    333 manage it carefully" to "this will cause an irreversible economic death spiral"
    334 isn't supported by the evidence, by economic history, or by what we know about
    335 how humans respond to technological change. The Citrini scenario requires every
    336 adaptive mechanism in the economy to fail simultaneously and completely within
    337 roughly two years. That's a very specific left-tail outcome.
    338 
    339 If you're building AI systems, if you're founding companies, if you're writing
    340 code that will shape how people experience the world, your psychological
    341 orientation toward the future is a variable that directly shapes // affects
    342 outcomes. Pessimistic builders build defensively. They hoard and hedge and make
    343 decisions based on fear. Optimistic builders build with ambition. They invest
    344 in safety because they believe safety is achievable. They take on hard problems
    345 because they believe hard problems have solutions.
    346 
    347 The tech industry is at a hinge point, and the narrative it tells itself will
    348 shape what it creates. If the dominant narrative is doom, the best people
    349 leave, the remaining people race to extract value before the collapse, and the
    350 governance frameworks get built by people who don't understand the technology.
    351 If the dominant narrative is cautious optimism, the best people stay, the work
    352 is good, and the institutions get built by people who know what they're
    353 building for.
    354 
    355 Ed Yardeni, the veteran Wall Street strategist, noted in the wake of the
    356 Citrini selloff that "the AI story has morphed from a Roaring 2020s
    357 productivity booster to an existential threat to our way of life." He found
    358 this striking. I find it absurd. The underlying technology hasn't changed, and
    359 the capabilities haven't shifted. What changed is the narrative, and narratives
    360 are always, always choices.
    361 
    362 I choose optimism. I choose it because the alternative is surrender as
    363 sophistication. And because every time I look at the historical record, the
    364 full record that includes both the disasters and the averted disasters, both
    365 the tragedies and the triumphs, the case for human ingenuity and resilience is
    366 stronger than the case against it.
    367 
    368 The doomers may have the best stories.
    369 
    370 I believe the optimists have the best evidence.
    371 
    372 I'll take the evidence.
    373 
    374 Everything is (going to be) awesome.
    375 
    376 $ cat ./comments
    377 $ cat ./subscribe.md
    378 
    379 Get updates
    380 
    381 Field Notes on Now.
    382 
    383 [23][                    ] SUBSCRIBE_
    384 [25]Home/// [26]About/// [27]RSS/// [28]Tools/// [29]YouTube/// [30]Forum/// 
    385 [31]Home/// [32]About/// [33]RSS/// [34]Tools/// [35]YouTube/// [36]Forum/// 
    386 [37]Home/// [38]About/// [39]RSS/// [40]Tools/// [41]YouTube/// [42]Forum/// 
    387 [43]Home/// [44]About/// [45]RSS/// [46]Tools/// [47]YouTube/// [48]Forum///
    388 © 2026 Westenberg. [49]Sign up
    389 Theme by [50]JA Westenberg x [51]Studio Self
    390 
    391 References:
    392 
    393 [1] https://www.joanwestenberg.com/
    394 [3] https://www.joanwestenberg.com/
    395 [4] https://www.joanwestenberg.com/about/
    396 [5] https://www.joanwestenberg.com/rss/
    397 [6] https://westenberg.gumroad.com/
    398 [7] https://www.youtube.com/@jawestenberg
    399 [8] https://westenberg.discourse.group/
    400 [9] https://www.joanwestenberg.com/everything-is-awesome-why-im-an-optimist/#
    401 [10] https://www.joanwestenberg.com/signin/
    402 [11] https://www.joanwestenberg.com/signup/
    403 [12] https://www.joanwestenberg.com/author/jawestenberg/
    404 [13] https://shumer.dev/something-big-is-happening?ref=joanwestenberg.com
    405 [14] https://www.citriniresearch.com/p/2028gic?ref=joanwestenberg.com
    406 [15] https://www.bloomberg.com/news/articles/2026-02-23/software-payments-shares-tumble-after-citrini-post-on-ai-risks?ref=joanwestenberg.com
    407 [16] https://www.noahpinion.blog/p/the-citrini-post-is-just-a-scary?ref=joanwestenberg.com
    408 [17] https://stratechery.com/2026/ai-and-the-human-condition/?ref=joanwestenberg.com
    409 [18] https://stratechery.com/2026/another-viral-ai-doomer-article-the-fundamental-error-doordashs-ai-advantages/?ref=joanwestenberg.com
    410 [19] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5637576&ref=joanwestenberg.com
    411 [20] https://klementoninvesting.substack.com/p/why-pessimists-make-more-money?ref=joanwestenberg.com
    412 [21] https://carvao.substack.com/p/the-problem-with-techs-latest-something?ref=joanwestenberg.com
    413 [22] https://x.com/jawestenberg/status/2021782902342922514?s=20&ref=joanwestenberg.com
    414 [25] https://www.joanwestenberg.com/
    415 [26] https://www.joanwestenberg.com/about/
    416 [27] https://www.joanwestenberg.com/rss/
    417 [28] https://westenberg.gumroad.com/
    418 [29] https://www.youtube.com/@jawestenberg
    419 [30] https://westenberg.discourse.group/
    420 [31] https://www.joanwestenberg.com/
    421 [32] https://www.joanwestenberg.com/about/
    422 [33] https://www.joanwestenberg.com/rss/
    423 [34] https://westenberg.gumroad.com/
    424 [35] https://www.youtube.com/@jawestenberg
    425 [36] https://westenberg.discourse.group/
    426 [37] https://www.joanwestenberg.com/
    427 [38] https://www.joanwestenberg.com/about/
    428 [39] https://www.joanwestenberg.com/rss/
    429 [40] https://westenberg.gumroad.com/
    430 [41] https://www.youtube.com/@jawestenberg
    431 [42] https://westenberg.discourse.group/
    432 [43] https://www.joanwestenberg.com/
    433 [44] https://www.joanwestenberg.com/about/
    434 [45] https://www.joanwestenberg.com/rss/
    435 [46] https://westenberg.gumroad.com/
    436 [47] https://www.youtube.com/@jawestenberg
    437 [48] https://westenberg.discourse.group/
    438 [49] https://www.joanwestenberg.com/everything-is-awesome-why-im-an-optimist/#/portal/
    439 [50] https://joanwestenberg.com/
    440 [51] https://thisisstudioself.com/