Silicon Valley Is Bracing for a Permanent Underclass

À retenir

1. Date. 30 avril 2026, NYT Opinion. Auteure : Jasmine Sun (Substack, AI et culture SF). Format audio : 28 min 34 s. URL : https://www.nytimes.com/2026/04/30/opinion/ai-labor-work-force-silicon-valley.html
2. Thèse. "the production of a social underclass is a policy choice. Instead of waiting for impact, we need to think seriously — now — about how we plan to support workers through A.I. disruption."
3. Phrase d'ouverture (aphorisme). "Most people I know in the A.I. industry think the median person is screwed, and they have no idea what to do about it."
4. San Francisco consensus. "Whether you talk with engineers, venture capitalists, founders or managers, or with doomers, accelerationists, lefties or libertarians... the so-called San Francisco consensus on the impact of A.I. for workers is bleak."
5. Permanent underclass — définition. terme repopularisé sur les réseaux pour décrire la théorie selon laquelle "people have a limited window of time to build wealth before A.I. and robotics are advanced enough to fully replace human labor. At that point, we will get frozen in our current class positions." Origine : années 1960 pour les ouvriers laissés pour compte par l'automatisation post-guerre.
6. Yash Kadadi (23 ans, Stanford dropout). "There's only a matter of time before GPT-7 comes out and eats all software and you can no longer build a software company. Or the best version of Tesla Optimus comes out... this year is a human's last chance to be a part of the innovation."
7. Dario Amodei.
8. "White-collar blood bath" déclaration publique.
9. 50% des emplois junior white-collar pourraient disparaître d'ici 2030. .
10. Essai blog ~20 000 mots, janvier 2026 : warns A.I. may create "an unemployed or very-low-wage 'underclass' for people with 'lower intellectual ability'".
11. Axios : "The balance of power of democracy is premised on the average person having leverage through creating economic value. If that's not present, I think things become kind of scary."
12. Sur les employés Anthropic : "It may be feasible to pay human employees even long after they are no longer providing economic value in the traditional sense. Anthropic is currently considering a range of possible pathways for our own employees."
13. Sam Altman.
14. 2021 blog post : "unstoppable" A.I. systems will do almost any job a human could → power shifts from labor to capital. "If public policy doesn't adapt accordingly, most people will end up worse off than they are today."*
15. Solution proposée : taxer agressivement assets (terres + actions A.I.).
16. Chris Lehane. (lobbyiste vétéran, OpenAI depuis avril 2024) : a "deprioritized" des projets de recherche aux résultats peu flatteurs (impacts environnementaux, gender gap, urban-rural divide ChatGPT, ChatGPT et décisions de carrière, prévisions économiques long terme). Citation employé OpenAI anonyme : "Whenever someone wrote a paper which talked about some negative aspect of A.I., he would say, 'We're not going to release something about a problem until we have a solution for it.'" Lehane lui-même : "We want to do applied physics, not theoretical physics."
17. OpenAI white paper avril 2026 — *"Industrial Policy for the Intelligence Age". * :
18. Justification Lehane : industrialisation "really threw off that relationship between capital and labor" et a facilité "the rise of fascism and communism".
19. Politiques radicales-progressistes proposées.
20. 32-hour workweek.
21. Hausse taxes corporations + capital gains
22. "Public wealth fund". equity stake citoyens dans les A.I. companies
23. Accélération expansion grid énergétique
24. National "right to A.I." : foundation models pour schools et libraries
25. Limite : vague sur l'implémentation, pas d'engagement concret. Spokesperson refuse de citer une législation spécifique.
26. OpenAI 2025. suppression du profit cap qui limitait les retours investisseurs/employés à 100x leur investissement initial.
27. Leading the Future PAC. (pro-IA, financé en partie par Greg Brockman, président OpenAI) : a dépensé plus de 2M$ en pubs contre Alex Bores, candidat congressional NY ayant introduit un projet de safety regulation pour grands développeurs IA + un plan de paiements directs aux Américains financés par la taxe sur l'IA.
28. Steven Adler. (ex-équipe safety OpenAI) : "I hope OpenAI is willing to fight for these prosocial ideas with policymakers. The A.I. industry is engaged in cutthroat competition over truly world-changing technology. Unless we change their incentives, we shouldn't be surprised when companies cut corners, even if they've said the right things."
29. Anthropic.
30. Anthropic Institute lancé mars 2026. , dirigé par Jack Clark (co-fondateur, ex-journaliste devenu billionaire IA). Réunit teams economics, societal impact, frontier safety.
31. ARR Anthropic. 30B$ annualisé (avril 2026) vs 9B$ fin 2025. Croissance massive enterprise agents.
32. Contribution politique. 20M$ à un groupe soutenant Alex Bores (counter-balance Leading the Future).
33. Pas (encore) de white paper économique ni d'endorsement législatif spécifique. Clark sur le lobbying : "the end of a very, very long chain of work."
34. Publication Anthropic janvier 2026 : "a small but increasing fraction of Claude users are delegating their most personal and consequential decisions to A.I. — a choice they often later regret. 'You made me do stupid things,' one such user told Claude."
35. Étude junior engineers. "junior engineers who relied on A.I. coding agents not only didn't complete tasks much faster; they also understood their work less when quizzed about it afterward." Implication : déskilling early-career.
36. Framing interne : "light and shade" de l'IA.
37. Employés ont pré-engagé plusieurs milliards de donations individuelles à des nonprofits (incluant prévention catastrophic AI outcomes).
38. Vision Clark sur l'avenir : "expand the share of human labor" en relational roles (teaching, nursing) — "What A.I. should allow us to do is pay these jobs way more and massively multiply the number of them."
39. Tejal Patwardhan. (lead frontier evaluations OpenAI) — sur GDPVal :
40. Benchmark sur 44 occupations (real estate broker → news analyst).
41. "When we originally released GDPVal, which was just a few months ago, none of the models were yet on par with human experts. Months later, we have over an 80 percent win rate compared to human professionals."
42. Anecdote : research colleague ex-banquière "keeps being shocked by how much of her old work the models can do."
43. A.I. Productivity Index. 4 jobs benchmarkés — investment banking associate, management consultant, Big Law associate, primary care physician.
44. Block / Jack Dorsey.
45. Mars 2026 : laid off nearly half of company employees.
46. Wired : coding agents "such as Anthropic's Opus 4.6 and OpenAI's Codex 5.3" "presented an option to dramatically change how any company is structured, and certainly ours."
47. Marché : +25% stock price surge en after-hours.
48. Mechanize. startup "once buzzy" avec mission "to enable the full automation of the economy". Blog post fondateurs : "the only real choice is whether to hasten this technological revolution ourselves, or to wait for others to initiate it in our absence."
49. Zoë Hitzig. (économiste, ex-OpenAI) : effet d'imitation. "When chief executives are 'saying they're cutting jobs because of A.I., other people feel like they have to too. That dynamic could make the changes happen sooner than efficiency would dictate."
50. St. Louis Fed. AI investments (logiciel + data centers) = 39% de la croissance US sur Q1-Q3 2025. Le gouvernement fédéral a donc un intérêt vital à soutenir le boom AI. Amodei : "the reluctance of tech companies to criticize the U.S. government, and the government's support for extreme anti-regulatory policies on A.I."
51. Stratégie politique David Shor. (event 25 février 2026, Dogpatch SF, sous une disco ball avec La Croix cans) :
52. Slide : 79% des voters worry "government not having a plan to protect workers".
53. 72%. concerned A.I. "drives down wages for people like you".
54. "Right now, the argument is, 'You're all about to lose your jobs, and the choice is either you get nothing and starve, or we do something fair.' People don't want to be members of the permanent underclass."
55. UBI impopulaire. , mais federal jobs guarantee a des jambes.
56. Voters s'en fichent de battre la Chine, mais excités par AI curing diseases.
57. Top ad testée : "We make the corporations and billionaires who profit from A.I. pay their fair share. They work for the bots. We work for you."
58. Pitch : "$700 billion a year is being spent" on A.I. transformation. Pour moins que ce que l'industrie dépense en une heure, donateurs peuvent armer les Démocrates en messaging.
59. Sur "crying wolf" : "People's bar is way too high on this. The reality is, if one concentrated industry with 1,000 people loses their jobs, it's going to be the biggest story of the century."
60. Données labor market.
61. Chômage en hausse pour young workers en occupations highly A.I.-exposed : software engineering, customer service.
62. Hiring slow attendu : accounting, marketing, design, administrative work.
63. Jobs à comparative advantage humain : entrepreneurship, care work, skilled trades, entertainment (sports, performing arts).
64. Économistes cités.
65. Anton Korinek. (UVA + Anthropic Institute) : "no human job is invulnerable in the long run, once A.I. can outperform humans at everything."
66. David Autor. (MIT) : new industries will emerge — "just as our ancestors could not have fathomed the modern roles of flight attendants and software salespeople."
67. Carl Benedikt Frey. (Oxford) : "Most economists will acknowledge that technological progress can cause some adjustment problems in the short run. What is rarely noted is that the short run can be a lifetime."
68. Bharat Ramamurti. (ex-deputy director Biden NEC) : "The China shock unfolded over several years, whereas this could happen over two years. These companies have spent so much money developing models that there's going to be immense pressure on them to generate revenue through quick adoption."
69. Molly Kinder. (Brookings senior fellow) : "I've interviewed so many college students who are super fearful about what the future means, and their narrative is exactly the same as those blue-collar guys in the heartland." Sur les narratifs IA : "Our economy grew extraordinarily and prices went down, but there were clear losers."
70. Class solidarity blue/white-collar. "For once, a rarefied class of employees — those used to being the automaters, not the automated — is reckoning with their potential obsolescence."
71. Politique 2028.
72. AI a monté plus vite que tout autre issue auprès des voters dans la dernière année (data Shor).
73. Senator Mark Kelly. + Rep. Ro Khanna ont annoncé sweeping AI agendas.
74. Démocrates exposés (voters jeunes + diplômés college plus exposés à l'IA que Républicains).
75. Sun : "The technology is an opportunity for gutsy politicians — especially populist candidates vying in a crowded 2028 presidential primary — to push ideas that are usually too radical for moderate voters to swallow."
76. Violence populiste émergente.
77. Avril 2026. tentative de firebombing du domicile de Sam Altman.
78. Autre attaquant accusé d'avoir ciblé un councilman d'Indianapolis ayant approuvé un projet local de data center.
79. Bans proposés : data center construction, self-driving cars, chatbots usage thérapie/droit.
80. Alex Karp (Palantir CEO). lors d'un panel mars 2026 avec Sean O'Brien (Teamsters) : "The biggest challenge to A.I. in this country is political unrest. If I were sitting here in private with my peers, I'd be telling them the country could blow up politically and none of us are going to make any money when the country blows up."
81. Articulation dossier veille.
82. Article-cadre politico-économique qui complète et nationalise les fiches brain fry (Bedard/BCG/HBR 2026-03-05), Silicon Valley boîte de Petri agents (Dèbes/Les Echos 2026-04-22), Sierra AI-native interview (Taylor 2026-04-20).
83. Corrobore Mollick Théorie organisationnelle pour l'IA agentique (2026-02), MIT NANDA 95% AI Pilots Fail (2025-08-23), MIT Iceberg Index (2025-11-26).
84. Triangulation des chiffres.
85. Anthropic ARR 30B$ (Sun) ↔ Stripe Minions 1300+ PRs/semaine (2026-02-19) — confirme le boom enterprise.
86. 50% jobs junior white-collar disparus d'ici 2030 (Amodei via Sun) ↔ thèses Andreessen (2026-02), Karpathy (2026-04-29), Rohit (2026-04-29) sur la transition prompt engineer → systems architect.
87. Codifie politiquement. ce que les fiches techniques décrivent au niveau workflow (Karpathy "never felt more behind", Block -50% Dorsey, GDPVal 80% win rate).
88. Pose la séquence narrative : labs reconnaissent le risque (Amodei) → corporate accélère layoffs (Block) → métriques l'orientent (GDPVal) → réponse politique embryonnaire (white paper OpenAI, Anthropic Institute) → mais lobbying contraire (Leading the Future PAC) → réponse populiste (Shor strategy, candidats 2028, violence émergente).
89. À mobiliser pour. présentations cadrant le contexte societal-policy de l'agentic engineering ; risque réputationnel des déploiements IA en France/Europe ; stratégie hiring défensif (Sierra-style refonte) face à l'obsolescence early-career.