DeepSeek

The AGI Chronicles

by Kevin Roose  · 481pp  · 151,124 words

small, relatively unknown research unit in Hangzhou that was a subsidiary of an algorithmic trading firm called High-Flyer. Some AI insiders had been tracking DeepSeek’s models for months, but its sudden arrival at the frontier blindsided most of the Western AI establishment. For years, China’s AI advances

…

—Baidu, Tencent, Alibaba—and their models tended to be unremarkable, lagging a year or two behind what the leading American companies were putting out. DeepSeek was different. It was started by Liang Wenfeng, a forty-year-old trader who had gotten bitten by the AGI bug. Unlike American AI companies

…

, which tethered themselves to hyperscalers like Microsoft, Google, and Amazon, Liang had funded DeepSeek’s research himself, hiring a team of young researchers and giving them a stash of GPUs from his trading firm to run their experiments. Liang

…

sounded like a Chinese version of Ilya Sutskever. “Our destiny is AGI,” he told a Chinese-language interviewer in 2024. In the mission statement on DeepSeek’s website, there was no mention of safety, competition, or the stakes for humanity—only “unraveling the mystery of AGI with curiosity.” Liang was

…

also seemingly uninterested in turning a profit; rather, he framed DeepSeek’s emergence as a symbolic revival of Chinese ambition. “We believe that as the economy develops, China should gradually become a contributor instead of freeriding

…

culture in China, noting, “In the past 30 years, we’ve emphasized only making money while neglecting innovation.” The second mystery was the cost. DeepSeek claimed that V3, the base model underlying R1, had been trained for $5.6 million on a cluster of 2,048 Nvidia GPUs. The reinforcement

…

a large, high-performance model to train a smaller, more efficient one. Some American AI leaders, like Alexandr Wang of Scale AI, speculated that DeepSeek had access to more powerful Nvidia GPUs than it claimed, and had gotten them through smuggling, in violation of U.S. export controls. But the

…

real, unsatisfying answer appeared to be that DeepSeek’s developers were simply “cracked”—the AI industry’s term of highest praise for elite technical talent. They’d read the same papers as the

…

Satya Nadella, Microsoft’s CEO, said in an interview, adding, “We should take the developments out of China very, very seriously.” The most obvious question DeepSeek raised was whether the Western AI industry’s entire theory of progress had been wrong. For years, the industry had operated under the assumption that

…

time, you could watch an AI model think, see it consider and discard approaches, and catch it correcting itself midstream. When a user asked DeepSeek-R1 for its opinion of a version of the “trolley problem”—the classic philosophy thought experiment that asks whether you’d divert a runaway trolley

…

seemed more promising than other methods they’d tried. Not everyone was sold. Users noticed that on sensitive topics like the Tiananmen Square massacre, DeepSeek’s models often refused to answer or parroted Chinese Communist Party talking points. And many users worried that the app was funneling their data to

…

Other AI-related stocks cratered, too. Broadcom dropped 17 percent. Dell and Oracle each fell at least 8.7 percent. By the close of trading, DeepSeek had wiped out close to a trillion dollars of value in the American technology sector. At the frontier AI labs, leaders rushed to reassure nervous

…

investors and employees that DeepSeek’s breakthroughs weren’t as meaningful as they looked. Demis Hassabis called R1 “an impressive piece of work,” but suggested that the $5.6

…

this time is that the company that was first to demonstrate the expected cost reductions was Chinese.” The American AI company most directly threatened by DeepSeek was Meta. For years, Mark Zuckerberg had been spending billions of dollars developing Llama, a series of open-source LLMs, and betting that releasing

…

releases. While the markets reeled and the labs scrambled, Washington was taking notice. President Trump, speaking at a House Republican retreat in Florida, called DeepSeek “a wake-up call for our industries.” His press secretary blamed the Biden administration for allowing China to develop the technology. David Sacks, the newly

…

appointed AI czar, echoed the critics who accused DeepSeek of stealing from American models through distillation. At the U.S. AI Safety Institute, a small team led by Paul Christiano—Dario Amodei’s

…

illustrated how tangled the AI world had become, had become the U.S. government’s top AI safety expert—spent a frenetic weekend reverse-engineering DeepSeek’s models, trying to figure out whether the company’s claims of efficiency gains were real, and assessing the models for evidence of security

…

issues and ideological bias. Within days, DeepSeek had managed to do what years of homegrown innovation had not: focus the entire American government on the dynamics of the AGI race. Lawmakers and

…

, or a reason to accelerate. Europeans saw it as an opening—if China could train frontier models on a budget, maybe they could, too. “DeepSeek has shown that all countries can be part of AI, which wasn’t obvious before,” said Clément Delangue, the French-born chief executive of Hugging

…

plans for the build-out. The stock market recovered within weeks, and American consumers mostly went back to using ChatGPT, Claude, and Gemini. But DeepSeek had broadened the race and hardened its terms. The nuanced, careful debate about how to build powerful AI safely—the debate that people like Altman

…

, Noam Shazeer, Dario Amodei, Daniela Amodei, Jared Kaplan, Sam McCandlish, Chris Olah, Tom Brown, Jack Clark, Amanda Askell, Jensen Huang, the cracked engineers at DeepSeek, the authors of “Attention Is All You Need,” and hundreds of others whose names may be forgotten by history but who were no less instrumental

…

status/1882786965608894629. “AI’s Sputnik moment”: Marc Andreessen (@pmarca), X, January 26, 2025, https://x.com/pmarca/status/1883640142591853011. On January 27, the Monday after DeepSeek’s app: Samantha Subin, “Nvidia Sheds Almost $600 Billion in Market Cap, Biggest One-Day Loss in U.S. History,” CNBC, January 27, 2025.

…

, “Nvidia Passes Apple Again to Become World’s Most Valuable Company,” CNBC, January 21, 2025. “an impressive piece of work”: Quoted in Arjun Kharpal, “DeepSeek’s AI Model Is ‘the Best Work’ out of China but the Hype Is ‘Exaggerated,’ Google DeepMind CEO Says,” CNBC, February 9, 2025, https://www

…

.cnbc.com/2025/02/09/deepseeks-ai-model-the-best-work-out-of-china-google-deepmind-ceo.html. “an impressive model, particularly around”: Sam Altman (@sama), X, January 27, 2025,

…

the Big A.I. Summit in Paris,” New York Times, February 10, 2025, https://www.nytimes.com/2025/02/10/technology/ai-summit-paris.html. “DeepSeek has shown that all countries”: Roose, “5 Notes from the Big A.I. Summit in Paris.” the ascent of accelerationists: Kyle Wiggers, “Sriram Krishnan

…

AI researcher livestream; board coup and firing; capped-profit structure creation; ChatGPT release; ChatGPT success reaction; ChatGPT superapp vision; chip venture fundraising; compute spending spree; DeepSeek R1 assessment; early career and networking; “The Gentle Singularity”; Google rivalry motives; GPT-3 API release decision; GPT-3 hype response; GPT-5 promotion;

…

of Compute”; Brockman conflicts; career pivot to AI; celebrity status; clash with Huang; Claude commercialization decision; compute infrastructure negotiations; countries plan objections; Dario Vision Quests; DeepSeek R1 assessment; Delano house and “Concrete Problems in AI Safety”; early life and education; effective altruism connections; export controls advocacy; export controls memo; FTX funding

…

arms race artificial general intelligence: alignment problem; AlphaFold; AlphaGo; AlphaZero; “Big Blob of Compute” hypothesis; Claude; Claude Code; commercialization of AI research; Dario Vision Quest; DeepSeek-R1; effective accelerationism; effective altruism; emergent abilities; existential risk; Gemini; GPT-2; GPT-4; GPT-5; Great AI Build-Out; jagged intelligence; next-token prediction

…

victory; 2020 AlphaFold2 victory censorship Center for AI Safety Center for Security and Emerging Technology CEO summit, see White House: CEO meeting chain of thought: DeepSeek-R1 traces; Q* scratchpad; reasoning models; reinforcement learning application Chalmers, David: “Could a Large Language Model Be Conscious?”; hard problem of consciousness Chaminade Character

…

GPT-4o integration; growth slowdown; internal OpenAI debate over; military contract backlash; user growth Chen, Mark Cherny, Boris; see also Claude: Claude Code release China: DeepSeek breakthrough; export controls; OpenAI countries plan China hawks Chinchilla compute-optimal discovery Chinese Communist Party Christiano, Paul: “Aligning the API”; “Concrete Problems in AI Safety

…

deep neural network: pattern discovery; reinforcement learning combination (see also reinforcement learning: DeepMind’s AlphaGo); YouTube recommendation system application DeepSeek: DeepSeek-R1; DeepSeek-V3; distillation accusations; export controls debate; market impact of app release DeepSeek-R1 release: distillation accusations; export control circumvention; industry reaction; mixture-of-experts architecture; origins and capabilities; stock market crash

…

Hellmann, Sue Detroit Devlin, Jacob dialogue agent; see also Meena dictionary learning digital life; see also Page, Larry digital neurons, see features disinformation Disney distillation: DeepSeek-R1 controversy; Gemini efficiency dogfooding doomers: Amodei’s rejection of label; economic disruption predictions; existential risk focus; political influence (see also effective accelerationism: Trump administration

…

tribe’s focus on exodus from OpenAI, see Anthropic split expert systems exponential progress; see also takeoff export controls: Buchanan’s role in chip restrictions; DeepSeek-R1 evasion; GPU restrictions on China; political debates; supply chain risk designation extinction risk, see existential risk Extropians Facebook Facebook AI Research (FAIR) fair

…

GPT-4o GPT-5; GPT-5.3-Codex release; see also generative pretrained transformer: GPT-5 GPU, see graphics processing unit GPU poor: compute inequality; DeepSeek-R1 efficiency GPU rich: infrastructure build-out; tech giant hoarding Gradient Canopy gradient clipping Graham, Logan; Claude 3.7 red-teaming Graham, Paul grand challenge

…

Grand Palais graphics processing unit: cloud platform rentals (see also Amazon: Anthropic partnership); DeepSeek-R1 efficiency; export controls; infrastructure build-out; neural network training (see also deep learning: AlexNet image recognition); Nvidia dominance; OpenAI internal capacity limits Great AI

…

and; AlphaFold project (see also DeepMind: AlphaFold); AlphaGo development (see also DeepMind: AlphaGo); AlphaZero development (see also DeepMind: AlphaZero); DeepMind cofounding; DeepMind spin-off attempt; DeepSeek R1 assessment; early life and education; Evil Genius; Google AI ambitions expansion; Google Brain rivalry; Google DeepMind CEO appointment; Google management pressures; Google personnel management

…

alignment problem: Anthropic’s “helpful, honest, harmless” framework; Askell, Amanda: “helpful, honest, harmless” framework Hendrycks, Dan Her (Jonze) Hessian-free optimizers High-Flyer; see also DeepSeek Hill, Kashmir hill-climbing h-index Hinton, Geoffrey: AlexNet development (see also deep learning: AlexNet image recognition); deep learning pioneering; loss-of-control scenario warnings

…

; Sydney encounter misinformation Mission Bay Mission Terrace Mississippi River misuse risks MIT Mitchell, Margaret; see also Stochastic Parrots controversy MIT Technology Review mixture-of-experts: DeepSeek-R1 cost savings; Gemini efficiency; Shazeer’s architecture ML, see machine learning Model Evaluation & Threat Research; see also red-teaming model welfare Modi, Narendra

…

chatbot demo; SpaceX job offer to Amodei; Stargate project opposition; Trump administration influence; xAI Gigafactory construction Nadella, Satya: Altman hiring offer; Bing chat mode launch; DeepSeek R1 reaction; GPT-2 demonstration; OpenAI $1 billion investment; OpenAI board crisis intervention; White House CEO meeting; see also Microsoft Narayanan, Arvind Narayanan, Srinivas narrow

…

shift to corporate era; Sutskever role in; WilmerHale investigation open letter on AI development pause; see also AI safety movement Open Philanthropy open-source AI: DeepSeek-R1 release; democratization vs. risks; Meta’s defense against regulation; Meta’s Llama models Open University of Israel open weights, see open-source AI Oppenheimer

…

influence rationalism movement: alignment problem focus; existential risk focus; LessWrong forum; Yudkowsky leadership RBMs, see restricted Boltzmann machines reasoning models: Apollo Research o1 evaluation; DeepSeek R1 (see also DeepSeek: DeepSeek-R1); Google DeepMind thinking strike; IMO competition; Narayanan and Kapoor critique; o1 release; o3 integration in GPT-5; Q* breakthrough; Strawberry development reasoning

…

San Jose Santa Cruz SAP Saunders, William “Say AGI” memo Scale AI scaling laws: Amodei’s realization of implications; Chinchilla correction; compute and data ratios; DeepSeek-R1 disruption; Gemini models; Google Brain adoption; GPT-3 scaling implications; industry fixation on; inference scaling axis; Kaplan and McCandlish equations; scaling laws (cont.)

…

University of Toronto University of Washington uplift trials Upwork U.S. AI Safety Institute U.S.-China AI rivalry: Amodei’s fears of Chinese advancement; DeepSeek R1 shock; GPU export controls; Trump administration China hawks U.S. Congress U.S. presidential election: Altman lobbying efforts; Altman’s political views; Bannon

…

superintelligence: existential risks); rationalism movement founding; views on AI safety; see also existential risk Zaremba, Wojciech Zhang, Hugh Zhao, Shengjia Zoph, Barret Zuckerberg, Mark: DeepSeek response war rooms; GPU stockpiling; Llama open-source model release; OpenAI researcher poaching Also by Kevin Roose Futureproof: 9 Rules for Surviving in the Age

The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence

by Sebastian Mallaby;  · 30 Mar 2026  · 607pp  · 161,998 words

on a fresh dimension. A new contender appeared out of left field: a Chinese AI lab called DeepSeek. An offshoot from a Chinese hedge fund, DeepSeek achieved instant celebrity with a reasoning system called R1. DeepSeek became a sensation for three overlapping reasons. The first was geopolitics. Chinese labs had been working on

…

in 2022, had been intended to forestall precisely this sort of catch-up; the embargo’s evident failure compounded the US sense of vulnerability. Whether DeepSeek had triumphed by buying smuggled semiconductors, or by accumulating powerful ones that the Biden team had initially not banned, was almost beside the point. One

…

actually worries me.”[12] Chinese strategists reveled in this discovery as much as American ones fretted about it. On the day of R1’s release, DeepSeek’s boss met China’s premier, Li Qiang, the country’s second-in-command, to discuss how Chinese labs could overtake US ones. The timing

…

of this conversation amplified its power: That same day, President Trump was inaugurated. Even without DeepSeek, the Trump administration would have been less inclined to restrain AI development than the Biden team had been. Now, with China’s AI momentum dramatically

…

Three Mile Island or Chernobyl—the prospects for slowing down the AI race had shrunk to roughly zero. The second reason for DeepSeek’s impact involved its supposed efficiency. DeepSeek said that it had built its models on a shoestring budget—its V3 system, which preceded R1, had allegedly cost less than

…

were falling steadily across all labs: Gemini’s jump from its expensive Ultra model to its nimble Flash models represented a huge gain in efficiency. DeepSeek’s engineers had added their own innovations, to be sure. But their achievement was a confirmation of the global trend toward better engineering and superior

…

cost-adjusted performance. It was not a disruption of it.[15] Yet whatever the truth of DeepSeek’s claims, many Western observers initially took them at face value. Commentators imagined that, since the US embargo had deprived the Chinese labs of chips

…

, DeepSeek had indeed been driven to squeeze miraculous performance out of limited hardware. This misapprehension triggered a sell-off in semiconductor stocks: If Chinese labs could

…

soon cut back on their orders from AI chipmakers. On January 27, Nvidia, the industry leader, saw its stock drop by a shocking 17 percent: DeepSeek was hailed as the disruptor of the disruptors. It took about a week for investors to absorb the real story, and for Nvidia’s stock

…

.[16] Even if R1 had not been miraculously cheap to build, it was certainly cheap to use the model. Following Meta’s open-weight example, DeepSeek allowed anyone to download the program and adapt it freely. Going beyond Meta, and indeed beyond the other Chinese labs

…

, DeepSeek allowed customers to use the full model, hosted at DeepSeek’s expense, in exchange for a paltry payment. This price-slashing persuaded many Western customers to embrace DeepSeek and ditch US providers: In late January and early February

…

, DeepSeek’s chatbot topped the free-apps download chart in Apple’s US app store.[17] China was now not only a contender in the

…

race to build frontier AI first. It was a plausible winner in the race to supply it globally. The third reason DeepSeek commanded public attention was perhaps the most interesting. Unlike frontier labs in the United States, which increasingly restricted what they published so as to protect

…

their competitive edge, DeepSeek was transparent. It posted a paper in English on the research site arXiv, explaining how R1 worked and highlighting a variant called R1-Zero. Scientists

…

, which had mastered the game of Go purely by playing against itself, R1-Zero grew strong by learning directly from trial and error—from experience. DeepSeek’s Zero system was not quite ready for prime time—it switched confusingly between languages, for example. But, guided by nothing but a reward signal

…

one reasoning path, backtrack if it hit a dead end, then take a fresh run at the problem. On mathematical benchmarks, and under certain conditions, DeepSeek’s Zero system actually beat o1. One time during training, R1-Zero interrupted itself midway through a reasoning problem. The stream of mathematical notation appearing

…

merely capable of thinking. It was capable of thinking about its thinking. For all intents and purposes, it was self-aware.[19] * * * • • • For Hassabis, the DeepSeek shock crystallized his wildly paradoxical experience of the AI race. On the one hand, he was succeeding in his mission. By the spring of 2025

…

AI development was spinning out of control. The bidding war for scientific talent and the scramble to build new data centers were increasingly wild. Following DeepSeek, a slew of Chinese labs released powerful models, mocking Hassabis’s long-ago hopes of a “singleton” scenario. It was a hard mixture to process

…

reach of Western regulatory restraint, even if one imagined that regulators had the gumption to restrain anything. In sum, and contrary to Hassabis’s hopes, DeepSeek and its followers signaled that the world would sprint over the threshold to AGI with no coordination whatsoever. Hassabis was honest about this predicament. A

…

month after the DeepSeek shock, on a panel at the World Economic Forum in Davos, he debated the dangers with Yoshua Bengio, acknowledging that RL agents of the sort

…

. BACK TO NOTE REFERENCE 12 Dave Lawler, “White House ‘Looking into’ National Security Implications of DeepSeek’s AI,” Axios, January 28, 2025, axios.com/2025/01/28/deepseek-ai-national-security-trump. BACK TO NOTE REFERENCE 13 DeepSeek claimed that it used a cluster of more than two thousand Nvidia chips to train

…

Huang, “Silicon Valley Is Raving About a Made-in-China AI Model,” The Wall Street Journal, January 27, 2025, wsj.com/tech/ai/china-ai-deepseek-chatbot-6ac4ad33. BACK TO NOTE REFERENCE 14 Between ImageNet in 2012 and 2024, frontier AI models were estimated to have become roughly ten times more

…

efficient every two years. See Leopold Aschenbrenner, “Situational Awareness,” June 2024, situational-awareness.ai. BACK TO NOTE REFERENCE 15 Following the DeepSeek shock, Nvidia’s stock experienced further turbulence relating to the Trump administration’s tariff offensive. BACK TO NOTE REFERENCE 16 Lawler, “White House ‘Looking into

…

’ National Security Implications of DeepSeek’s AI.” BACK TO NOTE REFERENCE 17 Daya Guo et al., “DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning,” arXiv, January 22, 2025, arxiv.org/abs/2501.12948. BACK

…

Veo 2 and, 367, 369 walk-away plan of, 234–36 women at, 89, 410n39 Zuckerberg and, 132–34 Deep-Q Network (DQN), 108–9 DeepSeek, 369–73, 439n14 Defense Production Act, 331 dense neural network, 347 D. E. Shaw Research, 263–64 Dewey, John, 227 digital brain, 56 direct folding

…

, 185 on deduction and induction, 28–29 DeepMind equity shares of, 83, 414n20–21 DeepMind split proposal and, 240, 246–48 on DeepMind valuation, 138 DeepSeek competition for, 373 Elixir, 405n2 AI ambitions for, 35 closing of, 42 cofounding, 32–34 DeepMind benefitting from, 109 at Electronic Entertainment Expo, 38–40

…

Putin, Vladimir, 246 Putnam, Hilary, 26, 404n10 Q Q-learning, 101–2, 106 Q* project, 356, 359 quantum mechanics, 389–94 R R1, DeepSeek, 369–72 R1-Zero, DeepSeek, 371–72 Radford, Alec, 208–10 radiology, 188–89 Rae, Jack, 346, 348–50, 368, 430n11, 437n10 Gopher frustrations of, 290, 293 OpenAI

The Reverse Centaur's Guide to Life After AI: How to Think About Artificial Intelligence—Before It's Too Late

by Cory Doctorow  · 22 Jun 2026  · 175pp  · 51,295 words

devaluation ever experienced by any corporation in human history. But Meta quickly lost its all-time-loser title. In 2025, there was the release of DeepSeek, a chatbot backed by the Chinese hedge fund High-Flyer. The chatbot, which runs on low-end chips, attained performance benchmarks heretofore associated with AI

…

a Walgreens. Even more so are the commercial operators who took these open-source models and ran with them. No model exemplifies this more than DeepSeek, the Chinese model that was created as a side hustle by High-Flyer, a hedge fund. Constrained by a U.S. blockade that limits Chinese

…

to the best cloud-based foundation model with open-source code that could be run on commodity hardware. The release of DeepSeek in 2025 sent shock waves through AI investors. DeepSeek laid bare the incredible laziness of the giant U.S. AI companies, who solved all their scaling issues by throwing money

…

are very expensive are difficult to replicate and thus the more an AI model costs to produce, the fewer competitors it is likely to have. DeepSeek’s debut sent a cold chill up the spine of every investor in a big U.S. AI company. If their $100 billion models can

…

’s pretty cool to see it fly, but that thing isn’t going to be able to compete with an Airbus. The sell-off that DeepSeek prompted was spectacular. Nvidia’s valuation alone fell by two-thirds of a trillion dollars, in one day. Remember: a growth stock has a very

The Age of Extraction: How Tech Platforms Conquered the Economy and Threaten Our Future Prosperity

by Tim Wu  · 4 Nov 2025  · 246pp  · 65,143 words

high-performance graphic processing units (GPUs) and that GPT-4 required over 10,000 petaflop/s-days.[14] (That said, the debut of China’s DeepSeek in 2025 suggested LLMs might be built with fewer computing resources.) And of course, at the frontier, you needed talent. Geoffrey Hinton won the Nobel

…

emotional intimacy or companionship. But emotions are irresistibly powerful drivers of business models, and it seems reasonable to imagine that a firm like Google or DeepSeek may find itself tempted to make subtle but deliberate efforts to cultivate bonds between us and our robot helpers. * * * — We have largely dwelled on the

…

, 159–71, 173–74 Chamberlin, Edward, 137 Charles River Bridge, Boston, 168–69 chatbots, 20, 88, 90, 101 Anthropic, 99 ChatGPT, 94–95, 98–99 DeepSeek, 94–95, 101 ELIZA, 90 Grok, 79, 99 LLaMA, 99 Replika, 100–101 Chavez, Hugo, 124, 155 Chicago School of Antitrust, 27, 30 China, 13

…

, 68, 85–88 The Death and Life of Great American Cities (Jacobs), 19 Deep Blue computer, 92 deep learning, AI, 93–94 DeepMind firm, 98 DeepSeek chatbot, 94–95, 101 de Garton, Ms., 162–63 Denmark, land ownership in, 127–29 dependence, 72, 140 convenience vs., 80 emotion-based, 100–102

The Economic Consequences of Mr Trump: What the Trade War Means for the World

by Philip Coggan  · 1 Jul 2025  · 96pp  · 36,083 words

June 2024, ourworldindata.org/data-insights/solar-panel-prices-have-fallen-by-around-20-every-time-global-capacity-doubled 53 David Pierson and Berry Wang, ‘DeepSeek is a win for China in the A.I. race. Will the party stifle it?’, The New York Times, 2 February 2025, nytimes.com/2025

…

/02/02/world/asia/deepseek-china-ai-censorship.html 54 Rogier Creemers and Louise Marie Hurel, ‘Limits of economic deterrence in the US–China tech competition’, RUSI, 14 March 2025

After Nations: The Making and Unmaking of a World Order

by Rana Dasgupta  · 27 Apr 2026  · 686pp  · 188,424 words

what it would mean for the global system if Silicon Valley firms fell behind their Chinese rivals – as was demonstrated on 20 January 2025, when DeepSeek, a Chinese start-up, launched what seemed to be a more efficient competitor to the ‘large language models’ run by American AI firms. US stock

…

D Daoism, 325, 343 Dark Enlightenment, 77*, 77 Darwinian concepts, 196–7 David (biblical character), 15 Dawes Plan, 243 De Gaulle, Charles, 268 Deepmind, 168* DeepSeek (Chinese start-up), 168 Defoe, Daniel, Robinson Crusoe (1719), 104*, 193* democracy: Britain as not democratic in ‘long 18th century’, 111, 113–14, 117*, 117

The Revolutionary Center: The Lost Genius of Liberalism

by Adrian Wooldridge  · 7 Apr 2026  · 342pp  · 129,097 words

the giants will be able to survive. The supposedly revolutionary start-ups that come along occasionally and shake up stock markets, such as China’s DeepSeek, are eventually eaten up by the giants: to survive, you need thousands of computers linked to the cloud, thousands of highly paid researchers who can

…

despair xxi, 246 Deaton, Angus 246 debate, open/freedom of xvii, 40–41, 46, 54, 56, 92, 184, 187, 214–16, 313, 317 deepfakes 215 DeepSeek 204 defence spending x, 293–4, 300, 301 deference, collapse of 141 Defoe, Daniel 26 ‘defund the police’ 191 DEI (diversity, equity and inclusion) 137

Gilded Rage: Elon Musk and the Radicalization of Silicon Valley

by Jacob Silverman  · 9 Oct 2025  · 312pp  · 103,645 words

that the United States seemed at risk of losing. While Silicon Valley was floundering, their competitors in China shocked the world with the release of DeepSeek, an AI model reportedly as strong or better than its Western counterparts, made at a fraction of the cost. China’s BYD was also producing

The Means of Prediction: How AI Really Works (And Who Benefits)

by Maximilian Kasy  · 15 Jan 2025  · 209pp  · 63,332 words

years. It is this scale that enables transformers to effectively store a compressed copy of the entire internet. (Recent successes of the Chinese AI company DeepSeek suggest that this scale might be reduced somewhat without loss of performance.) Figure 6 The size of large language models Source: Wikipedia, “Large Language Model

Enshittification: Why Everything Suddenly Got Worse and What to Do About It

by Cory Doctorow  · 6 Oct 2025  · 313pp  · 94,415 words

. It gave us the TikTok ban, and it’s giving America’s bloated, overcapitalized AI giants political cover as they attack nimbler Chinese rivals like DeepSeek. Antitrust is global, but it’s especially powerful in the United States, thanks to a mix of both government and private action. While I was

The Rise and Fall of the Artificial State

by Jill Lepore  · 24 Aug 2026  · 363pp  · 122,400 words