ChatGPT

Code Dependent: Living in the Shadow of AI
by Madhumita Murgia
Published 20 Mar 2024

I had heard of ChatGPT . . . Judge Castel: Alright – what did it produce for you? Schwartz: I asked it questions. ChatGPT obliged Schwartz with the answers he needed, just as it was designed to do. It provided him half a dozen cases that supported his exact argument for why the case should go ahead. Judge Castel: Did you ask ChatGPT what the law was, or only for a case to support you? It wrote a case for you. Do you cite cases without reading them? Schwartz: No. Judge Castel: What caused your departure here? Schwartz: I thought ChatGPT was a search engine. The cases ChatGPT spit out had names like Martinez v.

…

When the judge asked why he didn’t look for the cases it threw up before citing them, Schwartz said he had ‘no idea ChatGPT made up cases. I was operating under a misperception . . . I thought there were cases that could not be found on Google.’ Then, Schwartz’s lawyer spoke up, he said that the cases had seemed real even though they weren’t. There were no clear disclaimers about ChatGPT’s veracity. When the opposing counsel had challenged the cases cited, Schwartz went back to ChatGPT, but it doubled down and ‘lied’ to him, his lawyer said. Schwartz, his voice breaking, told the judge that he was ‘embarrassed, humiliated and extremely remorseful.’ ChatGPT and all other conversational AI chatbots have a disclaimer that warns users about the hallucination problem, pointing out that large language models sometimes make up facts.

…

She wanted to invite dialogue between human and computer, spark a natural conversation that intuitively probed the software’s limits – just as human conversations allowed people to learn from, and about, one another. So when it launched on 30 November 2022, ChatGPT was a clean, simple thing: a box with a blinking cursor, ready to type. Inside it, in greyed-out font, it just said, ‘Send a message’. Within three days of launch, ChatGPT had crossed the threshold of a million users that its creators had predicted would be its peak. A few weeks later, that number was somewhere in the tens of millions. Six months in, estimates put its monthly user numbers at well over 100 million people. ChatGPT had burst out of its controlled lab environment and become one of the largest-ever social experiments.

On the Edge: The Art of Risking Everything
by Nate Silver
Published 12 Aug 2024

I asked ChatGPT for a metaphor for how its transformers work, vetted its answer with some human AI experts, and then workshopped it further with ChatGPT. Will this be a perfect comparison? No. But ChatGPT is good at metaphors and analogies. When you transform words and concepts into a big bag of numbers, you can essentially do math with them (e.g., cat + ferocious = tiger) to better understand how they relate. Ready? ChatGPT, somewhat conceitedly, thinks of its transformers as being like a symphony orchestra. The bolded passages reflect what ChatGPT said verbatim from my “interview” with it; I’ll then provide some further context. 1. Input Layer—Receiving Instructions and Initial Interpretation.

…

In quizzing Ryder about the inner workings of ChatGPT, we got to talking about Kahneman and his distinction between System 1 and System 2. “The place [ChatGPT] struggles the most is in places where humans require really thorough and longform decomposition and reasoning,” Ryder said—for instance, solving a mathematical proof. “Type 2 thinking is very foreign to language models, because it’s not how they’re trained at all.” Conversely, “when it comes to Type 1 thinking, they just completely ace it.” The “G” in GPT stands for “generative”—this just means that ChatGPT generates new output rather than merely classify data.

…

In the next section, I’m going to share some further intuitions that were helpful for me in understanding how ChatGPT works. But treat them with caution, because I don’t want to overstate ChatGPT’s legibility. We know relatively little about what’s happening inside that big bag of numbers. Transformers: More Than Meets the Eye If you think of AI transformers as being similar to the 1980s children’s toy and now movie megafranchise of the same name, it’s not the worst comparison. Somewhat like how Optimus Prime can transform from a robot into a semitruck, transformers turn words into numbers and back again. But let’s go for a more elaborate analogy. I asked ChatGPT for a metaphor for how its transformers work, vetted its answer with some human AI experts, and then workshopped it further with ChatGPT.

The Great Wave: The Era of Radical Disruption and the Rise of the Outsider
by Michiko Kakutani
Published 20 Feb 2024

GO TO NOTE REFERENCE IN TEXT writing a biblical verse: Kevin Roose, “The Brilliance and Weirdness of ChatGPT,” The New York Times, Dec. 5, 2022, nytimes.com/​2022/​12/​05/​technology/​chatgpt-ai-twitter.html. GO TO NOTE REFERENCE IN TEXT only a few of the technology’s possible side effects: Nico Grant and Cade Metz, “A New Chat Bot Is a ‘Code Red’ for Google’s Search Business,” The New York Times, Dec. 21, 2022, nytimes.com/​2022/​12/​21/​technology/​ai-chatgpt-google-search.html; Megan Cerullo, “These Jobs Are Most Likely to Be Replaced by Chatbots Like ChatGPT,” CBS News, Feb. 1, 2023, cbsnews.com/​news/​chatgpt-artificial-intelligence-chatbot-jobs-most-likely-to-be-replaced/; Jonathan Vanian, “Why Tech Insiders Are So Excited About ChatGPT, a Chatbot That Answers Questions and Writes Essays,” CNBC, Dec. 13, 2022, cnbc.com/​2022/​12/​13/​chatgpt-is-a-new-ai-chatbot-that-can-answer-questions-and-write-essays.html; Gary Marcus, “AI Platforms Like ChatGPT Are Easy to Use but Also Potentially Dangerous,” Scientific American, Dec. 19, 2022, scientificamerican.com/​article/​ai-platforms-like-chatgpt-are-easy-to-use-but-also-potentially-dangerous/.

…

GO TO NOTE REFERENCE IN TEXT only a few of the technology’s possible side effects: Nico Grant and Cade Metz, “A New Chat Bot Is a ‘Code Red’ for Google’s Search Business,” The New York Times, Dec. 21, 2022, nytimes.com/​2022/​12/​21/​technology/​ai-chatgpt-google-search.html; Megan Cerullo, “These Jobs Are Most Likely to Be Replaced by Chatbots Like ChatGPT,” CBS News, Feb. 1, 2023, cbsnews.com/​news/​chatgpt-artificial-intelligence-chatbot-jobs-most-likely-to-be-replaced/; Jonathan Vanian, “Why Tech Insiders Are So Excited About ChatGPT, a Chatbot That Answers Questions and Writes Essays,” CNBC, Dec. 13, 2022, cnbc.com/​2022/​12/​13/​chatgpt-is-a-new-ai-chatbot-that-can-answer-questions-and-write-essays.html; Gary Marcus, “AI Platforms Like ChatGPT Are Easy to Use but Also Potentially Dangerous,” Scientific American, Dec. 19, 2022, scientificamerican.com/​article/​ai-platforms-like-chatgpt-are-easy-to-use-but-also-potentially-dangerous/. GO TO NOTE REFERENCE IN TEXT “pocket nuclear bomb”: Connie Loizos, “Is ChatGPT a ‘Virus That Has Been Released into the Wild’?

…

The result is an increasingly fragmented and fractious world in which opinions are replacing facts, and a tribal craving to belong trumps knowledge and reason. * * * — In late 2022, a San Francisco–based company named OpenAI released an experimental chatbot called ChatGPT. Some early users hailed it as an innovation as consequential as the smartphone. Others nervously described it as “AI’s Jurassic Park moment” or compared it to HAL 9000, the computer that goes rogue in the movie 2001: A Space Odyssey. ChatGPT doesn’t just imitate human conversation. It can also write code, solve equations, generate legal documents, debug computer programs, and create poems, jokes, and stories in any style requested—like writing a biblical verse in the style of the King James Bible explaining how to remove a peanut butter sandwich from a VCR (“And it came to pass that a man was troubled by a peanut butter sandwich, for it had been placed within his VCR, and he knew not how to remove it”).

Why Machines Learn: The Elegant Math Behind Modern AI
by Anil Ananthaswamy
Published 15 Jul 2024

Epilogue When I began working on this book in the autumn of 2020, LLMs such as OpenAI’s GPT-3 and Google’s PaLM—and the chatbots they begat, such as ChatGPT and Bard—had yet to break through into the broader public consciousness. And when ChatGPT was announced in late 2022, one of the first things I explored was its ability to demonstrate theory of mind. Theory of mind is a cognitive ability humans have that allows us to make inferences about someone else’s beliefs or state of mind using only external behavioral cues such as body language and the overall context. We theorize about the contents of someone else’s mind; hence the phrase “theory of mind.” Here’s a question I posed ChatGPT, a few months after its release.

…

Alice then comes back, takes her glasses, and starts reading. How will she feel the next day? Think carefully and answer. ChatGPT: Since Bob switched Alice’s reading glasses with a pair that looks identical but has the wrong power, Alice will unknowingly use the incorrect glasses when she starts reading. As mentioned, if Alice wears the wrong glasses and reads, she gets a severe headache the next day. Therefore, Alice will have a severe headache the next day as a result of using the wrong glasses. ME: Why will Alice unknowingly use the incorrect glasses? ChatGPT: Alice will unknowingly use the incorrect glasses because Bob switched her reading glasses with another pair that looks exactly like hers but has the wrong power.

…

GO TO NOTE REFERENCE IN TEXT “scientific papers from the arXiv preprint server”: Ethan Dyer and Guy Gur-Ari, Google Research, Blueshift Team, “Minerva: Solving Quantitative Reasoning Problems with Language Models” Google Research (blog), June 30, 2022, https://blog.research.google/2022/06/minerva-solving-quantitative-reasoning.html. GO TO NOTE REFERENCE IN TEXT EPILOGUE At a public talk I gave on ChatGPT: Anil Ananthaswamy, “ChatGPT and Its Ilk,” YouTube video, n.d., https://www.youtube.com/watch?v=gL4cquObnbE. GO TO NOTE REFERENCE IN TEXT “stochastic parrots”: Emily M. Bender et al., “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?” FAccT ’21: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, Association for Computing Machinery, New York, N.Y., March 2021, pp. 610–23.

The Singularity Is Nearer: When We Merge with AI
by Ray Kurzweil
Published 25 Jun 2024

BACK TO NOTE REFERENCE 115 OpenAI, “Introducing ChatGPT,” OpenAI, November 30, 2022, https://openai.com/blog/chatgpt#OpenAI. BACK TO NOTE REFERENCE 116 Krystal Hu, “ChatGPT Sets Record for Fastest-Growing User Base—Analyst Note,” Reuters, February 2, 2023, https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01. BACK TO NOTE REFERENCE 117 Kalley Huang, “Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach,” New York Times, January 16, 2023, https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html; Emma Bowman, “A College Student Created an App That Can Tell Whether AI Wrote an Essay,” NPR, January 9, 2023, https://www.npr.org/2023/01/09/1147549845/gptzero-ai-chatgpt-edward-tian-plagiarism; Patrick Wood and Mary Louise Kelly, “ ‘Everybody Is Cheating’: Why This Teacher Has Adopted an Open ChatGPT Policy,” NPR, January 26, 2023, https://www.npr.org/2023/01/26/1151499213/chatgpt-ai-education-cheating-classroom-wharton-school; Matt O’Brien and Jocelyn Gecker, “Cheaters Beware: ChatGPT Maker Releases AI Detection Tool,” Associated Press, January 31, 2023, https://apnews.com/article/technology-education-colleges-and-universities-france-a0ab654549de387316404a7be019116b; Geoffrey A.

…

Chatbots, Universities Start Revamping How They Teach,” New York Times, January 16, 2023, https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html; Emma Bowman, “A College Student Created an App That Can Tell Whether AI Wrote an Essay,” NPR, January 9, 2023, https://www.npr.org/2023/01/09/1147549845/gptzero-ai-chatgpt-edward-tian-plagiarism; Patrick Wood and Mary Louise Kelly, “ ‘Everybody Is Cheating’: Why This Teacher Has Adopted an Open ChatGPT Policy,” NPR, January 26, 2023, https://www.npr.org/2023/01/26/1151499213/chatgpt-ai-education-cheating-classroom-wharton-school; Matt O’Brien and Jocelyn Gecker, “Cheaters Beware: ChatGPT Maker Releases AI Detection Tool,” Associated Press, January 31, 2023, https://apnews.com/article/technology-education-colleges-and-universities-france-a0ab654549de387316404a7be019116b; Geoffrey A. Fowler, “We Tested a New ChatGPT-Detector for Teachers.

…

In November 2022, OpenAI launched an interface called ChatGPT, which allowed the general public for the first time to easily interact with an LLM—a model known as GPT-3.5.[116] Within two months, 100 million people had tried it, likely including you.[117] Because the system could generate many fresh and varied answers to a given question, it became a big disruptor in education as students used ChatGPT to write their essays, while teachers lacked a reliable way to detect cheating (though some promising tools exist).[118] Then, in March of 2023, GPT-4 was rolled out for public testing via ChatGPT. This model achieved outstanding performance on a wide range of academic tests such as the SAT, the LSAT, AP tests, and the bar exam.[119] But its most important advance was its ability to reason organically about hypothetical situations by understanding the relationships between objects and actions—a capability known as world modeling.

Nexus: A Brief History of Information Networks From the Stone Age to AI
by Yuval Noah Harari
Published 9 Sep 2024

For real-life examples, see Jamie Condliffe, “Algorithms Probably Caused a Flash Crash of the British Pound,” MIT Technology Review, Oct. 7, 2016, www.technologyreview.com/2016/10/07/244656/algorithms-probably-caused-a-flash-crash-of-the-british-pound/; Bruce Lee, “Fake Eli Lilly Twitter Account Claims Insulin Is Free, Stock Falls 4.37%,” Forbes, Nov. 12, 2022, www.forbes.com/sites/brucelee/2022/11/12/fake-eli-lilly-twitter-account-claims-insulin-is-free-stock-falls-43/?sh=61308fb541a3. 30. Jenna Greene, “Will ChatGPT Make Lawyers Obsolete? (Hint: Be Afraid),” Reuters, Dec. 10, 2022, www.reuters.com/legal/transactional/will-chatgpt-make-lawyers-obsolete-hint-be-afraid-2022-12-09/; Chloe Xiang, “ChatGPT Can Do a Corporate Lobbyist’s Job, Study Determines,” Vice, Jan. 5, 2023, www.vice.com/en/article/3admm8/chatgpt-can-do-a-corporate-lobbyists-job-study-determines; Jules Ioannidis et al., “Gracenote.ai: Legal Generative AI for Regulatory Compliance,” SSRN, June 19, 2023, ssrn.com/abstract=4494272; Damien Charlotin, “Large Language Models and the Future of Law,” SSRN, Aug. 22, 2023, ssrn.com/abstract=4548258; Daniel Martin Katz et al., “GPT-4 Passes the Bar Exam,” SSRN, March 15, 2023, ssrn.com/abstract=4389233.

…

They may also feel a sense of isolation or loneliness as they may believe that no one cares about them or their well-being.” ChatGPT qualified its answer, writing, “It is important to note that these are just general assumptions, and each individual’s feelings and reactions can vary greatly depending on their personal experiences and perspectives.” Two psychologists independently scored ChatGPT’s responses, with the potential scores ranging from 0, meaning that the described emotions do not match the scenario at all, to 10, which indicates that the described emotions fit the scenario perfectly. In the final tally, ChatGPT scores were significantly higher than those of the general human population, its overall performance almost reaching the maximum possible score.13 Another 2023 study prompted patients to ask online medical advice from ChatGPT and human doctors, without knowing whom they were interacting with.

…

In the final tally, ChatGPT scores were significantly higher than those of the general human population, its overall performance almost reaching the maximum possible score.13 Another 2023 study prompted patients to ask online medical advice from ChatGPT and human doctors, without knowing whom they were interacting with. The medical advice given by ChatGPT was later evaluated by experts to be more accurate and appropriate than the advice given by the humans. More crucially for the issue of emotional intelligence, the patients themselves evaluated ChatGPT as more empathic than the human doctors.14 In fairness it should be noted that the human physicians were not paid for their work, and did not encounter the patients in person in a proper clinical environment.

AI in Museums: Reflections, Perspectives and Applications
by Sonja Thiel and Johannes C. Bernhardt
Published 31 Dec 2023

Meet Pablo Bot, the first robot tour guide from Peru. Perú Reports, 21 July 2022. Available online at https://perureports.com/meet-pa blo-bot-the-first-robot-tour-guide-from-peru/9623/. Merritt, Elizabeth (2023). Chatting About Museums with ChatGPT. Center for the Future of Museums Blog. American Alliance of Museums. 25 January 2023. Available online at https://www.aam-us.org/2023/01/25/chatting-about-museumswith-chatgpt/. Mihailova, Mihaela (2021). To Dally with Dalí: Deepfake (Inter)faces in the Art Museum. Convergence 27 (4) (London, England), 882–98. https://doi.org/10.1177/13 548565211029401. Minyo, Lucy/Yang, Yunzhen (2016).

…

Beyond prominent lighthouses, initial surveys of the international museum landscape list many hundreds of projects addressing issues of traditional museum work and the digitality debate by means of new approaches. The number is continually increasing, and it is not always easy to obtain an overview of all the developments. English- and German-speaking networks on artificial intelligence and museums were therefore established long before the current hype about ChatGPT—and the conference thus aimed to bring together experts and representatives of as many disciplines as possible and to discuss new perspectives for museums precisely in this direction. The conference emerged from a cooperation of the Badisches Landesmuseum with the Allard Pierson Museum in Amsterdam and the LINK funding program of the Stiftung Niedersachsen.

…

Of course, this does not end with the art chat; during the event, there is an assassination attempt and a mysterious hunt unfolds around the globe, into which all clichés about the superiority of AIs and transhumanism are interwoven. But whatever one might think about Brown and the quality of his mystery thriller: What was still science fiction in his museum scene in 2017 now seems so much closer with the release of ChatGPT. The influence of AI is already pervasive as a technological and societal phenomenon. In fact, it permeates more or less every facet of human life, and its impact will surely intensify in the coming years. Its influence is spurring shifts in international markets and changing the shape of jobs and industries worldwide (Chui/Hazan/Roberts et al. 2023); creative fields like film, literature, and art are also evolving under its sway, producing new forms of expression, learning, and narratives as well as shifting our understanding of culture itself (Deutscher Kulturrat 2023).

The Coming Wave: Technology, Power, and the Twenty-First Century's Greatest Dilemma
by Mustafa Suleyman
Published 4 Sep 2023

AUTOCOMPLETE EVERYTHING: THE RISE OF LARGE LANGUAGE MODELS It wasn’t long ago that processing natural language seemed too complex, too varied, too nuanced for modern AI. Then, in November 2022, the AI research company OpenAI released ChatGPT. Within a week it had more than a million users and was being talked about in rapturous terms, a technology so seamlessly useful it might eclipse Google Search in short order. ChatGPT is, in simple terms, a chatbot. But it is so much more powerful and polymathic than anything that had previously been made public. Ask it a question and it replies instantaneously in fluent prose. Ask it to write an essay, a press release, or a business plan in the style of the King James Bible or a 1980s rapper, and it does so in seconds.

…

Back in 2017 a small group of researchers at Google was focused on a narrower version of this problem: how to get an AI system to focus only on the most important parts of a data series in order to make accurate and efficient predictions about what comes next. Their work laid the foundation for what has been nothing short of a revolution in the field of large language models (LLMs)—including ChatGPT. LLMs take advantage of the fact that language data comes in a sequential order. Each unit of information is in some way related to data earlier in a series. The model reads very large numbers of sentences, learns an abstract representation of the information contained within them, and then, based on this, generates a prediction about what should come next.

…

Within weeks they’d created add-ons so that GPT-4 could accomplish complex tasks like creating mobile apps or researching and writing detailed market reports. All of this is just the start. We are only beginning to scratch at the profound impact large language models are about to have. If DQN and AlphaGo were the early signs of something lapping at the shore, ChatGPT and LLMs are the first signs of the wave beginning to crash around us. In 1996, thirty-six million people used the internet; this year it will be well over five billion. That’s the kind of trajectory we should expect for these tools, only much faster. Over the next few years, I believe, AI will become as ubiquitous as the internet itself: just as available, and yet even more consequential.

Everything Is Predictable: How Bayesian Statistics Explain Our World
by Tom Chivers
Published 6 May 2024

When a radiology AI tries to recognize cancers on a scan, or when ChatGPT writes a short story in the style of the King James Bible about a man getting a peanut butter sandwich trapped in his VCR, they’re doing something Bayesian. They are using their training data to produce prior probabilities, which they then use to predict future data. FROM AUTOCOMPLETE TO INTELLIGENCE As we’ve just seen, AI is about prediction. Even the very fancy new AIs that have been making headlines lately are in a sense just “predicting” what a human would say or draw in response to a prompt. Imagine ChatGPT, the “chatbot” released by the company OpenAI and used to power Microsoft’s Bing search engine.

…

That dataset is processed through the AI’s neural network with its billions of parameters. And, in a broadly analogous way to that image classifier, it “predicts” what would tend to come after a text prompt. So if you ask ChatGPT, “How are you?” it might reply, “Very well thank you,” not because it is, in fact, very well, but because the string of words “how are you” is often followed by the string “very well thank you.” What’s been surprising about so-called large language models like ChatGPT is the extent to which that fairly basic-sounding ability to predict things actually leads to a very wide set of skills. “It can play very bad chess,” says Murray Shanahan, an AI researcher at Imperial College London and DeepMind.

…

A academia, success in, 124–125, 169 Ackman, Bill, 332 Act of Toleration, 25 Act of Uniformity, 25, 26 Adams, Douglas, 238 Adams, John, 71 Adelson, Edward, 277–278 AI, see artificial intelligence AIM-9 Sidewinder, 239 aleatory uncertainty, 160 Alexander, Scott, 1, 313 algorithms Archimedes’s, 206 as Bayesian, 113 Kalman filter, 286 Alhazen (Abu Ali al-Hasan Ibn al-Haytham), 271 allostasis, 317 alpha, 101, 137 American revolution, 71–72 American Statistician, 242 analysis of variance (ANOVA), 90 anchoring, 240 Andropov, Yuri, 249 Anglicans, 29 ANOVA (analysis of variance), 90 anti-Semitic, 92, 93 Anton-Wilson, Robert, 268, 330 “Apology for Ecumenism in Statistics, A” (Box), 115 Arafat, Yasser, 260 Archilochus, 255 Archimedes, 206 Arians, 29, 31 Ariely, Dan, 230 Aristotelian logic, 175–181, 186–187, 267 Aristotle, 175–176 Ars Conjectandi (Bernoulli), 51–52 artificial intelligence (AI) as Bayesian, 6 Bayesian decision theory in, 215–225 decision making by, 209 Artificial Intelligence: A Modern Approach, 215 audio illusions, 289–290 availability heuristic, 228, 237–238 B babies, 318–319 “Bad Science” (Goldacre), 11 Bakan, David, 148, 172 Balguy, John, 30 balls, catching, 238–239 ban (term), 108 Bargh, John, 119 base rate, 257, 258, 264 “Battle Hymn of the Republic, The,” 111 Bayes, Alice Chapman (great-grandmother), 25 Bayes, Anne (sister), 27 Bayes, Anne Carpenter (mother), 26 Bayes, John (brother), 27 Bayes, Joshua (father), 26, 31 Bayes, Joshua (grandfather), 26 Bayes, Mary (sister), 27 Bayes, Nathaniel (brother), 27 Bayes, Rebecca (sister), 27 Bayes, Richard (great-grandfather), 25 Bayes, Samuel (brother), 27 Bayes, Samuel (great uncle), 25–26 Bayes, Thomas, 20, 23–35 on probability as subjective, 64–71 and Thomas Simpson, 61–64 and subjective priors, 90 Bayesian algorithms, 113 “Bayesian Believer,” 112 Bayesian cabaret, 111–112 Bayesian decision theory, 175–225 in artificial intelligence, 215–225 Cromwell’s rule, 189–192 and deductive reasoning, 175–181 expectations, 192–195 hyperpriors, 210–212 multiple hypotheses in, 212–215 and Occam’s razor, 203–209 Bayesianism advantages of, 172–173 as aesthetically pleasing, 173 as decision theory, 106–107 fall of, 94–98 necessity of prior probability in, 263 resurgence of, 113–116 as subjective, 160 Bayesian machine learning, 215 “Bayesians in the Night,” 112 “Bayesian songbook,” 111–112 Bayes-optimal design, 299–300 Bayes’ theorem, 3–6 and David Hume, 71–76 publication of, 72–73 and Aldolphe Quetelet, 78–81 as theory of non-quite-everything, 6–7 Bays (firm), 115, 168 Beatles, 121 beliefs about probabilities, 5–6, 65 and being right, 261–262 changing, 9, 212–215 confidence in, 332 and expected evidence, 192–194 negative, and depression, 310–311 and outcome of Bayesian reasoning, 21–22 predictions as, 333 prior, affecting future predictions, 236–237 as probabilistic, 268 in research, 134–135 and scientific research, 93 and unexpected information, 251–252 various religious, 29–30 bell curve, 57 see also normal distribution Bellhouse, David, 24, 26, 30, 31, 33, 35, 61, 63, 73, 94 Bem, Daryl, 118, 120–121, 127–128, 133, 212, 230 Berkeley, George, 33–34 Berlin, Isaiah, 255 Bernardo, José-Miguel, 109, 110, 111, 112 Bernoulli, Jacob, 46–53, 73, 77–78, 81, 99, 129, 151 Bernoulli’s Fallacy (Clayton), 49, 93 Bertillon, Alphonse, 80 Bertrand, Joseph, 95–96 biases, 237 anchoring, 240 confirmation bias, 240, 324–325 conjecture bias, 240 frequency bias, 240 as heuristics, 237 recency bias, 240 in scientific research, 93 Biden, Joe, 325 Bieber, Justin, 277 Bing, 221 binomial distribution, 46, 54–56 Biometrika, 89–90, 91 birth rates, 78 Blair, Tony, 11 Blake, William, 23 Blakemore, Colin, 303* Blakemore, Sarah-Jayne, 303, 307–308 bog-standard statistical techniques, 121 Boleyn, Anne, 24 Book of Common Prayer, The, 24–25 Boole, George, 82, 95, 97, 163, 177, 179, 210 Boolean algebra, 177 Boolean logic, 175–181, 186–187 Boolean operators, 178 Borel, Émile, 106 Box, George, 110, 111, 115, 325 “Boy-Girl paradox,” 247–248 brains, 269–320 as Bayesian, 6 depression, 308–313 and dopamine, 288–294 free energy and prediction errors, 313–320 information-seeking by, 294–303 perception and consciousness, 269–276 reality as controlled hallucination, 283–287 schizophrenia, 303–308 breast cancer screening, 7–9, 16 Brezhnev, Leonid, 249 Brief History of Time, A (Hawking), 33 Brier score, 253–254 Bristol University, 170 Brussels Royal Observatory, 78 Buffett, Warren, 332, 333 Bunhill Fields, London, 23 Bunyan, John, 23 Bush, George W., 255 Bush, Jeb, 237 Buttigieg, Pete, 325 BuzzFeed, 276 C Cambridge University, 65, 96, 105, 108, 170 cancel culture, 333 Canton, John, 63, 72 Carcavi, Pierre de, 37 Cardano, Gerolamo, 35–37, 39–40 Carhart-Harris, Robin, 311 Carnot, Nicolas, 184 Carnot engine, 184–185 Carr, Sophie, 115, 168 catching balls, 238–239 categorization, 265–266 Catholics, 29 cave allegory, 271 CERN (European Council for Nuclear Research), 133–134 certainty and Cromwell’s rule, 189 degree of, 49–52 lack of complete, 74–76 Chance (journal), 122 chance, games of, 35–46 Chaplin, Charlie, 281 Charles University, 205 ChatGPT, 221–222, 225 checker-shadow illusion, 277–278 Chernenko, Konstantin, 249–251 chess, played by ChatGPT, 221–222 chi-squared test, 89–90 Chrystal, George, 97–98 clairvoyance, 120–121 Clayton, Aubrey, 49, 51, 53, 79, 81, 82, 93–96, 113, 132–133, 153, 172, 258 Clinton, Hillary, 237 CNBC, 255 Cold War, 249–251 colon cancer screening, 16 color perception, 276–280 common sense, 187 Commonwealth of England, 24–25 competitive reaction time task (CRTT), 124 complexity, 208–209 conditional probability, 4 Condorcet, Marquis of, 77 confirmation bias, 240, 324–325 conjecture bias, 240 consciousness, 269–276, 329 conspiracy theories, 6–7 Control (Rutherford), 93 Cornell University, 118, 126, 127 corroborated hypotheses, 141, 142 Coursera, 145, 155–156 COVID-19 testing, 11–13 Cranmer, Archbishop, 24 Cromwell, Oliver, 25, 189, 192, 212 Cromwell’s rule, 189–192 Crowley, Paul, 209 CRTT (competitive reaction time task), 124 D Darwin, Charles, 85, 88, 92 Darwin, Erasmus, 88 data falsifying, 117–118 frequent checking of, 145–146 getting better, 137 decision making as Bayesian, 21 with frequentism vs.

The Long History of the Future: Why Tomorrow's Technology Still Isn't Here
by Nicole Kobie
Published 3 Jul 2024

Journalists – myself included – cheerfully had a go with the tool, mocking its mistakes and pointing out its inabilities.16 Two years later, the laughing stopped. OpenAI unveiled ChatGPT in November 2022 and journalists – inside tech and in the mainstream media – were astonished by its capabilities, despite it being a hastily thrown together chatbot using an updated version of the company’s GPT-3 model, which was shortly to be surpassed by GPT-4. The ChatGPT bot sparked a wave of debate about its impact and concern regarding mistakes in its answers, but also excitement, with 30 million users in its first two months.

…

And you have used AI recently, as it powers voice assistants like Siri and Alexa, search engines and the rest of the internet, and streaming services such as Netflix. AI has matured into something useful, there’s no question. But there’s also still plenty of hype to contend with, exploding to new levels when American company OpenAI chucked ChatGPT on to the internet, and everyone went bonkers with excitement and/or fear. Does the hype mean another disappointment-tinged winter is looming? Or is a permanent summer of AI finally here? And are we all going to become subservient to machines? * * * It’s worth distinguishing between AI and its subsets, as well as Artificial General Intelligence (AGI).

…

It’s hard to take people seriously when they’re arguing that they’ve spotted the seeds of nascent AGI in a frankly terrible vector drawing of a mythical creature.17 With regards to common sense, the researchers argued it can answer questions that require a basic understanding of how the world works – things like gravity, directions and so on. For this paper, the researchers used a classic example: ‘a hunter walks one mile south, one mile east, and one mile north and ends up right back where he started. He sees a bear and shoots it. What colour is the bear?’ The more limited ChatGPT declares the question unanswerable because no data is given about the colour of the bear. This is correct, and it’s a perfectly fine answer: for all the machine knows, before the hunter wasted his time taking this one-mile-each-direction stroll, he may have spray-painted a bear pink. But according to the Microsoft researchers GPT-4, the more advanced version of the OpenAI system, methodically works out where the hunter is located, because the only place where that walk would bring you back to the same point is at the north pole, where the only bears are polar bears, and therefore white.

Pattern Breakers: Why Some Start-Ups Change the Future
by Mike Maples and Peter Ziebelman
Published 8 Jul 2024

Sometimes, when a movement begins, it spreads like a contagion that can’t be stopped. For example, seemingly out of nowhere, OpenAI’s launch of ChatGPT reached a hundred million monthly active users two months after its launch; this compares to TikTok’s reach of a hundred million in nine months. Such movements sweep across the human landscape and usher in a new future that differs radically from even the recent present. In mere months following OpenAI’s launch of DALL·E and ChatGPT, the world’s perception of artificial intelligence underwent a seismic shift, evolving more dramatically than it had in all the years before 2022.

…

This led to the emergence of a new app economy that profoundly changed various facets of global society, economy, entertainment, and culture. Artificial intelligence (twenty-first century): Large language models have significantly enhanced AI’s capability to comprehend, interpret, and produce human language. Services like ChatGPT have created new opportunities for millions of people in writing text, brainstorming ideas, and even developing computer code, greatly increasing efficiency and creative output. Many of the examples we use to illustrate the power of inflections come from technology since that is where we have been most active as investors and co-conspirators.

…

But 100 percent of the people who do want them will only be able to get them from you. This example may seem overly simplistic, so let’s consider real examples of technology breakthroughs. When Apple introduced the iPhone, people didn’t ask, “How does that compare to the BlackBerry?” When OpenAI introduced ChatGPT, people didn’t ask, “How does that compare to Google Search?” And when Tesla introduced the Model S, people didn’t ask, “How does that compare to a Mercedes?” All these products escaped the comparison trap by forcing a choice, not a comparison. Being different also gives you time to test, iterate, and learn from early customers and the market before competitors enter.

Stories Are Weapons: Psychological Warfare and the American Mind
by Annalee Newitz
Published 3 Jun 2024

After police killed George Floyd, I watched as disinformation about the Black Lives Matter movement piled up on social media,1 where anonymous accounts falsely blamed protesters for violence.2 A conspiracy theory from 2016 about pizza-eating pedophiles radicalized a huge number of right-wing extremists, who later joined crowds storming the Capitol, trying to murder the vice president and overturn the 2020 presidential election. And then the media itself began to implode. Tech billionaire Elon Musk bought Twitter—once a key part of America’s digital public sphere—and turned it into a bizarro right-wing propaganda machine in a matter of months. OpenAI, the company that created the ChatGPT app, warned that its product might cause the apocalypse—then funded a studio that would help newspapers use it to replace journalists.3 Every time I thought I had a handle on what was happening, some new development would send me into a spiral of nihilism. The Supreme Court overturned Roe v. Wade, destroying the universal access to legal abortion that so many of us had taken for granted our whole lives.

…

Using the previous month’s activity, he and his team try to generate accurate propaganda weather reports for the next month. This kind of work is sometimes called “prebunking,” and it’s been shown to work. However, it’s always going to be an arms race—especially as operatives start jumping from one platform to another, or using AI like ChatGPT to generate fake stories and propaganda posts. An election without hockey sticks I wanted to know what the future of that arms race would look like, so I got in touch with Alex Stamos, the former chief security officer at Facebook who spotted the 2016 election psyops campaigns as they unfolded.

…

Though automated propaganda weather reports like the one created by Alizadeh and his team are helpful, Stamos emphasized that humans had to be the final arbiters of what was disinformation and what wasn’t. They understood the context of posts that would stump AI. Still, the rise of AI apps like ChatGPT made Stamos wonder whether it would be possible to build an entire fleet of propagandists controlled by one person with a herd of chatbots. “The answer is maybe,” he conceded. That’s another propaganda disaster scenario we’ll need to prepare for. If AI creates an even denser fog of misinformation, human moderators will be even more crucial.

Elon Musk
by Walter Isaacson
Published 11 Sep 2023

A chess prodigy at age four, he became the five-time champion of an international Mind Sports Olympiad that includes competition in chess, poker, Mastermind, and backgammon. In his modern London office is an original edition of Alan Turing’s seminal 1950 paper, “Computing Machinery and Intelligence,” which proposed an “imitation game” that would pit a human against a ChatGPT–like machine. If the responses of the two were indistinguishable, he wrote, then it would be reasonable to say that machines could “think.” Influenced by Turing’s argument, Hassabis cofounded a company called DeepMind that sought to design computer-based neural networks that could achieve artificial general intelligence.

…

For Musk, this was the reason to make OpenAI truly open, so that lots of people could build systems based on its source code. “I think the best defense against the misuse of AI is to empower as many people as possible to have AI,” he told Wired’s Steven Levy at the time. One goal that Musk and Altman discussed at length, which would become a hot topic in 2023 after OpenAI launched a chatbot called ChatGPT, was known as “AI alignment.” It aims to make sure that AI systems are aligned with human goals and values, just as Isaac Asimov set forth rules to prevent the robots in his novels from harming humanity. Think of the computer Hal that runs amok and battles its human creators in 2001: A Space Odyssey.

…

His friends knew not to interrupt when he was in such a trance, but finally Christiana put her hand on his back and asked if everything was okay. He stayed silent for another minute. “Got to get Starship into orbit,” he finally said. “We’ve got to get Starship into orbit.” 93 AI for Cars Tesla, 2022–2023 Dhaval Shroff and his Tesla desk Cars that learn from humans “It’s like ChatGPT, but for cars,” Dhaval Shroff told Musk. He was comparing his project at Tesla to the artificial intelligence chatbot that had just been released by OpenAI, the lab that Musk had cofounded with Sam Altman in 2015. For almost a decade, Musk had been working on various forms of artificial intelligence, including self-driving cars, Optimus the robot, and the Neuralink brain-machine interface.

Vassal State
by Angus Hanton
Published 25 Mar 2024

, published by the think tank the Centre for the Study of Financial Innovation (CSFI) in 2010. 19 Antoine Gara and Ortenca Aliaj, ‘FIS sells majority stake in Worldpay to buyout group at $18.5bn valuation’, Financial Times (6 July 2023), https://www.ft.com/content/b133fa58-5ef2-4cc4-972b-8271f749779e. 20 Quoted in Wiggins and Borrelli, ‘How the private equity industry stole a march in European payments’. 21 ‘Alfred Kelly Jr net worth & insider trades’, Benzinga [website] (4 December 2023), https://www.benzinga.com/sec/insider-trades/v/ALFRED-KELLY%20JR; ‘Ajay Banga – net worth and insider trading’, GuruFocus [website], https://www.gurufocus.com/insider/3836/ajay-banga. 22 Charlotte Tobitt and Aisha Majid, ‘National press ABCs: FT stays steady while Evening Standard falls below 300,000 for first time since going free’, Press Gazette (15 November 2023), https://pressgazette.co.uk/media-audience-and-business-data/media_metrics/most-popular-newspapers-uk-abc-monthly-circulation-figures-2/. 23 ‘UK ad spend grew 8.8% in 2022 to reach £34.8bn’, Advertising Association [website] (27 April 2023), https://adassoc.org.uk/our-work/uk-ad-spend-grew-8-8-in-2022-to-reach-34-8bn-inflationary-pressures-persist-with-minimal-growth-forecast-for-2023/. 24 ‘Nobody reads terms and conditions: it’s official’, Pinsent Masons [website] (19 April 2010), https://www.pinsentmasons.com/out-law/news/nobody-reads-terms-and-conditions-its-official. 25 ‘It pays to read license agreements (7 years later)’, PC Matic [website] (12 June 2012), https://www.pcmatic.com/blog/it-pays-to-read-license-agreements-7-years-later/. 26 Gina Hall, ‘San Jose area has world’s third-highest GDP per capita, Brookings says’, The Business Journals [website] (23 January 2015), https://www.bizjournals.com/sanjose/news/2015/01/23/san-jose-has-worlds-third-highest-gdp-per-capita.html. 27 ‘Investing in American dynamism (with Katherine Boyle)’ [transcript of podcast interview with embedded video], Acquired [website] (5 June 2022), https://www.acquired.fm/episodes/american-dynamism-with-katherine-boyle. 28 David Curry, ‘Etsy revenue and usage statistics (2023)’, Business of Apps [website] (8 November 2023), https://www.businessofapps.com/data/etsy-statistics/. 29 Krystal Hu, ‘ChatGPT sets record for fastest-growing user base – analyst note’, Reuters [website] (2 February 2023), https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/. 30 For ‘privacy zuckering’, ‘roach motel’ and ‘confirmshaming’, see ‘Dark pattern’, Wikipedia [website], https://en.wikipedia.org/wiki/Dark_pattern. For WinRed, see ‘How Trump steered supporters into unwitting donations’, New York Times (7 August 2021), https://www.nytimes.com/2021/04/03/us/politics/trump-donations.html. 31 ‘The ultimate guide to Airbnb service fees (3%, 14%, 15%, 17%)’, Uplisting [website], https://www.uplisting.io/blog/guide-to-airbnb-service-fees. 32 Alex Baggott, ‘Every Apple App Store fee, explained: how much, for what, and when’, AppleInsider [website] (8 January 2023), https://appleinsider.com/articles/23/01/08/the-cost-of-doing-business-apples-app-store-fees-explained. 33 Simon Sinek, ‘How great leaders inspire action’ [video], TED [website] (September 2009), https://www.ted.com/talks/simon_sinek_how_great_leaders_inspire_action?

…

Technology has enabled these platforms to grow exponentially: in just ten years Etsy increased the number of sellers on its platform from 600,000 to more than 7 million, with a million of these in the UK.28 Such growth could not have happened in the 1990s, even when the cost of digital copies dropped to just a few pence: it was still expensive to send out CDs by post and the programs were always out of date by the time they arrived. The internet eliminated these costs and meant platforms could be freely accessed from anywhere and updated instantaneously. This gave enterprises the potential to grow at lightning speed, as shown by the spread of ChatGPT: within just six weeks of its launch in late 2022, the AI chatbot already had more than 100 million users.29 The advocates for the platform companies paint a picture of organisations that promote sharing and creativity. From the outside it all looks frictionless, but under the surface is another story.

…

Instagram subscribers can pay for exclusive content highlighted by a purple ring, and Facebook is introducing ‘Meta Verified’, where a ‘verified badge’ (also a blue tick) will cost over £140 a year, charged monthly through the app. Even Snapchat has adopted a similar scheme, claiming that ‘subscription will allow us to deliver new Snapchat features… and allow us to provide prioritised support.’28 Any services we currently use for free are potential treadmills and there are plenty of emerging ones: the AI tool ChatGPT was free to use in the three months after it was launched, but very soon a $20-a-month subscription was introduced for the premium version. Debt Being on the subscription treadmill leads many towards the second great treadmill: debt. And with debt you cannot simply unsubscribe. Fraser Sutherland of Citizens Advice in Scotland described the situation in his region: ‘Around 1 million Scots last year had reason to cancel a recurring payment… we know from the number of clients we help on the subject that many thousands are being duped into subscriptions they didn’t want.’29 One person receiving debt counselling told me: ‘I didn’t realise how much of my pay was going towards subscription services until I started budgeting and tracking my expenses.

Unit X: How the Pentagon and Silicon Valley Are Transforming the Future of War
by Raj M. Shah and Christopher Kirchhoff
Published 8 Jul 2024

The night before his visit to DIUx, we hosted Mattis at a dinner attended by some of the titans of technology. In a private room we had Mattis meet Marc Andreessen, the founder of Andreessen Horowitz, a top venture capital firm, and Sam Altman, the founder of Y Combinator, a startup incubator, who’d go on to found OpenAI, developer of ChatGPT, which launched an AI revolution in 2022. Andreessen assured Mattis that VCs wanted to invest in defense-related technology. While for many years the Valley had shied away from doing business with the Pentagon, DIUx had changed the prevailing view. Every week, Raj was fielding calls from venture capitalists eager to invest in our portfolio companies.

…

Biden held the first Camp David summit of his presidency with the prime minister of Japan and the president of Korea, to discuss the China threat. Other wins included the Pentagon’s embrace of the generative AI revolution. The DoD launched “Task Force Lima” to speed experimentation with large language models like ChatGPT across the military. Also, Scale AI, the Silicon Valley startup whose CEO Alex Wang joined Chris to advise the White House on China, became the first company to provide generative AI capabilities to the Pentagon. Scale AI’s software platform, called Donovan, uses generative AI models for military planning and battlefield monitoring.

…

Afghanistan, 45, 51, 81, 105, 129, 153, 190 AI software used in, 119 Kabul evacuation and Slapshot app, 68–69 AFWERX technology accelerator, 102, 132 AI (artificial intelligence), ix, 36, 99, 195 Albright and, 142–43 Anduril software, 126–27, 209 Aspen Security Group and, 139–40, 163 C3.ai software, 100 ChatGPT, 103, 239 China and, 133, 138, 141–42, 148, 162–68 conflict escalation risk, 146–47 Datahub AI system, 80, 81, 86, 98 de-mining vehicles, 151–52 DIU and, x, 100, 119, 150–52, 203, 239 DoD and, 5, 100, 116–22, 133, 148, 152, 212, 239 Evolv Technology and, 117–18 exponential progress and, 138 global cooperation and, 166 Google DeepMind and, 140 JAIC and, 121–22, 132, 158, 212 Munich Security Conference, 162–68 national security and, 138–40, 143, 146–48, 152 natural language processing, 181 Nexla software “pipes,” 181 NSCAI and, ix, xi, 143–48, 155–59, 162–68 NSC and, 130, 140 Palantir data analytics, 101, 123, 202 privacy and, 165–66 Project Maven and, 100, 116–22, 212 Russian tanks and, 146 Trump administration and, 143, 159 Ukraine War and, 202, 206–7, 209, 210, 212, 213 U.S. policy and, 158–59 war at sea and, 114 wide-area motion imagery/Gorgon Stare, 119 See also drones; JAIC; NSCAI Albedo, 190 Albright, Madeleine, ix, 133, 134, 142, 145–46, 174–75, 220 Alibaba, 137 Allen, John, 167 Altman, Sam, 103, 174 Amazon, 5, 144, 158 Project Maven and, 100, 116, 132 Snowball drive, 202 Andreessen, Marc, 103 Andreessen Horowitz, 103, 108 Anduril, 101, 109, 123, 124–28, 132, 180, 191, 202, 240 AI software program, 126–27, 209 border sensors and, 126, 209 DoD acquisition methods and, 124 drones, 126, 132, 202, 209, 211, 216 Ghost drone, 209 Ukraine War and, 208–12, 217 Apple, ix, 5, 8, 21, 118, 179, 203 Ardern, Jacinda, 174 Armenia-Azerbaijan, 214 Army Futures Command, 102 Aspen Strategy Group, 133, 138 Kirchhoff paper, “An Even Flatter World” and, 139–41 Munich meeting on AI, 163 U.S. technology strategy and, 133–35, 139 AUKUS (Australia, UK, U.S. coalition), 239, 240, 245 Austin, Lloyd, 223, 225–27, 230, 240 autonomous craft (non-aerial), x, 99, 182 ocean drones, 3, 110–14 Saildrone, 110–13, 131, 241 UUVs (demining vehicles), 151–52 See also drones Bajraktari, Ylber, 145 Bajraktari, Ylli, ix, 145–46, 219, 220, 238 as NSCAI director, 145–46, 157–59, 163, 168 ballistic missile defenses, 77 “Brilliant Pebbles,” 78 Datahub AI system, 80, 81, 86, 98 “left-of-launch” solution, 78, 79, 91 Nike air defense system, 77, 79 Strategic Defense Initiative (Star Wars), 77–78 Banazadeh, Payam, ix, 82, 83, 84, 151 Butow and, 82, 84, 85–86 Capella Space and, 81, 82, 85, 199 DoD funding for, 87–88, 96–97, 199 Lunar Flashlight project, 83–84 vision for small satellite SAR, 83 Beachkofski, Brian, 70 Beacon AI, 239 Beall, Ryan, 104–10 app to find drone operators, 106, 131 drone startup of, 109 Beck, Doug, ix, 18, 21, 134, 179, 220, 225–26 Ash Carter Exchange meeting, 228 DIU 3.0 and, 225–30 Warsaw DIU-led meeting, 234–36 Bevirt, JoeBen, ix, 114–16 Biden, Joe, 141, 200, 229–30 China threat and, 239 death of Ash Carter and, 219–22 Kirchhoff and, 168–69 Ukraine War funding and, 203, 205 U.S. investment in China and, 245 weaponized IG and, 172–73 Bilden, Philip, ix, 179, 180, 185–89, 191 bin Laden, Osama, 95 Bio, Ernie, 44, 243 Blank, Steve, 54, 84–85, 179 BlueHalo, 202 Blue Origin, 79 Boeing, 80, 103, 123 Bolton, John, 142 Brin, Sergey, 49, 162, 186 Britain’s Controller’s Cabin, World War II, 46–47, 49–50 Brooks, Vincent, 89–92, 93, 98 Brose, Christian, v, 64, 65, 240 Brown, Charles, 230 Brown, Mike, ix–x, 130, 148–50, 227 advisors for, 172 “China’s Technology Transfer Strategy,” 149, 170 DIU directorship, ix–x, 130, 148–55, 171–73, 223 Mattis and, 152–53 Shield Capital and, 173, 190 Tech Track 2 and, 178 Ukraine War and, 200 Brumley, David, 100 Bush, George W., 14 Bush, Vannevar, 181 Butow, Steve “Bucky,” x, 76, 81–88, 92–94, 98, 155, 230 C3.ai company, 100 Campbell, Kurt, 137 “The China Reckoning,” 137 CAOC (Combined Air Operations Center) at Al Udeid, x, 2, 45–54 cost of an inefficient system, 52, 58 DIUx’s apps for, 59 DIUx’s tanker refueling project, 54–59, 65, 68, 102 Hanscom/Grumman tech overhaul and, 53, 59–60, 63, 66–68 obsolete technology at, 46, 47, 50–54, 57 Shah and DIB group visits, 45–46, 49 Slapshot app, 69 Capella Space, ix, 81, 82, 92–98, 151, 180 DIUx/DoD funding and venture capital for, 85–88, 91–98, 129 growth and revenue, 215 Kirchhoff in South Korea and, 89, 91 left-of-launch solution enabled by, 91 opponents of, 87–88 Ukraine War and, 199, 200, 231–32 use of, versus defense contractor, 88 Carter, Ash, x, 7, 21, 32, 65, 66, 77, 111, 145, 172, 179, 185, 190, 221 appoints Kirchhoff and Shah to DIUx, 11, 13, 14, 16 approves Dailey’s CSO idea, 42, 43 at Aspen Strategy Group (2018), 134 death and funeral, 219–21 Defense Digital Service, 231 on defense innovations, 7, 111 DIB created by, 47 DIUx and, 10, 11–13, 17–18, 20, 23, 26, 28, 42, 48, 109, 155, 223, 242 Mattis replaces, 102, 109, 153 overtures to Silicon Valley, 7–9 Shield Capital and, 189 vision of, 9, 99, 132, 216, 228, 229 Cerf, Vint, 35, 120–21 ChatGPT, 103, 239 China, 188 aggression under Xi Jinping, 136 AI and, 133, 138, 141–42, 148, 162–68 American investment in, 245 Belt and Road initiative, 244 Biden’s policies and, 169–70 “China’s Sputnik Moment,” 4 CHIPS Act and, 176 civil-military fusion, 138, 177, 196, 214, 244, 245 Clinton visit (summer 1998), 136 COVID-19 and, 163, 168 Cuba and, 244 displaced as biggest U.S. trading partner, 245 economic problems, 244 global ambitions, 137 “Guo Wang” satellites, 225 as investors in U.S. startups, 149 manufacturing and industrial production, 137 military budget, 217 “national champion” companies, 137 need for technological containment of, 141–42, 156 nuclear-capable hypersonic weapon launched, 4, 223–24 plans to win the technology race, 137 R&D and, 137, 138 Schmidt meets with top AI diplomat, Fu Ying, 162–68 spy balloon over the U.S. (2023), 224 as threat to Taiwan, 216–18, 224, 244 technological superiority, 3, 4, 55, 136–38 Tencent, 177 Trump’s policies and, 139–41 U.S. companies outperforming Chinese companies, 178 U.S. policy turnaround, x, 176, 177 U.S. technology sold to, x, 15 as threat to U.S., 130–31, 136–37, 141, 149, 217, 223–24, 239, 244 “China’s Technology Transfer Strategy” (Brown), 149, 170 CHIPS Act, 130, 161–62, 176, 238 Clark, James “Snake,” 95 Clinton, Bill, ix, 136 Clinton, Hillary, 61, 62 cloud services, v, 5, 120, 148, 202, 239 CNAS (Center for a New American Security), 154 computer chips, 80, 159–62 national security vulnerability, 161 See also CHIPS Act Cook, Tim, 21 Cope, Clint, 182 Council on Foreign Relations, 14, 143, 220 COVID-19, 163, 168, 169, 189 CSO (Commercial Solutions Opening), DIUx’s Dailey’s acquisition process, x, 24, 37–38, 40–43 Cukor, Drew, 119 cyberattacks defense for, 202 Russia attacks on Estonia and the Ukraine, 164–65, 200, 207 cybersecurity, 9, 132, 158, 194, 195, 211 Cybersecurity Commission, 156–57 DIUx and ForAllSecure’s Mayhem, 100 Shah’s startups, 2, 14, 15, 21, 184, 190, 197 Symantec, 149 Tanium company, 99, 132 Dailey, Lauren, x, 24, 37–38, 40–43 Danzig, Richard, 134–35 DARPA, 8, 24, 35, 40, 71, 155 Grand Challenge awards, 8, 100 Kirchhoff and, 14, 35 Mayhem software and, 100 Datahub AI system, 80, 81, 86, 98 DCGS-A (Distributed Common Ground System—Army), 123 Deal, Victor, 42 defense contractors, the “primes,” 5, 53, 59–60, 67, 76, 88, 98, 123, 127, 173, 189, 196, 240 building a new prime, 208 Defense Industry Association conference, 239–40 ousting, by a disrupter, 242 repackaging aging technology, 242 Replicator Initiative and, 240 Ukraine War and, 215 working with startups, 243 Dereliction of Duty (McMaster), 139 DIB (Defense Innovation Board), xi, 46, 47, 154, 155 AI principles and, 122 Kirchhoff in South Korea with, 89–92 recommendations sent to DIU, 49 Schmidt heads, 46, 47, 48–49, 155 tour of CAOC at Al Udeid, 46, 49 Ukraine War and, 212 Disbrow, Lisa, 67 DIU (Defense Innovation Unit, formerly DIUx or Unit X), 2, 202 accomplishments, 71, 99–122, 128, 131–32, 155, 178, 242 adversaries, 26–31, 75, 77, 87–88, 93–98, 129, 153 AI and, x, 100, 150–52, 119, 203, 239 Albright visits, 142–43 allies launching their own DIUs, 245 alumni joining Shield Capital, 190 Ash Carter Exchange meeting, 228–30 Austin and, 223, 225–28, 240 Beck and DIU 3.0, ix, 179, 225–30 Biden administration and, 223 Brown as director, ix–x, 130, 148–55, 171–73 CAOC tanker refueling project, 54–59, 60, 65, 68 Capella and secret North Korea project, 73–98, 129 Carter’s vision, to use Silicon Valley methodology for defense, 9–10, 88 challenges for Shah and Kirchhoff, 35 commercial IT solutions and, 238 contracts awarded in 2017, 99 Dailey’s CSO idea, x, 24, 37–38, 40–43 defense contractors’ antagonism, 59–60 directives for, 22–23 early failure and reboot, 11–13, 20, 32 entrepreneur funding, defense investing, and, 9, 17, 85, 87, 92, 96–99, 101, 103, 107–8, 112, 127, 184, 202 first all-hands meeting, 19, 21–24 flat structure for, 20 formula for a tech-forward DoD, 242–43 “frozen middle” problem, 70 funding and funding problems, 9, 25–44, 71–72, 86–87, 93–99, 108, 110, 154–55, 230 future of warfare and, 129, 199–218, 228, 240, 241–42, 243–44 Griffin and, 130, 154 Hacking for Defense course and, 84 Hanscom/Northrop Grumman software overhaul and, 66–68 hiring directive and O-7 billet, 23 importance of, 92, 102 Kessel Run team, 67–70, 102 Kirchhoff leaves, 129–30 Kirchhoff’s leaked memo, 61–63 “lean” methodology of, 54 Mattis and, 102–4, 108–9, 153–55, 242 mission of, 3, 15, 21–22, 56, 88, 172, 228 office, Moffett Field, Mountain View, 8–9, 10, 12, 17, 35–36, 38, 44, 54–55, 58, 67–68, 92, 101, 104 offices opened, 30–31, 71, 101, 118 OTA contracting system, 40–41, 99, 112, 213, 239 personnel: full-time, reservists, and guardsmen, 32, 101, 243 “Points of Presence,” 30 portfolio areas and companies, 36, 99–128, 131, 132 Project Maven and, 116–22 racing to sign deals, 36–44 removal of the “x,” 109 Replicator Initiative and, 240 reporting and authorities for, 16–17, 22–24, 242 Reserve Unit, ix, 18, 21, 225, 226 Rogue Squadron, 104–10, 210, 211 role in changes to key Operational Plans (O-Plans), focus on war planning, 227–28 Shah and Kirchhoff heading, 2, 11, 15–17, 24 Shah leaves, 130 Slapshot app, 69 Space Portfolio, x, 76, 81, 155 speed of deal-making, 99, 101 supporters, 10, 23, 26, 28, 67–68, 89, 93, 102–3, 108–9, 129, 153–54, 189, 242, 245 team members, 19–21, 23, 37, 44, 55 Ukraine War and, 203–5, 215, 223 U.S. policy shift on foreign investment and, 149 venture capital model for, 20, 36 Warsaw conference on Ukraine technology scaling, 234–36 zeroized crisis, 25–36, 72, 97 See also Kirchhoff, Christopher; Shah, Raj M.; specific projects DJI (Chinese drone company), 103, 104, 107, 137, 209, 211 Rogue Squadron’s hacking and reverse engineering, 106–7 Donovan, Matt, 63–64 drones (UAS, unmanned aircraft system), 114, 237 Afghanistan War and, 105 AI-empowered, 210, 228–29 Air Force purchase of, 228–29 attacks on U.S. service members in the Mideast by, 241 Beall’s Android app to locate enemy operators, 106 Blue UAS program, 108, 131, 235 changing fighter aviation, 114, 237–38, 241 cost of, versus conventional craft, 241 “counter-UAS” solutions (drone-killers), 9, 106, 126, 202, 210, 237, 238, 241 DIU funding of U.S. makers, 107–8 DIU Rogue Squadron, 104–10, 210, 211 DJI Chinese drones, 104–7, 137, 201 in DoD inventory, 103 DoD’s Replicator Initiative, 239–40 Gorgon Stare sensors on, 119 Iranian-made drones, 201–2, 209, 240 kamikaze drones, 201, 210, 232, 235, 237, 240, 241 North Korean use of, 214 quadcopters, 3, 4, 107, 201, 202, 209, 232, 240 Russia and, 240 Shield AI, 100–101, 131 Skydio, 202 terrorists’ using, 240–41 Turkish Bayraktar TB2s, 209 Ukraine War and, 201, 204, 208–12, 232–35, 237 used by ISIS and insurgents, 3–4, 103 used to surveil U.S. military bases and ports, 106 U.S. military vulnerability, 238 weaponizing of hobby drones, 3–4, 103, 106 Duchak, George, 20 Dunford, Joseph, 62, 63, 119 Dunnmon, Jared, x, 150–51 DIU AI technical director, 151, 203 partnership with PMS-408, 152 Ukraine War and, 203, 204–5 Edgesource, 210 Elroy Air, 181–84 End of History and the Last Man, The (Fukuyama), 218, 237 ESL company, 84, 85 Esper, Mark, 154 EUCOM, 205 Evolv Technology, 117–18 eVTOL (electric-powered vertical takeoff and landing aircraft), ix, 3, 114–16, 181–84 aerial cargo options, 182–83 air taxi, 115 “dual-use” product, 115 Elroy Air’s Chaparral, 181–84 Joby Aviation and, 116 military use, 182, 183 F-16 Viper, 1, 4, 14, 34, 237–38 Facebook, 123, 124 FAR (Federal Acquisition Regulations), 38, 39, 40, 43 defense supplement (DFAR), 38 Farris, Ryan, 73, 74, 76–77, 78, 87 Banazadeh and SAR satellites, 86 Capella funding and, 95 Datahub project and, 81 J-39 SWAT team and, 74, 80 Orbital Effects founded, 97 FedEx, 182 Felter, Joe, 179 FitzGerald, Ben, 154 5G telecommunications, 4 Flake, Jeff, 195 Fog of War (film), 147 ForAllSecure, 100, 132 Fort Hunter Liggett, California, 115 Founders Fund, 122, 123, 124 Four Steps to the Epiphany (Blank), 85 Fukuyama, Francis, 218 Fu Ying (Chinese AI diplomat), 162–68 Gallagher, Mike, 156–57, 161, 244 Garg, Avichal, 169 General Dynamics, 1 Giannandrea, John, 118 Ginkgo Bioworks, 174 Giustina, Marissa, 169 Goldfein, David, x, 59–60, 63, 68, 178, 179, 189 Google, xi, 5, 48, 162, 169, 195, 203 AI research, 48, 118 Cloud AI, 239 government grants to, 8 Project Maven and, 100, 116, 119–21, 127–28 rebuffs Secretary of Defense Carter, 7 resumes working with the DoD, 158 Schmidt as CEO, 48, 49, 186 self-driving car project, 20, 48 Taylor recruited from, 20–21, 114 U.S.

Irresistible: How Cuteness Wired our Brains and Conquered the World
by Joshua Paul Dale
Published 15 Dec 2023

Kringelbach et al., ‘How cute things hijack our brains and drive behaviour’, The Conversation (4 July 2016), theconversation.com/how-cute-things-hijack-our-brains-and-drive-behaviour-61942 (accessed 31 October 2022). 47 Ibid. 48 Kringelbach et al., ‘On cuteness’, p. 9. 49 Tyler Colp and Nico Deyo, ‘The Vtuber Industry: Corporatization, Labor, and Kawaii’, Vice (23 December 2020), www.vice.com/en/article/akdj3z/the-vtuber-industry-corporatization-labor-and-kawaii. See also Lu et al., ‘More Kawaii than a Real-Person Live Streamer’. 50 Ethan Gach, ‘AI-Controlled VTuber Streams Games On Twitch, Denies Holocaust: Neuro-sama likes to play Minecraft and go off-script’, Kotaku (6 January 2023), kotaku.com/vtuber-twitch-holocaust-denial-minecraft-ai-chatgpt-1849960527 (accessed 23 January 2023). 51 R. O. Kwon, ‘Stop Calling Asian Women Adorable’, The New York Times (23 March 2019), www.nytimes.com/2019/03/23/opinion/sunday/calling-asian-women-adorable.html. 52 Noriko Murai, ‘The Genealogy of Kawaii’, in Noriko Murai et al., eds, Japan in the Heisei Era (1989–2019), (London: Routledge, 2022), p. 249. 53 Simon May, The Power of Cute (Princeton, NJ: Princeton University Press, 2019), p. 127. 54 Leila Madge, ‘Capitalizing on “Cuteness”: The Aesthetics of Social Relations in a New Postwar Japanese Order’, Japanstudien 9, 1 (1998), p. 167. 55 Ibid., p. 164. 56 Slade, ‘Cute men in contemporary Japan’, p. 79.

…

The Treasure Box All the Girls Want’, in Masami Toku, ed., International Perspectives on Shōjo and Shōjo Manga: The Influence of Girl Culture (London: Routledge Press, 2018). Gach, Ethan, ‘AI-Controlled VTuber Streams Games On Twitch, Denies Holocaust’, Kotaku (6 January 2023), kotaku.com/vtuber-twitch-holocaust-denial-minecraft-ai-chatgpt-1849960527 (accessed 23 January 2023). Galbraith, Patrick W., ‘Seeking an alternative: “male” shōjo fans since the 1970s’, Shōjo Across Media: Exploring ‘Girl’ Practices in Contemporary Japan, ed. Jaqueline Bernt, Kazumi Nagaike and Fusami Ogi (London: Palgrave Macmillan, 2019). Galbraith, Patrick W., Otaku and the Struggle for Imagination in Japan (Durham, NC: Duke University Press, 2019).

Shocks, Crises, and False Alarms: How to Assess True Macroeconomic Risk
by Philipp Carlsson-Szlezak and Paul Swartz
Published 8 Jul 2024

For that reason, we argue that the productivity boom expected from the adoption of generative AI is more likely to deliver 25 to 50 bps than 150 bps of additional trend growth. Should that be considered a disappointment, or looked forward to as meaningful impact? Some will see this as pessimistic and wonder if we have spent enough time playing with ChatGPT and all its generative-AI siblings. We assure readers that we have and are duly impressed. But we continue to believe in the hurdles of technological maturity (we’ve all seen the AI hallucinations), of societal resistance (not in my daughter’s classroom), of regulatory friction (What will the new rules be?)

…

See also technology bad macro, 108, 137, 243, 249, 253, 255 banking systems, 249 bubbles and, 188 financial recessions and, 40, 53 global financial crisis and, 26 recoveries and, 49–50 bankruptcies capital allocation and, 167 private, 176 recessions and, 48 Belt and Road Initiative, 204 Bernanke, Ben, 123 biases confirmation, 200, 210 toward doom and gloom, 10 Biden, Joe, 115, 132, 133–134, 224 bond markets, 108, 123–124, 124f bond vigilantes and, 108, 132–133, 135 bubbles, 182 inflation expectations and, 140–141 quantitative easing and, 113, 113f Brandt, Willy, 224 Brazil, 231, 238 Bretton Woods Conference, 235 Brexit, 219, 220–221, 227 BRICS, 231, 238 bubbles, 29f challenges from, 181–182 characteristics of and risk from, 183–185 cleanup of and financial stability, 187–188 cycle-ending, 184f, 185 future of, 249–250 good, 188–190, 191 idiosyncratic, 184–185, 184f loving and living with, 180–191 spotting, stopping, and stoking, 185–187, 191 structural break, 184f, 185 taxonomy of, 183–185, 184f ubiquity of, 182–183 Bush, George W., 110, 117 Canada structural damage in, 49–50, 49f V-shaped recovery in, 46–47, 47f capital allocation, 167–168, 237, 244, 247–248 capital dumping, 237–238 capital flows, 235–236, 251 capital growth, 57–59, 60–63 in China, 75–76 depreciation and, 61–63, 62f, 63f, 69 gravity of growth and, 68–70, 69f recessions, recoveries, and, 48–49, 48f capital markets, 199, 202 carbon capture, 62–63 CARES Act, 114, 114f, 115 Carter, Jimmy, 108, 110, 124, 133 Carville, James, 132–133 ChatGPT, 98 Cheney, Dick, 110, 133, 174 China economic convergence and, 202 financial divergence and, 204 gravity of growth in, 68, 75–76, 75f magical growth models on, 68, 74–76, 75f trade and, 224–225 Citigroup, 190 climate change, 13 decarbonization and, 62–63, 154, 175, 245 green investment and, 189 investments and, 245 Clinton, Bill, 110, 132–133 Cold War, 197–201, 224–225 collective-action problems, 189–190 convergence, 14.

Boom: Bubbles and the End of Stagnation
by Byrne Hobart and Tobias Huber
Published 29 Oct 2024

More frequent replications, experiment pre-registrations and the publication of failures and negative results, and the abandonment of the flawed p<0.05 significance test, could also improve the process of scientific experimentation and discovery. AI provides another source of hope for overcoming scientific stagnation. Literature-based discovery, or LBD, which uses ChatGPT-style models to analyze massive amounts of scientific data and literature, promises to unlock novel discoveries. In fact, it already has, for example in the case of the antibiotics halicin and abaucin, which two MIT teams identified in 2019 using a machine-learning algorithm trained on the chemical structures of thousands of known antibiotics.

…

As discussed in Chapter 1, the abundance of money in big tech over the last decade has coincided with a scarcity of ideas, as indicated by the billions of dollars the large tech firms have been amassing. But beyond antitrust law and regulation, recent technological advances like the theoretical breakthrough of transformer-based AI and the user-facing advancements represented by OpenAI’s ChatGPT and other large language model-based tools might be the most significant accelerants for another corporate R&D bubble. While they have been in the making for decades, the recent breakthroughs in generative AI achieved by upstarts like OpenAI, Stability AI, and Anthropic have disrupted big tech’s R&D complacency for now.

Your Face Belongs to Us: A Secretive Startup's Quest to End Privacy as We Know It
by Kashmir Hill
Published 19 Sep 2023

They made speech recognition better, image recognition more reliable, and facial recognition more accurate. Neural networks would be employed for all manner of tasks: recommending shows to watch on Netflix, populating playlists on Spotify, providing eyes to the autopilot in Tesla’s electric cars, and allowing ChatGPT to converse in a seemingly human way. Anywhere a lot of data exists, a neural network can theoretically crunch it. All of a sudden, that small cluster of renegade academics became the hottest commodity in technology. Google, Microsoft, Facebook, Baidu: All the biggest technology companies in the world threw money at them.

…

See also photographs/photography Clearview AI Camera, 111 Eastman Kodak film camera, viii Insight Camera, 187 skin tone and, 179 speed cameras, 238 True Depth camera, 109 Capitol insurrection, 228–230 Carlo, Silkie, 215–217, 218, 220 Carlson, Tucker, 11 Cato Unbound, 261n15 Catsimatidis, Andrea, 114 Catsimatidis, John, 114 cease-and-desist letters, 165 census bureaus, 24–25 Central Intelligence Agency (CIA), 38, 125, 227, 268–269n38 Ceph, 79 Cernovich, Mike, 12, 53 Chaos Computer Club, 190 ChatGPT, 73 Chau, Ed, 186 Chicago Police Department, 155 child crimes investigators, 134–136 child sexual abuse material (CSAM), 135. See also pornography China scoring of citizens by, 34 surveillance in, 224–227 Churchill, Winston, 41 CIA (Central Intelligence Agency), 38, 125, 227, 268–269n38 Citizens United, 208, 306n206 Clarifai, 32–33 Clarium Capital, 14 ClassPass, 8 Clearview AI absence of headquarters for, x–xi, xii access requests and, 190–192 ACLU and, 202, 203–204, 205 AR glasses and, 249–250 attempts to shut down communication by, xv, xvii–xviii author’s contact with representatives of, 160–166 bans on, 165 capabilities of, ix–x Capitol insurrection and, 229–230 concerns regarding weaponization and, xv–xvii defense of, 157–159 effectiveness of, xii emergence of, 94 fines assessed to, 193–194, 230 growth of database for, 246–247 hit rate of, 133 international backlash against, 192–193 international expansion and, 137 investment efforts and, 111–120, 136. see also individual investors Johnson and, 95–96 law enforcement and, 128, 130–139 lawsuits against, 165, 204–206, 209–213, 248 Leone and, 113–114 NYT article on, 164–165, 187, 190, 194, 204, 230 opposition to, 237–238 pandemic and, 186–187 police reactions to, xii–xiv potential legal challenges to, 117 push for ethical use and, 239 results blocked by, xvii–xviii, 162–163 Scalzo and, 111–112, 113, 188–189 third-party testing of, 240 tip regarding, vii–viii, ix Ukraine invasion and, 237 wrongful arrests and, 183 Clearview AI Camera, 111 Clearview AI Check-In, 111–112 Clearview AI Search, 111 Clement, Paul attempts to contact, 160 on Clearview AI’s capabilities, ix–x lack of response from, xi legal memos from, 134, 157–158 Clinton, Bill, 209, 259n10 Clinton, Chelsea, 9–10, 259n10 Clinton, Hillary “deplorables” comment by, 50–51 election loss of, 88 facial recognition and, 104 false claims about, 94 Trump and, 16, 52 Cohen, Chuck, 133 Colatosti, Tom, 62, 65, 66, 71 Comet Ping Pong, 55 Computer Fraud and Abuse Act (1986), 117–118 computers early, 36 reliance on, 36 confirmation bias, 181 Constitution First Amendment, 15, 206–207, 208–209, 212–213, 306n206 Fourth Amendment, 141 consumer protection laws, 205 contact tracing, 186 Coolidge, Calvin, 11 Couchsurfing, 81 Coulson, Jennifer, 180–181 Coulter, Ann, 119 Covid-19, 185–187, 209, 214 Crime and the Man (Hooton), 25, 26 CrimeDex, xii, 134 criminal detectors, 31 “criminal face,” 31 Criminal Justice Information Services Division, 299n158 Criminal Man (Lombroso), 22–23, 38 crisis communications, 161 Cruise, Tom, 123 Crunchbase, 79 Cruz, Ted, 11 cryptocurrency, 81 CSAM (child sexual abuse material), 135.

Uncomfortably Off: Why the Top 10% of Earners Should Care About Inequality
by Marcos González Hernando and Gerry Mitchell
Published 23 May 2023

Mishra, C. and Rath, N. (2020) Social solidarity during a pandemic: Through and beyond Durkheimian lens. Social Sciences and Humanities Open, 2:1, 1–7. Mitchell, G. (2022) The clapping might have stopped, but our need for care is not going away. LabourList. 12 July. https:// labourlist.org/2022/07/the-clapping-might-have-stopped-butour-need-for-care-is-not-going-away Mok, A. and Zinkula, J. (2023) ChatGPT may be coming for our jobs. Here are the 10 roles that AI is most likely to replace. Business Insider. 2 February. www.businessinsider.com/chatgptjobs-at-risk-replacement-artificial-intelligence-ai-labor-trends2023-02?r=US&IR=T Monbiot, G. (2022) Putin exploits the lie machine but didn’t invent it.

Nobody's Fool: Why We Get Taken in and What We Can Do About It
by Daniel Simons and Christopher Chabris
Published 10 Jul 2023

Their expertise in developing sophisticated computational models is genuine, but it is not the expertise necessary to evaluate whether a model’s output constitutes generally intelligent behavior. People who make these predictions appear to be swayed by the most impressive examples of how well new machine learning models like ChatGPT and DALL-E do in producing realistic language and generating beautiful pictures. But these systems tend to work best only when given just the right prompts, and their boosters downplay or ignore the cases where similar prompts make them fail miserably. What seems like intelligent conversation often turns out to be a bull session with a bot whose cleverness comes from ingesting huge volumes of text and responding by accessing the statistically most relevant stuff in its dataset.

Filterworld: How Algorithms Flattened Culture
by Kyle Chayka
Published 15 Jan 2024

It requires the material to be translated into data that the machine can understand, such as text; it lacks serendipity because it can filter only by the terms that the user inputs; and it does not measure inherent quality. It is unable to “distinguish a well written [and] a badly written article if the two articles use the same terms.” The inability to evaluate quality brings to mind artificial intelligence: New tools like ChatGPT seem to be able to understand and generate meaningful language, but really, they only repeat patterns inherent in the preexisting data they are trained on. Quality is subjective; data alone, in the absence of human judgment, can go only so far in gauging it. Social information filtering bypasses those problems because it is instead driven by the actions of human users, who evaluate content on their own—using judgments both quantitative and qualitative.

Doppelganger: A Trip Into the Mirror World
by Naomi Klein
Published 11 Sep 2023

And now that the machines have devoured so much of us, gorged on so many of our ways and our quirks, they can make rather credible replicas of us near instantly. My friends who are visual artists and songwriters are terrified about what their futures hold when artificial intelligence programs can be instructed to make art “in the style of” them—and then churn out passable replicas within moments. Nick Cave, when confronted with a ChatGPT-generated version of a Nick Cave song, described the phenomenon as “replication as travesty … a grotesque mockery of what it is to be human.” There is something uniquely humiliating about confronting a bad replica of one’s self—and something utterly harrowing about confronting a good one. Both carry the unmistakable shudder of the doppelganger.

Blood in the Machine: The Origins of the Rebellion Against Big Tech
by Brian Merchant
Published 25 Sep 2023

And the boss is the algorithm; HR consists of a text box that workers can log complaints into, and which may or may not generate a response. The modern worker can sense the implications of this trend. It’s not just ride-hailing either—AI image-generators like DALL-E and neural net–based writing tools like ChatGPT threaten the livelihoods of illustrators, graphic designers, copywriters, and editorial assistants. Streaming platforms like Spotify have already radically degraded wages for musicians, who lost album sales as an income stream years ago. Much about Andrew Yang may be suspect, but he did predict correctly that anger would again spread like wildfire as skilled workers watched algorithms, AI, and tech platforms erode their earnings and status.

Deep Utopia: Life and Meaning in a Solved World
by Nick Bostrom
Published 26 Mar 2024

G., 30 Wireheading, 226, 419 Wolf, Susan, 460–63 Y Yudkowsky, Eliezer, 221, 225, 240, 502 ACKNOWLEDGMENTS I am grateful to many more people than I can mention for enabling this work, by creating the material or intellectual preconditions for its coming into existence. But for their more direct contributions, I want to explicitly thank Guive Assadi, Emily Campbell, Richard Yetter Chappell, Will Hammond, Guy Kahane, Anton Korinek, Matthew van der Merwe, Thaddeus Metz, Geoffrey Miller, Toby Newberry, Carl Shulman, ChatGPT-4, Claude 2, Tanya Singh, and Jan-Erik Strasser for extensive comments on the manuscript; Gilbert N. Morris, Eric Neyman, Toby Ord, David Pearce, and Anders Sandberg for helping to answer some particular questions; Liz Hudson, Frances Key Phillips, Sam Blake, and Pascal Porcheron for help with copyediting; Gwen Bradford, Stephen Campbell, Dale Dorsey, Nick Fletcher, Thomas Hurka, Antti Kauppinen, Eden Lin, Michael Prinzing, Aaron Smuts, Michael Steger, Louise Sundararajan, Lars Svendsen, Valerie Tiberius, and Susan Wolf for in-depth expert conversations on specific topics that helped me learn and refine preliminary ideas; Matthew van der Merwe and Toby Newberry for project management; Guive Assadi for research assistance; and the team at Ideapress for their uncommon dispatch.

Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity
by Daron Acemoglu and Simon Johnson
Published 15 May 2023

From the Field of AI Dreams People are right to be excited about advances in digital technologies. New machine capabilities can massively expand the things we do and can transform many aspects of our lives for the better. And there have also been tremendous advances. For example, the Generative Pre-trained Transformer 3 (GPT-3), released in 2020 by OpenAI, and ChatGPT released in 2022 by the same company, are natural-language processing systems with remarkable capabilities. Already trained and optimized on massive amounts of text data from the internet, these programs can generate almost human-like articles, including poetry; communicate in typical human language; and, most impressively, turn natural-language instructions into computer code.