Timeline of the AGI Movement
An ideological history that underpins today's AI development
To even make sense of what is happening in AI, it helps to understand how we got here. In this post, we present an ideological history of AI over the last 25 years, which–at least in Western AI labs and policy circles–has centered on the development of AGI.
The most important week in AI
There’s a running joke in the AI world that every week is “The most important week in AI!” – and people have been saying it for five years straight.
But the news in the past few weeks has been a lot. Since mid-January, we’ve seen AI agents form their own religions on Moltbook; developers flock to powerful new AI coding tool Claude Code (while the CEO of the same company warns about existential risks from the same technology); xAI undergo the largest merger ever with SpaceX at a $1.25T valuation (allegedly to build datacenters in space); and prominent D.C. policy analyst Dean Ball write that AI “recursive self-improvement” is on the horizon.
Here’s Dean Ball:
There is one assumption I’ll ask you to make with me, which is that substantial automation of AI research is a near-term possibility. This requires believing a few things. First, that AI research and engineering is substantively composed of work like: finding optimizations in various complex software systems; designing and testing experiments for AI model training and posttraining; and creating software interfaces to expose AI model capabilities to users. Second, that a great deal of this work is essentially reducible to the engineering of software. Third, that AI models, while not yet geniuses, are reaching quite high levels of human competence. Fourth, that frontier lab leadership and staff are serious when they describe AI research automation as a near-term goal, and that frontier lab research staff are telling the truth when they say that AI is already writing a large fraction of their code.
The tech world has been shocked in the last few weeks by how useful AI has become at coding. But what Dean Ball is writing about is something far more serious: if AI technology can automate AI research and development, it could mean an exponential acceleration in AI capabilities. We would be on the cusp of building what’s often called Artificial General Intelligence (AGI) - essentially the holy grail of AI development.
What is AGI?
Artificial General Intelligence (“AGI”) is conceived as a powerful AI that can match or exceed human capabilities on any task. This type of general purpose AI is often contrasted with “narrow artificial intelligence” or AI systems that are extraordinarily good at just one thing–for example, playing chess or go or coming up with novel ways of folding proteins. Popularly illustrated by Anthropic CEO Dario Amodei as a “country of geniuses in a datacenter,” AGI could have the power to rival not just individuals, but entire companies or even nation-states.
Whether or not it is possible to build AGI–and if so, how soon it may arrive–is highly contentious. Over the last 10 years, however, informed estimates about the timeline to AGI have dramatically compressed from a century or more to a decade or less. It is also debated whether or not AGI would be dangerous, with some warning that it would lead to human extinction, others writing that it will lead to abundance and human flourishing, and still others speculating that it will help the U.S. (or China) secure the geopolitical upper-hand in “the race to achieve global dominance in AI.”
In just a few weeks (March 27!), a feature film “The AI Doc” will be released sharing perspectives of many of the people working directly in this field. While the film is open-ended, there is a clear consensus that the “default path” we are headed down leads to danger.
What does this have to do with Buddhism?
In this post, we present an ideological history of AI over the last 25 years, which–at least in Western AI labs and policy circles–has centered on the development of AGI.
We do not directly offer Buddhist angles on AI below. But as we shall see, tracking the ideological underpinnings of the current AI movement makes a strong case for why Buddhists should be paying attention to AI and how it is being developed and deployed.
As AI shapes and reshapes the world, wisdom traditions like Buddhism and other systems of moral authority are in strong positions to clarify and critically engage the underlying philosophical tenets of the current movement, point out their flaws, and help course-correct AI for the better. In Buddhist terms, this is a matter of identifying “wrong view” and offering wiser alternatives.
The Buddhism and AI Initiative will soon begin offering education courses and workshops to Buddhist Sanghas and Teachers interested in learning more about AI, how Buddhism can help shape AI for the better, and ways to get directly involved. If you or your sangha is interested in connecting about this, please reach out to hello@engagedbuddhists.ai.
Timeline of the AGI Movement
Fortunately, even those who disagree about today’s AI situation agree on the ideological history behind the AGI movement. The history below draws on three resources: (1) the TESCREAL Bundle paper which tends to take the view that AGI is largely a fantasy used to justify capitalist ends, (2) The Compendium, which takes the view that AGI is possible in the near future and a genuine risk to human civilization, and (3) the meticulous archiving of Issa Rice, an independent researcher who has compiled events and dates related to the history of AI safety and various AI companies.
All three sources agree on the following high-level history, which we chart in more detail below:
1950 - 1960: Birth of the AI field
1960 - 1990: Origin of ideas around AGI through a number of sci-fi stories
1990 - 2000: AGI as a niche interest; a few individuals take AGI seriously
2000 - 2010: AGI as a research interest; first institutions formed to build AGI
2015 - 2023: AGI as competitive interest; multiple labs now pursuing AGI
2023 - 2024: AGI as a commercial interest; explosion of interest post ChatGPT
2024 - 2025: AGI as a political interest; founding of AI safety inst., regulation
2025 - now: AGI as a geopolitical interest; “live players” includes labs and T1 governments.
Taking this view deemphasizes the scientific and technical history of AI, as well as the academic and activist history of AI Safety and Ethics, in favor of drawing attention to one of the more prevalent ideological streams driving the Western AI movement.1
The birth of AI (1950s)
The field of AI kicked off in the 1950’s with an attempt to create “human-like intelligence.” Rather than AGI being a completely new idea, the seeds of as-smart-as or smarter-than-human AI has been with computer scientists from the beginning.
In 1950, Alan Turing published “Computing Machinery and Intelligence,” proposing the “Turing Test” as a way to evaluate machine intelligence, and Norbert Weiner published “The Human Use of Human Beings,” which was the first attempt to ground scenarios of an automated AI future in societal considerations. Six years later, the 1956 Dartmouth Summer Research Project established the field of AI, bringing together experts with a bold premise: “Every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it.”
Science fiction plants the seeds (1960s–1990)
What followed was decades of slow scientific and technical progress, but rapid imaginative progress. Late-20th-century science fiction explored concepts of AGI, technological supremacy, and human extinction, planting ideas that would later be taken literally by the people spearheading the industry-scale building of AI.
In 1965, I.J. Good coined the term “intelligence explosion,” arguing that a sufficiently intelligent machine could design even better machines, triggering a recursive loop: “The first ultraintelligent machine is the last invention that man need ever make.” (Compare this idea with that of Dean Ball’s above!). Through the 1970s and 80s, stories like 2001: A Space Odyssey, Terminator, and Vernor Vinge’s A Fire Upon the Deep depicted AGI explicitly as a civilizational turning point–for better or worse.
A niche subculture takes it seriously (1990s)
Then something shifted. The sci-fi concepts of Verner Vinge and others were picked up by a small group of hobbyists who took them seriously, contemplating for the first time: “what if the ‘singularity’ - or the arrival of AGI - was closer than we think?” This marks a turning point from fiction to theory, with many leading intellectuals and builders of AGI companies descending from these circles.
In 1990, the Extropy Institute was founded, promoting ideals of life extension and transhumanism2; it’s mailing list included many who would become leading AI researchers and thinkers, include Eliezer Yudkowsky, Nick Bostrom, Marvin Minskey, Eric Drexler, and others. That same year, Ray Kurzweil published “The Age of Intelligent Machines,” predicting that AGI was possible and would have huge impacts on work, education, medicine, and warfare. In 1993 Vinge published “The Coming Technological Singularity,” arguing superintelligence–an even more powerful form of AI than AGI–may arrive within decades.
And in 1997, Yudkowsky launched the SL4 mailing list, one of the first online hubs for “singularity” discussions. Future AGI company founders participated, including Shane Legg of DeepMind. From Yudkowsky’s “Future Shock Levels” post: “When I first ran across the idea of the Singularity I knew immediately that Vernor Vinge was perfectly right; I felt my entire ethical system restructuring over the course of about five seconds.”
From mailing lists to institutions (2000–2010)
Over the next decade, the people in these early discussion groups formed, for the first time, institutions dedicated to building or protecting against AGI. In 2000, Yudkowsky founded the Singularity Institute for Artificial Intelligence (SIAI, later renamed MIRI), writing about how to build “Friendly AI” and later “The Sequences” which became a foundation for concerns that AGI poses existential risk. Today, Yudkowsky’s views have shifted even more pessimistic and his latest 2025 book “If Anyone Builds It, Everyone Dies” is a NYT bestseller.
In 2005, Nick Bostrom founded Future of Humanity Institute (FHI) at Oxford, and would later write the influential book “Superintelligence” in 2014. Then came the pivotal moment: in 2010, DeepMind was founded by Demis Hassabis, Shane Legg, and Mustafa Suleyman–the first company explicitly built to create “general-purpose AI.” Funding came from Peter Thiel and Elon Musk, via connections to Yudkowsky’s network. From Legg’s PhD thesis: “If our intelligence were to be significantly surpassed, it is difficult to imagine what the consequences of this might be. It would certainly be a source of enormous power, and with enormous power comes enormous responsibility.”
The race begins (2010–2023)
What DeepMind was doing worked. Their remarkable research successes–teaching AI to master early video games, and later the sophisticated game of Go–led other researchers, companies, and countries to take notice. Driven by further progress on neural networks and large language models, and increasing tension between Silicon Valley leaders, multiple elite “Tier 1” labs were formed, each with the explicit purpose of building AGI, by founders connected to transhumanist values.
In 2014, Google acquired DeepMind for ~$650M after numerous research advances. Today, the Google DeepMind team states “AI — and ultimately artificial general intelligence — has the potential to drive one of the greatest transformations in history.”
In 2015, OpenAI was founded by Sam Altman and Elon Musk to create AGI. Musk was driven partially by losing his stake in Deepmind after the Google sale, and grew further wary of Google after a conflict with Larry Page over the future of AGI. Altman was a subscriber to transhumanist ideas, writing about the risk of “superhuman machine intelligence” in 2015. This is the origin of commercial/industrial AGI competition, the belief that whoever controls AGI controls the future. From early emails between Musk and Altman (Musk resigns in 2018):
In 2016, China entered the race with State Council releasing the New Generation AI Development Plan, setting the goal of becoming the world’s leading AI power by 2030. This followed from DeepMind’s AlphaGo defeating world Go champion Lee Sedol, an event watched by over 280 million people in China which Chinese government insiders later described this as a “Sputnik moment” for AI.
In 2021, Anthropic was founded by ex-OpenAI researchers (Amodei et al.), who left over safety concerns and a desire to outcompete OpenAI to build AGI. Amodei and others have long been part of rationalist and effective altruist circles, which share an intellectual history with SL4 and Yudkowsky’s early writings. This was now the third major company competing for AGI.
AI goes mainstream (2023–2024)
The release of ChatGPT in November 2022 launched AI into the mainstream, with 100m users trying it in the first two months—the fastest growing app ever. Private investment flooded in and competition skyrocketed. Training state-of-the-art LLMs now required billions of dollars, meaning AGI labs had to partner with Big Tech for financial and technical support. This happened without exception, and while the AGI ideology became buried in a mix of commercial incentives, the tech companies were still led by and filled with true believers.
The deals came fast. In January 2023, OpenAI received $10B from Microsoft (total company funding today is near $60B, with a valuation over $830B being discussed for future rounds, with Microsoft partnership deepening). In March 2023, Elon Musk founded xAI stating its goal is to build “AGI with the purpose of understanding the universe.” In September 2023, Anthropic set up a $1-4B partnership with AWS (total company funding is soon projected to be ~$50B at a valuation of $350B). In January 2024, Meta released its largest open-weight AI models, with CEO Mark Zuckerberg announcing that AGI was the company’s top priority. In April 2024, Deepmind and Google’s research teams consolidated into single AI lab. And in June 2024, Ilya Sutskever–ex Chief Scientist at OpenAI–launched Safe Superintelligence Inc. raising $1B with no commercial goal, explicitly to train superhuman AI.
Governments take notice (2024–2025)
What had been a commercial matter now became increasingly political. Governments began to recognize the potential economic, social, and defense implications of AI. Countries formed AI safety institutes and reoriented national strategy. Underneath this, the story of AGI gained further weight, with leading labs pairing up with governments and urging them towards action–framing the future as holding “all or nothing” tradeoffs.
In July 2023, hundreds of leaders, including leaders of all major AI labs, signed the CAIS statement: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” In October 2023, Biden signed an AI Executive Order framing AI as both opportunity and national-security risk. The following month, the Bletchley Declaration was signed at the UK AI Safety Summit–the first multilateral statement that frontier AI poses “potentially catastrophic” risks requiring international coordination. Between 2023 and 2025, the UK, US, EU, Japan, Singapore, Canada, Australia, France, South Korea, and Kenya all set up AI Safety institutes. The EU AI Act was published in July 2024, with heavy last-minute edits reflecting how fast the technology was moving.
During this period, AGI lab founders wrote a series of “manifestos” that offer prophecies of different AI futures. OpenAI published “Planning for AGI and Beyond” (Feb 2023); Marc Andreesan’s “The Techno-Optimist Manifesto” (Oct 2023); Anthropic CEO Amodei’s “Machines of Loving Grace,” (Oct 2024); Sam Altman’s “Who Will Control the AI Future” (July 2024): “The rapid progress being made on artificial intelligence means that we face a strategic choice about what kind of world we are going to live in: Will it be one in which the United States and allied nations advance a global AI that spreads the technology’s benefits and opens access to it, or an authoritarian one, in which nations or movements that don’t share our values use AI to cement and expand their power? There is no third option — and it’s time to decide which path to take.” Many of these manifestos frame an “entente” strategy, suggesting that the US needs to outcompete China to develop superhuman AGI first. By late 2024 governments were funding heavy infrastructure to support AGI labs and blocking chip exports to China in progressively more stringent regulation.3
AGI becomes geopolitics (2025–now)
Today, AI is seen as a decisive factor in the future of nations. A year ago at Davos 2025, AI was the dominant conversation–perhaps only overshadowed by the Russia-Ukraine war (and this year by US foreign policy). At private events, heads of state discussed AI strategy, with a general air of “how can our small nation maintain sovereignty when the US controls all the tech?”
In January 2025, the release of China’s DeepSeek model triggered a global AI-stock selloff, with the NASDAQ dropping 3%—a flash crash that demonstrated just how sensitive financial markets had become to the AI race. The Trump Administration pivoted from Biden’s focus on safety to a race frame with its January 2025 Executive Order “Removing Barriers to American Leadership in AI,” followed by its AI Action Plan in July 2025, whose first sentence reads: “The United States is in a race to achieve global dominance in artificial intelligence (AI).”
Lab leaders normalized “superintelligence” rhetoric, with Meta announcing “Personal Superintelligence“ and Altman writing in “The Gentle Singularity“: “We are past the event horizon; the takeoff has started.” Military partnerships deepened, with Anthropic and OpenAI winning defense contracts, and AI defense companies Anduril and Palantir skyrocketing in value.
Today’s AI Situation
The above narrative poses the AGI and superintelligence ideology as central to the history of the modern AI movement, and indeed, it is this view of a “winner takes all [by building AGI]” scenario that is driving today’s government policy decisions around AI. AGI has gone from a niche interest to a geopolitical priority.
The companies in the focus have consolidated globally:
Google DeepMind → US
OpenAI (Microsoft, Apple) → US
Anthropic (AWS, Google) → US
xAI (Elon Musk) → US
Meta (Mark Zuckerberg) → US
Safe Superintelligence Inc. → Israel
Multiple Chinese labs, many with state support → China4
At these labs, discussion is on autonomous AI agents–AI software that can complete tasks over longer periods of time without human oversight. Returning to Dean Ball’s comments on recursive self-improvement, of particular interest and concern are AI agents that can code at or beyond human proficiency. If these agents are capable of writing machine learning code and automating the process of AI research itself, some believe an “exponential take off” of AI progress is possible, far beyond what we have seen today.
Making sense of it all
There are other stories that can be told about the history of AI–such as capabilities advancements over the years, or the dedicated efforts of AI ethicists5, or even comments that the AI world is a “bubble about to burst”–but the story of AGI is one that helps best paint the ideological underpinnings of today’s efforts.
The belief in superintelligence, the ideology of beyond-human capabilities, is at the heart of the AI race, and it’s not hard to see why. It seductively promises technology capable of “solving civilizational problems” or even “ending death.” For those who zealously hold this belief, pursuing AGI is all-consuming, especially as the goal feels increasingly within reach. And every day, it sucks in more of the world’s attention.
In fact, we can view the AGI movement’s progression as the steady expansion of a meme (in the Richard Dawkins sense of an idea that has its own life) that is claiming an ever-larger share of global attention and resources in service of its goal. Yes, there are commercial interests that do not have any “vision” for where this all goes. But those leading the labs–and an increasing number of believers from the world’s most powerful politicians (see Xi and Putin on immortality) to technologists–believe in AGI and its transhumanist implications (or at least its ability to shift the global balance of power permanently). Trillions of dollars, state-backed infrastructure, and the collective attention of some of the world’s brightest scientists and coders are now being deployed to build AGI.
This leads to an assessment of the present situation:
The AGI ideology is a core driver of today’s efforts. The creation of AGI is an act of hyperstition - a belief that makes itself true - as more people believe in it and then work to create it.
Believing in AGI leads to a worldview where most human concerns take a backseat - discrimination, climate, even global peace - in a race to global dominion.
Many powerful leaders and technologists now argue for viewing the world in this way, and are using it to justify their actions. Whether some truly believe it vs. are using it as a narrative tool for political motives is another question, though its origin in a niche subculture and early computer science research suggest this ideology was not initially power-or-profit-motive-in-disguise.
This assessment helps illustrate our previous description of AI as a karmic accelerator. Ideologies are based on value judgments: determinations of what matters and how to achieve pre-determined societal goals. These value judgments are what shape our activities, which then impact the world and feed back into our decision-making.
When we look at the karmic orientations of the AGI quest, and its underlying transhumanist or global-dominance values, we see a vision that runs counter to many Buddhist ideals. Even the “world of superabundance” that AGI proponents hold up in which you can always get whatever you want, whenever you want it, karmically generates a loop of deepening dissatisfaction. When we consider the darker vision of dominance, the karmic implications are graver.
A wise response is difficult.
The collective belief in this story, and the rapid technological progress, suggests that increasingly advanced AI (and potentially AGI) is around the corner, as humanity creates that which it believes in. And yet we must work against this ideological spiral: a narrative that leads us deeper into geopolitical collapse, enabling destructive means to justify a power-hungry end, siphoning attention away from the harms of today.
And yet there is a silver lining: many people are aware that there is something deeply amiss here - with AI, with the ideology behind it, and with the world today. This pressure building in the system, our collective sense of impending collapse, can become fuel for compassionate action.
It is in this spirit that everyone working on AI safety, and Buddhists alike, can find common ground to approach AI through a broad frame of positive change. The risks of AI are not reasons to turn away, but reasons it is paramount to get involved and help steer this trajectory towards more wholesome ends. The next chapter in the history of AI is being written as we speak.
It also largely ignores the intersections between AI research and the Cold War competition between ostensibly capitalist-democratic and socialist-autocratic ideologies, and the deep roots of AI in the military-industrial and later military-industrial-communications complex. But there are themes of these two stories that show up as we begin to talk more about geopolitical competition below.
Transhumanism is a philosophical movement that believes limitations in the human condition should be overcome with technology.
Buddhism & AI co-founder Peter Hershock has written more about some of the Chinese side to this story here.
China positions itself less strongly vis-a-vis AGI. Instead, Chinese development is characterized by massive open-source efforts and top-down state level control aimed at GDP growth. Nonetheless, they are still heavily swept up in the race to build increasingly powerful AI and secure a geopolitical advantage.
This group traces a history through Latanya Sweeney’s Discrimination in Online Ad Delivery (2013), Cathy O’Neil’s Weapons of Math Destruction (2016), Safiya Noble’s Algorithms of Oppression (2018), and other works, many of which demonstrate how algorithmic bias in AI entrench unfairness and discrimination, particularly against minorities. Much of this group’s view on modern AI can be characterized by Emily Bender’s On the Dangers of Stochastic Parrots (2021), which argues “Contrary to how it may seem when we observe its output, an [LLM] is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning: a stochastic parrot.”
Great piece. You may be interested in my aligned work at https://technologicalmetamodernism.substack.com/
Nice to find this Substack. It's like a Buddhist complement to my idealist / Kashmir Shaivist approach to AI!