# Artificial Intelligence Timeline A comprehensive chronological reference tracing artificial intelligence from its theoretical origins in wartime mathematics through the present-day convergence of Big Tech, the [[Department of Defense]], and the [[Surveillance State]]. Every era follows the same pattern: military money seeds the research, corporate power captures the results, and the public arrives last — as consumers, as training data, and as labor to be displaced. ## Core Thesis AI is not a neutral technology that emerged from garages and dorm rooms. It is a product of [[Military-Industrial Complex]] funding, corporate consolidation, and geopolitical competition. Every major breakthrough — from neural networks to autonomous weapons — traces back to defense dollars. The companies that now control AI are worth hundreds of billions and deeply entangled with the Pentagon. The "safety vs. acceleration" debate is largely theater: the accelerationists won, the regulators lost, and the defense establishment got what it wanted. Understanding this timeline means understanding that AI's future was shaped by its past — and its past is the [[Military-Industrial Complex]]. → For the defense-tech thread, see [[Defense Tech Timeline]] · [[DARPA Timeline]] · [[In-Q-Tel Timeline]] → For individual company arcs, see [[OpenAI Timeline]] · [[Anthropic Timeline]] · [[Microsoft Timeline]] --- ## Era 1 — Theoretical Foundations (1943–1955) The ideas that will become artificial intelligence emerge from wartime mathematics and military-funded university labs. - **1943** — Warren McCulloch and Walter Pitts publish "A Logical Calculus of Ideas Immanent in Nervous Activity" — the first mathematical model of a neural network. Funded by the military research establishment. - **1950** — [[Alan Turing]] publishes "Computing Machinery and Intelligence" — introduces the Turing Test, asks "Can machines think?" The foundational philosophical text of AI. - **1950** — Isaac Asimov publishes "I, Robot" — the Three Laws of Robotics enter popular culture, framing the public imagination of AI as humanoid, obedient, and controllable. None of these assumptions will prove correct. - **1951** — [[Marvin Minsky]] and Dean Edmonds build SNARC — first neural network computer, simulating 40 neurons with vacuum tubes. Funded by the Office of Naval Research. - **1955** — Allen Newell and [[Herbert Simon]] create Logic Theorist — widely considered the first AI program, capable of proving mathematical theorems. > **RWKB angle:** From the very beginning, AI research runs on military money. The Office of Naval Research funds early neural network research. The pattern that defines the next 80 years starts here — the [[Department of Defense]] pays for the science, universities do the work, and private industry harvests the results. --- ## Era 2 — The Birth of AI (1956–1969) The field gets a name, a founding mythology, and its first government patron. Optimism is unbounded. - **Summer 1956** — The Dartmouth Conference — John McCarthy, Marvin Minsky, Claude Shannon, and Nathaniel Rochester organize the workshop that coins the term "artificial intelligence." Attendees include Newell and Simon. The field is born with a prediction: human-level AI within 20 years. - **1957** — Frank Rosenblatt builds the Perceptron — first artificial neural network capable of learning from examples. - **1958** — McCarthy develops the LISP programming language — becomes the standard for AI research for decades. - **1958** — ARPA (later DARPA) created within the [[Department of Defense]] — will become the single largest funder of AI research in history. See [[DARPA Timeline]]. - **1961** — Unimate — the first industrial robot deployed at a GM factory. Automation of physical labor begins. - **1964–66** — ELIZA chatbot created by Joseph Weizenbaum at MIT — the first natural language processing program, simulating a Rogerian therapist. People form emotional bonds with it, alarming its creator. The first demonstration that humans will anthropomorphize machines. - **1966** — SHAKEY the Robot — first general-purpose mobile robot that can reason about its own actions (SRI International, DARPA-funded). - **1969** — Minsky and Papert publish "Perceptrons" — proves mathematical limitations of single-layer neural networks, effectively killing neural network research for 15+ years. > **RWKB angle:** DARPA funds nearly everything. The university-military pipeline that will eventually produce the internet, [[Autonomous Weapons]], and the [[Surveillance State]] infrastructure is built during this decade. Researchers predict human-level AI by the 1980s. They are off by at least 40 years. --- ## Era 3 — First AI Winter (1970–1979) The promises collapse. Military funders demand results. When they don't come, the money disappears. - **1973** — The Lighthill Report — Sir James Lighthill's devastating critique of AI research for the British government. Argues the field has failed to deliver on its promises. The UK slashes AI funding. - **1974** — DARPA cuts AI funding — the agency that birthed the field withdraws support after projects fail to produce military applications. Cascading cuts follow across the US research establishment. - **1970s** — The "[[AI Winter]]" — funding dries up, researchers abandon the field, university programs shrink. The term won't be coined until later, but the phenomenon is real and devastating. > **RWKB angle:** First lesson in the hype-crash cycle that repeats throughout AI history. Military funders demand practical results; when they don't come, the money disappears. This pattern — overpromise, underfund fundamental research, blame the researchers — becomes a template. See also [[Creative Destruction]]. --- ## Era 4 — Expert Systems and the Second Winter (1980–1993) AI returns as rule-based expert systems. Governments panic about falling behind. The bubble inflates and pops. - **1980** — XCON (R1) expert system deployed at Digital Equipment Corporation — configures computer orders, saves $40M/year. The expert systems boom begins. - **1981** — Japan's Fifth Generation Computer Project — $850M government initiative to build AI supercomputers. Triggers panic in the US and UK about falling behind in the technology race. - **1983** — DARPA launches the Strategic Computing Initiative — a $1 billion program to develop AI for military applications: autonomous vehicles, battle management systems, and an aircraft pilot's associate. See [[DARPA Timeline]]. - **1986** — Rumelhart, Hinton, and Williams publish the backpropagation paper — enables training of multi-layer neural networks. The quiet revival of connectionism that will eventually win. - **1987** — Expert systems market collapses — LISP machine market crashes, companies fold. The technology can't scale beyond narrow domains. - **1987–93** — Second AI Winter — funding cuts again, corporate AI labs close, the field contracts. Japan's Fifth Generation project ends with little to show for $850M. - **1988** — Judea Pearl publishes "Probabilistic Reasoning in Intelligent Systems" — introduces Bayesian networks, shifts AI toward statistical and probabilistic methods and away from rule-based approaches. > **RWKB angle:** The expert systems boom is the first AI bubble. Companies overpromise, investors flood in, the technology can't deliver, everyone loses money. Japan's panic-driven government spending is the prototype for today's AI arms race between the US and [[China]]. The pattern recurs: geopolitical fear drives spending far past what the technology can support. --- ## Era 5 — Machine Learning Renaissance (1994–2010) Statistical methods replace hand-coded rules. DARPA seeds the autonomous vehicle industry. Deep learning begins its quiet ascent. - **1997** — [[IBM]] Deep Blue defeats Garry Kasparov at chess — first time a computer beats a reigning world champion. Front-page news worldwide, but a dead end: brute-force search, not learning. - **1998** — Yann LeCun develops convolutional neural networks (CNNs) for handwriting recognition — foundational architecture for modern computer vision. - **2000** — Kismet robot at MIT — demonstrates social and emotional interaction with humans. - **2002** — iRobot Roomba — AI enters homes as a consumer product. - **2004** — DARPA Grand Challenge — autonomous vehicle race in the Mojave Desert. No vehicle finishes the course. Seeds the entire self-driving industry. See [[DARPA Timeline]]. - **2005** — Second DARPA Grand Challenge — Stanford's "Stanley" completes the 132-mile desert course. 5 of 195 teams finish. The teams that compete go on to build Waymo, [[Tesla]] Autopilot, and the autonomous vehicle industry. - **2006** — [[Geoffrey Hinton]] introduces Deep Belief Networks — demonstrates that deep neural networks with many layers can be trained effectively. The deep learning revolution begins. - **2006** — Netflix Prize ($1M) — crowdsourced competition accelerates recommendation algorithm research, establishes the competitive benchmark paradigm that AI labs still use. - **2009** — ImageNet database created by Fei-Fei Li — 14 million labeled images. Will become the benchmark that proves deep learning works at scale. > **RWKB angle:** DARPA Grand Challenge is the template for how military R&D seeds commercial technology — a direct pipeline from [[Department of Defense]] money to [[Tesla]] Autopilot. Deep Blue is a spectacle; the real revolution is happening quietly in university labs still running on defense funding. The pattern holds: military money, university research, corporate capture. --- ## Era 6 — Deep Learning Breakthrough (2011–2016) The "ImageNet moment" proves deep learning works. Big Tech begins acquiring AI companies and researchers. The key players assemble. - **2011** — [[IBM]] Watson wins Jeopardy! against champions Ken Jennings and Brad Rutter — a showpiece, but Watson's later commercial failures at IBM are instructive. - **2011** — Apple launches Siri — AI assistant enters mainstream consumer technology. - **2012** — AlexNet wins ImageNet competition — Geoffrey Hinton, Alex Krizhevsky, and [[Ilya Sutskever]]'s deep convolutional neural network crushes the competition, reducing error rate by over 10 percentage points. The "ImageNet moment" — proves deep learning works at scale. Launches the modern AI era. - **2013** — Word2Vec ([[Google]]) — word embeddings that capture semantic relationships between words. Foundation for modern NLP. - **2013** — [[Google]] acquires DeepMind for ~$500M — [[Larry Page]] and [[Sergey Brin]] now own the world's leading AI research lab AND the world's largest [[Surveillance Capitalism]] platform. - **2013** — [[Boston Dynamics]] (DARPA-funded military robotics company) acquired by [[Google]]. Later sold to [[SoftBank]] (2017), then Hyundai (2020). The pipeline: defense R&D → Big Tech → global conglomerate. - **2014** — Ian Goodfellow invents Generative Adversarial Networks (GANs) — two neural networks compete to generate increasingly realistic synthetic data. Foundation for deepfakes and the [[Disinformation]] crisis that follows. - **2014** — [[Amazon]] Alexa launches — AI voice assistant enters millions of homes, creating a mass consumer [[Surveillance State]] appliance marketed as convenience. - **2015** — [[OpenAI]] founded by [[Sam Altman]], [[Elon Musk]], and others as a nonprofit AI safety lab. Mission: ensure AI benefits all of humanity. [[Peter Thiel]] is an early funder. (This origin story will age poorly.) See [[OpenAI Timeline]]. - **2016** — [[Google DeepMind]]'s AlphaGo defeats Lee Sedol at Go — a game with more possible positions than atoms in the universe. Watched by 200+ million people. A genuine capability leap, not brute force. - **2016** — ProPublica investigation reveals racial bias in criminal sentencing algorithms (COMPAS system) — AI bias becomes a public issue, revealing that "objective" algorithms encode the prejudices of their training data. > **RWKB angle:** The key players assemble. [[Ilya Sutskever]] goes from Hinton's lab to [[Google]] to [[OpenAI]]. [[Peter Thiel]] is an early [[OpenAI]] funder through the [[Thiel Network]]. DeepMind's acquisition means [[Google]] now owns frontier AI research AND the world's largest [[Surveillance Capitalism]] platform. The [[Boston Dynamics]] pipeline — DARPA → [[Google]] → [[SoftBank]] → Hyundai — shows how military robotics washes through corporate hands. --- ## Era 7 — The Transformer Revolution (2017–2020) A single paper changes everything. The geopolitical race begins. The safety researchers start getting fired. - **June 2017** — [[Attention Is All You Need]] — [[Google]] researchers publish the transformer paper. The self-attention mechanism processes entire sequences simultaneously. This single paper is the foundation of GPT, BERT, Claude, Gemini, Llama, and every modern large language model. The most consequential publication in AI history. - **2017** — [[China]] releases its New Generation AI Development Plan — government targets global AI leadership by 2030. The US-China AI race begins in earnest. See also [[CHIPS and Science Act]]. - **2018** — [[Google]] [[Project Maven]] controversy — [[Google]] employees protest the company's Pentagon drone imagery analysis contract. 4,000+ employees sign a petition. [[Google]] publicly pledges not to use AI for weapons. The contract quietly continues through subcontractors. See [[Defense Tech Timeline]]. - **June 2018** — GPT-1 ([[OpenAI]]) — 117 million parameters. Establishes the unsupervised pre-training paradigm. - **October 2018** — BERT ([[Google]]) — bidirectional transformer. Revolutionizes NLP benchmarks. - **2018** — Turing Award goes to Hinton, LeCun, and [[Yoshua Bengio]] — the "Godfathers of AI." - **February 2019** — GPT-2 ([[OpenAI]]) — 1.5 billion parameters. [[OpenAI]] initially withholds release, citing misuse concerns. First major [[AI Safety]] debate around a specific model. - **2019** — [[OpenAI]] transitions from nonprofit to "capped-profit" structure — [[Microsoft]] invests $1B. The mission shift from safety to profit begins. See [[OpenAI Timeline]]. - **June 2020** — GPT-3 ([[OpenAI]]) — 175 billion parameters. Demonstrates emergent few-shot learning — performs tasks from natural language descriptions alone. A qualitative capability discontinuity, not just scale. - **December 2020** — [[Timnit Gebru]] fired from [[Google]]'s AI ethics team — publishes paper on large language model risks (environmental cost, bias). [[Google]] fires her, sparking massive backlash about corporate control of [[AI Safety]] research. The message to safety researchers everywhere: your job security depends on not embarrassing the company. > **RWKB angle:** The transformer paper is the most consequential single publication in AI history. [[Project Maven]] shows the [[Revolving Door]] between [[Google]] and the Pentagon in action. [[OpenAI]]'s nonprofit-to-profit pivot is the canary in the coal mine for [[Regulatory Capture]]. [[Timnit Gebru]]'s firing reveals that Big Tech will suppress its own safety researchers when profits are at stake. [[China]]'s AI plan makes the race geopolitical, and the frame shifts from "can we build it?" to "can we build it before they do?" --- ## Era 8 — Generative AI Explosion (2021–2022) AI goes from industry conversation to civilization-level event. ChatGPT arrives. Nothing is the same after. - **January 2021** — DALL-E ([[OpenAI]]) — text-to-image generation. AI can now create images from text descriptions. The creative industries begin to feel the threat. - **2021** — GitHub Copilot launches — AI code assistant trained on public repositories. Copyright controversy begins: whose code trained this, and who owns the output? See [[Intellectual Property]]. - **2021** — [[Anthropic]] founded by [[Dario Amodei]], [[Daniela Amodei]], and other ex-[[OpenAI]] safety researchers. Motivated by concerns about [[OpenAI]]'s commercialization. Receives funding from [[Google]] and [[Spark Capital]]. See [[Anthropic Timeline]]. - **January 2022** — [[Microsoft]] invests another $10B in [[OpenAI]] — total investment reaches ~$13B. Exclusive cloud partnership. See [[Microsoft Timeline]]. - **Mid-2022** — Midjourney and [[Stable Diffusion]] launch — text-to-image AI goes mainstream. Artists revolt. Class-action lawsuits follow. - **July 2022** — Stability AI releases Stable Diffusion open-source — democratizes image generation, also enables deepfakes and nonconsensual imagery at scale. - **November 30, 2022** — ChatGPT launches — reaches 100 million users in 2 months, the fastest-growing software application in history. This is the "iPhone moment" for AI. The public finally understands what's coming. Every tech company scrambles to respond. - **2022** — Deepfakes proliferate — AI-generated content used in political campaigns, harassment, and fraud. The [[Disinformation]] problem enters a new phase where seeing is no longer believing. > **RWKB angle:** ChatGPT is the inflection point. Before November 2022, AI is an industry conversation. After, it's a civilization-level event. The speed of adoption means regulation can't keep up — [[Congress]] is still trying to understand what a large language model is. Artists, writers, and creators realize their work has been scraped to train the systems replacing them. [[Microsoft]]'s $13B [[OpenAI]] bet is the defining corporate power play. [[Anthropic]]'s founding as "the safety company" will be tested by the same market forces that corrupted [[OpenAI]]. --- ## Era 9 — The AI Arms Race (2023) Everything converges: safety collapses, labor fights back, autonomous weapons deploy, and the corporate capture of AI governance becomes permanent. - **March 2023** — GPT-4 released by [[OpenAI]] — multimodal (text + images), dramatically more capable. Integrated into [[Microsoft]] Bing. - **March 2023** — [[Future of Life Institute]] petition — 1,000+ AI researchers and tech leaders (including [[Elon Musk]]) sign open letter calling for a 6-month pause on AI development beyond GPT-4. Ignored completely. - **March 2023** — [[Anthropic]] launches Claude — positioned as the "safer" alternative to GPT. See [[Anthropic Timeline]]. - **May 2023** — "Mitigating the risk of extinction from AI should be a global priority" — statement signed by hundreds of AI researchers, including Hinton, Bengio, [[Sam Altman]], and [[Dario Amodei]]. Places [[Existential Risk]] from AI alongside pandemics and nuclear war. - **May 2023** — Geoffrey Hinton quits [[Google]] to warn about AI dangers — the "Godfather of AI" says he regrets his life's work. The most credible possible messenger delivers the most alarming possible message. - **July 2023** — [[Meta]] releases Llama 2 open-source — 70 billion parameter model available for commercial use. [[Mark Zuckerberg]] bets on open-source as competitive strategy. The open vs. closed AI debate intensifies. - **2023** — Hollywood writers and actors strike — WGA and SAG-AFTRA strike over AI use in entertainment. First major labor action against AI displacement. Writers win contractual protections against AI replacing them — the first negotiated guardrails in any industry. - **2023** — AI copyright lawsuits begin — [[New York Times]] sues [[OpenAI]] and [[Microsoft]]; Getty Images sues Stability AI; class-action suits from authors and artists. The legal framework for AI training data — who owns what, who gets paid — remains unresolved. See [[Intellectual Property]]. - **October 2023** — [[Joe Biden]] Executive Order on AI — sweeping safety requirements, reporting mandates for frontier models, federal [[AI Safety]] governance frameworks. The high-water mark of US AI regulation. - **November 2023** — [[OpenAI]] board crisis — the board fires [[Sam Altman]] over safety concerns; [[Microsoft]] offers to hire him and the entire staff; 95% of employees threaten to quit; Altman is reinstated in 5 days. Safety-oriented board members are replaced with corporate governance types. The nonprofit mission is effectively dead. See [[OpenAI Timeline]]. - **November 2023** — [[Ilya Sutskever]], whose vote helped fire Altman, stays briefly then leaves to found Safe Superintelligence Inc. The chief scientist of the world's leading AI company concludes that safety and profit cannot coexist under the same roof. - **2023** — [[Israel]] deploys AI targeting systems in Gaza — "Gospel" and "Lavender" AI systems generate 100+ bombing targets per day (vs. 50/year by human analysts). AI-assisted warfare becomes reality. [[Palantir]] expands its [[Israel]] contract. [[Autonomous Weapons]] are no longer hypothetical — they are killing people, powered by Big Tech infrastructure. > **RWKB angle:** 2023 is when everything converges. The [[AI Safety]] movement splits between "[[Existential Risk]]" (Hinton, [[Ilya Sutskever]]) and "present harms" (bias, labor displacement, [[Surveillance State]]). The [[OpenAI]] board crisis proves that corporate power defeats safety governance — $90B in value trumps mission every time. The Hollywood strikes are the first labor war of the AI era. [[Joe Biden]]'s Executive Order is the high-water mark for US AI regulation. [[Israel]]'s AI targeting in Gaza shows [[Autonomous Weapons]] aren't hypothetical — they're in the kill chain now, with [[Palantir]] infrastructure. See [[Defense Tech Timeline]]. --- ## Era 10 — Consolidation and Conflict (2024) The regulatory window closes. The defense-tech consolidation accelerates. AI job displacement goes from theoretical to measured. - **January 2024** — [[OpenAI]] valued at $80B+ — a nonprofit that became a profit machine. See [[OpenAI Timeline]]. - **March 2024** — EU AI Act adopted by European Parliament — the most comprehensive AI regulation in the world, with a risk-based framework and mandatory transparency requirements. The US has nothing comparable. [[Congress]] holds hearings but passes nothing. - **March 2024** — [[Anthropic]] raises $7.3B at $18.4B valuation — [[Google]] and [[Amazon]] are major investors. The "safety company" is now funded by the same [[Surveillance Capitalism]] platforms it was founded to check. See [[Anthropic Timeline]] and [[Amazon Timeline]]. - **April 2024** — DARPA ACE program — first AI vs. human F-16 dogfight. An AI-piloted fighter jet faces a human pilot. The age of autonomous air combat is tested. See [[DARPA Timeline]]. - **May 2024** — Jan Leike resigns from [[OpenAI]]'s superalignment team — says "safety culture and processes have taken a backseat to shiny products." Another safety researcher exits. The pattern from [[Timnit Gebru]]'s firing continues: raise concerns, get pushed out. - **June 2024** — Apple announces Apple Intelligence — AI integration across all Apple products. [[Tim Cook]] bets the company's future on on-device AI. - **September 2024** — [[Gavin Newsom]] vetoes SB 1047 — the most ambitious state [[AI Safety]] bill, killed after intense [[Lobbying]] from the tech industry. [[Effective Accelerationism]] and corporate interests prevail over [[AI Safety]]. [[Regulatory Capture]] in action. - **November 2024** — [[Donald Trump]] wins the [[2024 Presidential Election]] — promises to revoke [[Joe Biden]]'s AI regulations, accelerate AI development, eliminate safety guardrails. The [[Thiel Network]] and [[PayPal Mafia]] return to power with [[David Sacks]] as AI/crypto czar. See [[Defense Tech Timeline]]. - **December 2024** — [[Anduril]] and [[Palantir]] announce defense AI consortium — links [[Anduril]]'s Lattice with Maven Smart System and [[Palantir]]'s AI Platform. Military AI infrastructure consolidates under [[Thiel Network]] companies. See [[Defense Tech Timeline]]. - **2024** — AI job displacement accelerates — tech companies lay off tens of thousands of workers while investing billions in AI. Hollywood loses an estimated 41,000 film/TV jobs. Entry-level white-collar jobs begin disappearing. - **2024** — Pentagon awards contracts to [[Anthropic]], [[Google]], [[OpenAI]], and [[xAI]] for military AI models — the safety companies and the acceleration companies all take defense money. > **RWKB angle:** The regulatory window closes. [[Gavin Newsom]]'s SB 1047 veto and [[Donald Trump]]'s election mean the US will not regulate AI. Only the EU acts. The defense-tech consolidation ([[Anduril]] + [[Palantir]] + Pentagon) creates the [[Autonomous Weapons]] infrastructure under [[Thiel Network]] control. AI job displacement is no longer theoretical — entire industries are contracting. The safety researchers who warned about this are either fired ([[Timnit Gebru]], Jan Leike) or quit (Hinton, [[Ilya Sutskever]]). Every frontier AI company — including [[Anthropic]], the "safety company" — takes Pentagon money. --- ## Era 11 — DeepSeek Shock and Acceleration (2025) [[China]] demonstrates that export controls aren't working. [[Donald Trump]] deregulates. The Pentagon institutionalizes AI warfare. - **January 20, 2025** — [[Donald Trump]] takes office — revokes [[Joe Biden]]'s AI Executive Order on day one. Signals "AI dominance over safety." See [[Defense Tech Timeline]]. - **January 27, 2025** — DeepSeek R1 released — a Chinese AI lab (see [[China]]) demonstrates a frontier-level model trained for approximately $6M on restricted chips. Triggers the largest single-day stock crash in history: [[NVIDIA]] loses $589B in market value in one day. The entire AI infrastructure investment thesis is questioned. The US assumption that chip export controls would prevent [[China]] from competing is shattered. - **February 2025** — Markets recover — Big Tech doubles down on AI spending rather than pulling back. By late 2025, [[NVIDIA]] reaches $5 trillion market cap. The crash was a speed bump, not a reckoning. - **May 2025** — Pentagon raises [[Project Maven]] contract ceiling to $1.3B (up from $480M) — [[Palantir]]'s Maven Smart System becomes the military's AI backbone. See [[Defense Tech Timeline]]. - **2025** — AI Safety Index — [[Future of Life Institute]] grades all major AI companies. [[Anthropic]] gets the highest score but still a "D" for existential safety. [[OpenAI]], [[Meta]], [[xAI]] have "unacceptable" risk practices. The [[AI Safety]] movement produces report cards that nobody is required to read. - **2025** — [[In-Q-Tel]] makes 36 investments — the most active year on record for the [[CIA]]'s venture arm. See [[In-Q-Tel Timeline]]. - **December 2025** — [[Donald Trump]] signs executive order targeting state AI regulation — "minimally burdensome national policy framework," explicitly preempting California and other state AI laws. Federal deregulation meets state-level [[Regulatory Capture]]. - **2025** — Open-source AI accelerates — Llama, DeepSeek, [[Mistral]] models make frontier capabilities widely accessible. The [[AI Safety]] community's worst fear — powerful AI that no one controls — is realized not through AGI but through open distribution. > **RWKB angle:** DeepSeek proves the US doesn't have an AI monopoly and throws the trillion-dollar infrastructure thesis into question (temporarily). [[Donald Trump]]'s deregulation means the US AI industry operates with zero guardrails. The Pentagon's [[Project Maven]] expansion means AI is now embedded in the military kill chain. [[China]]'s open-source strategy democratizes AI while the US tries to maintain control through export restrictions that clearly aren't working. [[In-Q-Tel]]'s record investment year shows the [[CIA]] is scaling its AI portfolio faster than ever. --- ## Era 12 — The Present Moment (2026) The acceleration continues. The safety movement has lost. The question is no longer whether AI transforms everything — it's who controls the transformation and who benefits. - **Current frontier models** — Claude (Anthropic), GPT series ([[OpenAI]]), Gemini ([[Google]]), Grok ([[xAI]]) release updated versions every few months. The pace of capability development has no historical parallel. - **[[OpenAI]]** — valued at $300B+, preparing IPO. The nonprofit that was supposed to "benefit humanity" becomes one of the most valuable companies on Earth. See [[OpenAI Timeline]]. - **[[Anthropic]]** — valued at $61B+, weighing IPO options. The "safety company" navigates pressure to take military contracts while maintaining its differentiation. See [[Anthropic Timeline]]. - **[[Project Maven]]** — transitions to official program of record by September 2026. AI warfare is institutionalized as permanent Pentagon infrastructure. See [[Defense Tech Timeline]]. - **[[Safe Superintelligence]]** — [[Ilya Sutskever]]'s lab takes $5B from [[NVIDIA]] (July 2026) while still holding a ~$32B valuation with no product, no revenue, and no published research. The last frontier lab standing outside the circular chip-financing structure joins it. See [[NVIDIA-Safe Superintelligence Investment]]. - **70% of DARPA programs** now incorporate AI, machine learning, or autonomy. See [[DARPA Timeline]]. - **AI-generated content** floods the internet — detection tools fail to keep up, the information environment degrades. [[Disinformation]] at scale becomes trivially cheap. [[Information Warfare]] enters a new phase. - **AI job displacement** — continues across white-collar sectors. Creative industries, legal research, financial analysis, customer service, software development all report measured contraction. > **RWKB angle:** We are living in the acceleration. The AI companies are worth more than most countries' GDP. The safety movement lost — regulation was killed by corporate [[Lobbying]] and geopolitical competition. [[Autonomous Weapons]] are deployed. Creative workers are displaced. The [[Surveillance State]] infrastructure is complete. And the technology is still improving exponentially. The answer to "who controls this?" is: the same defense contractors, tech billionaires, and intelligence agencies — the [[Military-Industrial Complex]], the [[Tech Oligarchy]], and the [[Thiel Network]] — that have shaped the technology from the beginning. --- ## Why This Matters The arc of this timeline reveals a consistent pattern across eight decades: 1. **Military money seeds the research** — from the Office of Naval Research funding neural networks in the 1940s to DARPA funding autonomous vehicles in the 2000s to the Pentagon's [[Project Maven]] today 2. **University labs do the work** — then their best researchers get hired away by corporations (Hinton to [[Google]], [[Ilya Sutskever]] to [[OpenAI]], the entire deep learning faculty to Big Tech) 3. **Corporations capture the results** — [[Google]], [[Microsoft]], [[Amazon]], and [[Meta]] acquire or acqui-hire every breakthrough, creating an [[Antitrust]] problem no regulator has addressed 4. **The public arrives last** — as consumers, as training data, and as labor to be displaced The [[Revolving Door]] between the Pentagon, the intelligence community, and Big Tech is not a bug — it's the system working as designed. [[Peter Thiel]]'s [[Palantir]] was funded by [[In-Q-Tel]]. [[Google]]'s [[Project Maven]] contract became the template for every AI company's defense work. [[OpenAI]] went from "nonprofit for humanity" to "$300B corporation preparing an IPO" in under a decade. [[Anthropic]], founded to be the safety-first alternative, takes Pentagon money alongside everyone else. The "[[Effective Altruism]] vs. [[Effective Accelerationism]]" debate — the philosophical frame that dominated 2023–2024 — was always a distraction from the structural reality: AI development is controlled by a small number of [[Tech Oligarchy]] players with deep ties to the defense establishment, operating under zero regulatory constraint in the US, and racing against [[China]] under conditions where safety is a competitive disadvantage. Understanding this matters because the decisions being made right now — about [[Autonomous Weapons]], about [[Surveillance State]] infrastructure, about which jobs survive and which don't, about who owns the [[Intellectual Property]] created by AI — will shape the next century. And those decisions are being made by the people and institutions documented in this timeline, not by democratic process. --- ## See Also - [[Defense Tech Timeline]] — the Silicon Valley-to-Pentagon pipeline in detail - [[DARPA Timeline]] — the agency that funded AI from the beginning - [[In-Q-Tel Timeline]] — the CIA's venture capital arm - [[OpenAI Timeline]] — nonprofit to profit machine - [[Anthropic Timeline]] — the "safety company" under pressure - [[Microsoft Timeline]] — the $13B bet on [[OpenAI]] - [[Facebook Timeline]] — [[Meta]]'s open-source AI strategy - [[Amazon Timeline]] — cloud infrastructure and [[Anthropic]] investment