Timeline of Machine Intelligence Research Institute

From Timelines
Revision as of 10:32, 17 December 2024 by Hith (talk | contribs)
Jump to: navigation, search
The timeline currently offers focused coverage of the period until June 2024. It is likely to miss important developments outside this period (particularly after this period) though it may have a few events from after this period.

This is a timeline of Machine Intelligence Research Institute. Machine Intelligence Research Institute (MIRI) is a nonprofit organization that does work related to AI safety.

Sample questions

This is an experimental section that provides some sample questions for readers, similar to reading questions that might come with a book. Some readers of this timeline might come to the page aimlessly and might not have a good idea of what they want to get out of the page. Having some "interesting" questions can help in reading the page with more purpose and in getting a sense of why the timeline is an important tool to have.

The following are some interesting questions that can be answered by reading this timeline:

  • Which Singularity Summits did MIRI host, and when did they happen? (Sort by the "Event type" column and look at the rows labeled "Conference".)
  • What was MIRI up to for the first ten years of its existence (before Luke Muehlhauser joined, before Holden Karnofsky wrote his critique of the organization)? (Scan the years 2000–2009.)
  • How has MIRI's explicit mission changed over the years? (Sort by the "Event type" column and look at the rows labeled "Mission".)

The following are some interesting questions that are difficult or impossible to answer just by reading the current version of this timeline, but might be possible to answer using a future version of this timeline:

  • When did some big donations to MIRI take place (for instance, the one by Peter Thiel)?
  • Has MIRI "done more things" between 2010–2013 or between 2014–2017? (More information)

Big picture

Time period Development summary More details
1998–2002 Various publications related to creating a superhuman AI During this period, Eliezer Yudkowsky publishes a series of foundational documents about designing superhuman AI. Key works include "Coding a Transhuman AI," "The Plan to Singularity," and "Creating Friendly AI." These writings lay the groundwork for the AI alignment problem. Additionally, the Flare Programming Language project is launched to assist in the creation of a superhuman AI, marking the early technical ambitions of the movement.
2004–2009 Tyler Emerson's tenure as executive director Under Emerson’s leadership, MIRI (then known as the Singularity Institute) experiences growth and increased visibility. Emerson launches the Singularity Summit, a major event that brings together AI researchers, futurists, and thought leaders. MIRI relocates to the San Francisco Bay Area, gaining a strong foothold in the tech industry. During this period, Peter Thiel becomes a key donor and public advocate, lending credibility and significant financial support to the institute.
2006–2009 Modern rationalist community forms This period sees the formation of the modern rationalist community. Eliezer Yudkowsky contributes by founding the websites Overcoming Bias and LessWrong. These platforms become central hubs for discussions on rationality, AI safety, and existential risks. Yudkowsky's Sequences, a comprehensive collection of essays on rationality, are written and gain a wide following, helping shape the philosophy of many within the AI safety and rationalist movements.
2006–2012 The Singularity Summits take place annually The Singularity Summit takes place annually during this period, attracting both prominent thinkers and the general public interested in AI, technology, and futurism. In 2012, the organization changes its name from "Singularity Institute for Artificial Intelligence" to the "Machine Intelligence Research Institute" (MIRI) to better reflect its focus on AI research rather than broader technological futurism. MIRI also sells the Singularity Summit to Singularity University, signaling a shift toward a more focused research agenda.
2009–2012 Michael Vassar's tenure as president Michael Vassar serves as president during this period, continuing to build on the progress made by previous leadership. Vassar focuses on strategic development and positions MIRI within the broader intellectual landscape, further cementing its role as a leader in AI safety research.
2011–2015 Luke Muehlhauser's tenure as executive director Luke Muehlhauser takes over as executive director and is credited with professionalizing the organization and improving donor relations. Under his leadership, MIRI undergoes significant changes, including a name change, a shift in focus from outreach to research, and stronger connections with the Effective Altruism community. Muehlhauser builds relationships with the AI research community, laying the groundwork for future collaborations and funding opportunities.[1][2][3]
2013–2015 Change of focus MIRI shifts its research focus to AI safety and technical math-based research into Friendly AI. During this period, MIRI reduces its public outreach efforts to concentrate on solving fundamental problems in AI safety. It stops hosting major public events like the Singularity Summit and begins focusing almost exclusively on research efforts to address the alignment problem and existential risks from advanced AI systems.
2015–2023 Nate Soares's tenure as executive director Nate Soares, who takes over as executive director in 2015, continues to steer MIRI toward more technical and research-based work on AI safety. Soares expands MIRI’s collaboration with other AI safety organizations and risk researchers. During this time, MIRI receives major funding boosts from cryptocurrency donations and the Open Philanthropy Project in 2017. In 2018, MIRI adopts a "nondisclosed-by-default" policy for much of its research to prevent potential misuse or risks from the dissemination of sensitive AI safety work.
2023–present Leadership transitions at MIRI MIRI undergoes significant leadership changes in 2023. Nate Soares steps down as executive director and transitions to President, focusing on strategic oversight. Malo Bourgon becomes the new CEO, handling day-to-day operations and growth management. Alex Vermeer takes on the role of COO, providing internal support and leadership. The organization continues to prioritize AI safety research and collaborates with other AI safety organizations to address emerging challenges in the field.

Highlights by year (2013 onward)

Full timeline

Year Highlights
2013 MIRI (formerly SIAI) continues its focus on AI alignment research and community-building. Collaboration with the rationalist and Effective Altruism movements deepens. MIRI establishes itself as a key organization for long-term AI safety, setting the groundwork for its agent foundations research agenda.
2014 MIRI publishes several key technical research papers on decision theory and logical uncertainty. The Effective Altruism community increasingly recognizes MIRI's role in addressing existential risks from AI. The Intelligent Agent Foundations Forum is launched to foster collaboration among AI alignment researchers.
2015 MIRI co-organizes the Puerto Rico AI Safety Conference with the Future of Life Institute, a pivotal event that brings mainstream attention to AI risks. Nate Soares succeeds Luke Muehlhauser as MIRI’s Executive Director, signaling a new phase for the organization. MIRI holds multiple workshops on logical decision theory, logical uncertainty, and Vingean reflection, solidifying its research agenda.
2016 MIRI shifts its research focus toward highly reliable agent design and alignment for advanced AI systems. Scott Garrabrant and collaborators publish the "Logical Induction" paper, a major breakthrough in reasoning under uncertainty. Open Philanthropy awards MIRI a $500,000 grant for general support, acknowledging its role in reducing AI-related risks.
2017 Cryptocurrency donations surge, boosting MIRI’s funding, including a significant contribution from Ethereum co-founder Vitalik Buterin. Open Philanthropy grants MIRI $3.75 million, its largest single grant to date. The organization also begins exploring new research directions while maintaining its focus on AI safety and alignment.
2018 MIRI announces its nondisclosure-by-default research policy, marking a strategic shift to safeguard alignment progress. Edward Kmett, a prolific Haskell developer, joins MIRI to contribute to its research. The Embedded Agency sequence, exploring naturalized agency concepts, is published, becoming a foundational reference for AI alignment discussions.
2019 Open Philanthropy and the Survival and Flourishing Fund provide substantial grants to MIRI, supporting its ongoing AI safety research. MIRI’s research agenda focuses on building robust agents capable of reasoning under logical uncertainty, with continued emphasis on solving core AI alignment challenges.
2020 MIRI receives its largest grant to date—$7.7 million over two years—from Open Philanthropy, reinforcing its position as a leader in AI safety research. Internal discussions about relocating MIRI’s operations emerge but conclude with a decision to remain in Berkeley, California.
2021 Major cryptocurrency donations, including from Vitalik Buterin, provide critical funding for MIRI’s research. Scott Garrabrant introduces "Finite Factored Sets" as a novel approach to causal inference, generating interest in the alignment research community.
2022 Eliezer Yudkowsky publishes "AGI Ruin: A List of Lethalities," renewing discussions on catastrophic risks from AI systems. MIRI refines its internal strategy, pausing public communications to focus on advancing its research agenda amid a rapidly evolving AI landscape.
2023 MIRI undergoes leadership changes, with Malo Bourgon appointed CEO and Nate Soares transitioning to President. Eliezer Yudkowsky’s public advocacy for halting advanced AI development garners significant media attention, amplifying calls for stricter AI governance.
2024 MIRI launches a new technical governance research team, engaging with international AI policy initiatives and contributing to global discussions on AI safety. The organization announces the termination of the Visible Thoughts Project due to evolving research priorities and limited engagement.
Year Month and date Event type Details
1979 September 11 Eliezer Yudkowsky is born.[4]
1996 November 18 Eliezer Yudkowsky writes the first version of "Staring into the Singularity".[5]
1998 Publication The initial version of "Coding a Transhuman AI" (CaTAI) is published.[6]
1999 March 11 The Singularitarian mailing list is launched. The mailing list page notes that although hosted on MIRI's website, the mailing list "should be considered as being controlled by the individual Eliezer Yudkowsky".[7]
1999 September 17 The Singularitarian mailing list is first informed (by Yudkowsky?) of "The Plan to Singularity" (called "Creating the Singularity" at the time).[8]
2000–2003 Eliezer Yudkowsky's "coming of age" (including his "naturalistic awakening," in which he realizes that a superintelligence would not necessarily follow human morality) takes place during this period.[9][10][11]
2000 January 1 Publication "The Plan to Singularity" version 1.0 is written and published by Eliezer Yudkowsky, and posted to the Singularitarian, Extropians, and transhuman mailing lists.[8]
2000 January 1 Publication "The Singularitarian Principles" version 1.0 by Eliezer Yudkowsky is published.[12]
2000 February 6 The first email is sent on SL4 ("Shock Level Four"), a mailing list about transhumanism, superintelligent AI, existential risks, and so on.[13][14]
2000 May 18 Publication "Coding a Transhuman AI" (CaTAI) version 2.0a is "rushed out in time for the Foresight Gathering".[15]
2000 July 27 Mission Machine Intelligence Research Institute is founded as the Singularity Institute for Artificial Intelligence by Brian Atkins, Sabine Atkins (then Sabine Stoeckel) and Eliezer Yudkowsky. The organization's mission ("organization's primary exempt purpose" on Form 990) at the time is "Create a Friendly, self-improving Artificial Intelligence"; this mission would be in use during 2000–2006 and would change in 2007.[16]:3[17]
2000 September 1 Publication Large portions of "The Plan to Singularity" are declared obsolete following the formation of the Singularity Institute and a fundamental shift in AI strategy after the publication of "Coding a Transhuman AI" (CaTAI) version 2.[8] This marks a pivotal moment in MIRI's (then known as the Singularity Institute) focus. Earlier discussions about the Singularity give way to a more precise, strategic approach to developing safe, self-improving AI. The obsoletion reflects how new insights are rapidly reshaping the institute's path.
2000 September 7 Publication Version 2.2.0 of "Coding a Transhuman AI" (CaTAI) is published.[15] CaTAI is a detailed technical document outlining the architecture for creating a transhuman-level artificial intelligence. It covers key ideas on how an AI can be designed to improve itself safely without deviating from its original, human-aligned goals. This text serves as a core theoretical foundation for MIRI's mission, advocating for AI development grounded in ethical and rational decision-making frameworks.
2000 September 14 Project The first Wayback Machine snapshot of MIRI's website is captured, using the singinst.org domain name.[18] The launch of the website signals MIRI’s formal entry into the public discourse on AI safety and existential risks. It becomes a hub for sharing research, ideas, and resources aimed at academics, technologists, and the broader community interested in the ethical implications of advanced AI.
2001 April 8 Project MIRI begins accepting donations after receiving tax-exempt status.[19] Receiving tax-exempt status is a critical milestone for MIRI, allowing it to officially solicit and receive donations from the public. This status helps secure the financial support necessary to expand their research efforts and build a formal research team.
2001 April 18 Publication Version 0.9 of "Creating Friendly AI" is released.[20] This early draft outlines the first comprehensive framework for developing "Friendly AI" — an AI system designed to operate under constraints that ensure its goals remain aligned with human interests. It is an important early step in formalizing the institute’s approach to safe AI development.
2001 June 14 Publication The "SIAI Guidelines on Friendly AI" is published.[21] These guidelines serve as a set of ethical and technical principles meant to guide AI researchers in designing systems that prioritize human well-being. The guidelines represent MIRI’s effort to communicate the necessity of carefully managing AI's development and potential risks.
2001 June 15 Publication Version 1.0 of "Creating Friendly AI" is published.[22] This version is the first full publication of MIRI’s flagship research document. It provides a detailed analysis of how to design AI systems that remain aligned with human values, even as they gain the ability to self-improve. It is considered one of the key early texts in the AI safety field.
2001 July 23 Project MIRI formally launches the development of the Flare programming language under Dmitriy Myshkin.[23] The Flare project is conceived as a way to build a programming language optimized for AI development and safety. Though it is eventually canceled, it shows MIRI’s early commitment to exploring technical approaches to developing safer AI systems.
2001 December 21 Domain MIRI secures the flare.org domain name for its Flare programming language project.[23] This acquisition signifies MIRI's focus on developing tools that assist in the creation of AI, though Flare itself is eventually shelved due to technical challenges and shifting priorities.
2002 March 8 AI Box Experiment The first AI Box experiment conducted by Eliezer Yudkowsky, against Nathan Russell as gatekeeper, takes place. The AI is released.[24] This experiment involves testing whether a hypothetical AI can convince a human "gatekeeper" to let it out of a confined environment — highlighting the persuasive abilities that a sufficiently advanced AI might possess, even when theoretically controlled.
2002 April 7 Publication A draft of "Levels of Organization in General Intelligence" is announced on SL4.[25][26] This paper explores theoretical foundations for creating AI that mimics general human intelligence, contributing to the field’s understanding of how to structure and organize machine learning systems.
2002 July 4–5 AI box The second AI box experiment by Eliezer Yudkowsky, against David McFadzean as gatekeeper, takes place. The AI is released, showcasing the potential persuasive power of advanced AI in overcoming human-imposed restrictions, even in a controlled experimental setting.[27]
2002 September 6 Staff Christian Rovner is appointed as MIRI's volunteer coordinator, formalizing efforts to engage volunteers in advancing the institute's mission of Friendly AI development.[23]
2002 October 1 MIRI "releases a major new site upgrade" with various new pages, reflecting its growing presence and commitment to outreach and transparency in its research efforts.[23]
2002 October 7 Project MIRI announces the creation of its volunteers mailing list to streamline communication and foster collaboration among its expanding network of supporters.[23]
2003 Project The Flare Programming language project is officially canceled, marking a strategic shift in MIRI's focus to other priorities in the pursuit of advanced AI research.[28]
2003 Publication Eliezer Yudkowsky's "An Intuitive Explanation of Bayesian Reasoning" is published. This accessible explanation of Bayesian statistics becomes a foundational resource for those interested in rational decision-making and probability.[29]
2003 April 30 Eliezer Yudkowsky posts an update about MIRI to the SL4 mailing list. The update highlights the need for an executive director, bright programmers, and mentions plans for a rationality-focused book to attract talent.[30]
2004 March 4–11 Staff MIRI announces Tyler Emerson as executive director. Emerson's expertise in nonprofit management and leadership aims to strengthen the organization’s capacity to achieve its mission.[31][32]
2004 April 7 Staff Michael Anissimov is announced as MIRI's advocacy director. A dedicated volunteer and influential voice in the transhumanist community, Anissimov is tasked with leading advocacy initiatives.[33]
2004 April 14 Outside review The first version of the Wikipedia page for MIRI is created. This marks a step in MIRI's broader public visibility and transparency.[34]
2004 May Publication Eliezer Yudkowsky's paper "Coherent Extrapolated Volition" is published around this time, outlining a vision for aligning AI development with human values. Originally called "Collective Volition," it is later announced on the MIRI website on August 16.[35][31]
2004 August 5–8 Conference TransVision 2004 takes place. TransVision, the World Transhumanist Association's annual event, sees MIRI participating as a sponsor, reflecting its growing influence in the transhumanist and AI communities.[31]
2005 January 4 Publication "A Technical Explanation of Technical Explanation" is published.[36] Eliezer Yudkowsky explores the nature of technical explanations, emphasizing how we can communicate complex ideas with clarity and rigor. This work becomes foundational for those studying rationality and AI, offering insights into how we convey and understand deep technical topics. It plays an important role in grounding the theoretical framework behind AI safety research. MIRI announces its release, underlining its importance to their broader research agenda.[37]
2005 Conference MIRI gives presentations on AI and existential risks at Stanford University, the Immortality Institute’s Life Extension Conference, and the Terasem Foundation.[38] These presentations help MIRI broaden the conversation about the risks associated with AI development. By engaging academic audiences at Stanford and futurist communities at the Life Extension Conference, MIRI establishes itself as a critical voice in discussions about how AI can impact humanity’s future. These events also allow MIRI to connect its mission with broader existential concerns, including life extension and the future of human intelligence.
2005 Publication Eliezer Yudkowsky contributes chapters to Global Catastrophic Risks, edited by Nick Bostrom and Milan M. Ćirković.[38] Although the book is officially published in 2008, Yudkowsky’s early contributions focus on the potential dangers of advanced AI and global catastrophic risks. His chapters play a vital role in shaping the emerging field of AI safety, providing critical perspectives on how advanced AI could shape or threaten humanity’s future. This collaboration with prominent scholars like Nick Bostrom helps solidify MIRI's reputation within the existential risk community.
2005 February 2 Relocation MIRI relocates from the Atlanta metropolitan area in Georgia to the Bay Area of California.[31] This move is strategic, placing MIRI at the heart of Silicon Valley, where technological advancements are rapidly accelerating. By moving to the Bay Area, MIRI positions itself closer to influential tech companies and research institutions, allowing it to form stronger partnerships and participate more actively in the conversations around AI development and safety. The relocation also signals MIRI’s commitment to influencing the future of AI in a global technology hub.
2005 July 22–24 Conference MIRI sponsors TransVision 2005 in Caracas, Venezuela.[31] TransVision is one of the world’s leading transhumanist conferences, focusing on how emerging technologies, including AI, can impact humanity’s evolution. MIRI’s sponsorship of this event highlights its dedication to transhumanist goals, such as safe AI and human enhancement. The sponsorship also enables MIRI to reach new international audiences, solidifying its role as a global leader in the field of AI safety and existential risk.
2005 August 21 AI Box Experiment Eliezer Yudkowsky conducts the third AI Box experiment, with Carl Shulman as the gatekeeper.[39] This experiment explores the theoretical dangers of an advanced AI persuading a human to release it from confinement. Yudkowsky’s successful manipulation as the AI in this experiment further demonstrates the potential risks posed by highly intelligent systems. The AI Box experiment serves as a thought-provoking exercise in AI safety, illustrating the need for stringent safeguards in future AI development.
2005–2006 December 20, 2005 – February 19, 2006 Financial The 2006 $100,000 Singularity Challenge, led by Peter Thiel, successfully matches donations up to $100,000.[31][40] Peter Thiel’s donation marks the beginning of his significant financial support for MIRI, which continues for many years. The Singularity Challenge helps MIRI raise critical funds for its research, enabling the organization to expand its efforts in AI safety and existential risk mitigation.
2006 January Publication "Twelve Virtues of Rationality" is published.[41] This essay, written by Eliezer Yudkowsky, lays out twelve core principles or virtues meant to guide rational thinkers. It highlights values like curiosity, empiricism, and precision in thinking, which Yudkowsky frames as essential for clear, logical analysis. The publication is relatively short and structured as a set of concise principles, making it an easily digestible guide for those interested in improving their rational thinking skills.
2006 February 13 Staff Peter Thiel joins MIRI’s Board of Advisors.[31] Peter Thiel, the tech entrepreneur and venture capitalist, becomes a part of MIRI’s leadership by joining its Board of Advisors. Thiel’s addition to the board follows his growing interest in existential risks and advanced AI, which aligns with MIRI’s mission. His role primarily involves advising MIRI on its strategic direction and helping the organization secure long-term financial support for its AI safety research.
2006 May 13 Conference The first Singularity Summit takes place at Stanford University.[42][43][44] The Singularity Summit is held as a one-day event at Stanford University and gathers leading scientists, technologists, and thinkers to discuss the rapid pace of technological development and the potential for artificial intelligence to surpass human intelligence. The agenda includes a series of talks and panel discussions, with topics ranging from AI safety to the philosophical implications of superintelligent machines. Attendees include a mix of academics, entrepreneurs, and futurists, marking it as a landmark event for those interested in the technological singularity.
2006 November Project Robin Hanson launches the blog Overcoming Bias.[45] This project is a personal blog started by Robin Hanson, focusing on cognitive biases and rationality. It is a platform for Hanson and guest contributors to write about topics such as human decision-making, bias in everyday life, and how individuals can improve their thinking. Overcoming Bias quickly gains a readership among academics, technologists, and rationality enthusiasts.
2007 May Mission MIRI updates its mission statement to focus on "developing safe, stable, and self-modifying Artificial General Intelligence." This reflects the organization’s shift in focus to ensuring that future AI systems remain aligned with human values.[46]
2007 July Project MIRI launches its outreach blog. The blog serves to engage the public in discussions around AI safety and rationality. It provides a platform for MIRI staff and guest writers to share research updates, existential risk concerns, and general AI news.[38]
2007 August Project MIRI begins its Interview Series, publishing interviews with leading figures in AI, cognitive science, and existential risk. These interviews offer insights into AGI safety and foster connections within the academic community.[38]
2007 September Staff Ben Goertzel becomes Director of Research at MIRI, bringing formal leadership to MIRI’s research agenda. He focuses on advancing research in AGI safety, leveraging his expertise in cognitive architectures.[47]
2007 May 16 Project MIRI publishes its first introductory video on YouTube.[48] The video is created as an introduction to MIRI’s mission and the field of AI safety. It explains the basic concepts of AI risk and outlines MIRI’s role in researching the challenges posed by advanced AI systems. The video is designed to be accessible to a general audience, helping MIRI reach people who might not be familiar with the nuances of AI development.
2007 July 10 Publication The oldest post on MIRI’s blog, "The Power of Intelligence", is published by Eliezer Yudkowsky.[49] This blog post explores the fundamental concept of intelligence and how it shapes the world. It discusses the role of intelligence in achieving goals and solving problems, emphasizing its potential impact on the future. The post serves as an introduction to Yudkowsky’s broader work on AI safety and rationality, marking the start of MIRI’s ongoing blog efforts.
2007 September 8–9 Conference The Singularity Summit 2007 is held in the San Francisco Bay Area.[42][50] The second Singularity Summit takes place over two days and features presentations from leading thinkers in AI and technology. Topics include the future of artificial intelligence, the ethics of AI development, and the technological singularity. The event builds on the success of the previous year’s summit, expanding in both size and scope, and attracting a broader audience from academia and the tech industry.
2008 January Publication "The Simple Truth" is published. This short, fictional story by Eliezer Yudkowsky explains the basic concepts of truth and rationality, illustrating how humans can understand objective reality through evidence and reasoning. It serves as an introduction to epistemology, making complex ideas about knowledge more accessible to a general audience.[51]
2008 March Project MIRI expands its Interview Series, broadening its scope to include a wider range of experts in AI safety, cognitive science, and philosophy of technology. This expansion provides a more comprehensive view of the diverse research efforts and opinions shaping AGI and existential risk discussions.[38]
2008 June Project MIRI launches its summer intern program, engaging young researchers and students in AI safety research. The program allows participants to work with MIRI’s research staff, contributing to ongoing projects and gaining hands-on experience in AGI research. It becomes a key method for developing talent and integrating fresh perspectives.[38]
2008 July Project OpenCog is founded with support from MIRI and Novamente LLC, directed by Ben Goertzel. OpenCog receives additional funding from Google Summer of Code, allowing 11 interns to work on the project in the summer of 2008. The initiative focuses on cognitive architectures and remains central to Goertzel's research efforts at MIRI until 2010.[52][53]
2008 October 25 Conference The Singularity Summit 2008 takes place in San Jose.[54][55]
2008 November–December Outside review The AI-Foom debate between Robin Hanson and Eliezer Yudkowsky takes place. The blog posts from the debate would later be turned into an ebook by MIRI.[56][57]
2009 Project MIRI launches the Visiting Fellows Program in 2009. This initiative allows individuals from various backgrounds to spend several weeks at MIRI, engaging in collaborative research and contributing to projects related to Friendly AI and rationality. The program becomes a key method of recruitment for future MIRI researchers.[38]
2009 (early) Staff Tyler Emerson, who served as executive director of MIRI, steps down early in the year. His departure marks a leadership transition that eventually sees Michael Vassar take on a more prominent role within the organization.[58]
2009 (early) Staff Michael Anissimov is hired as Media Director. Having served as MIRI’s Advocacy Director in previous years, it is unclear whether he briefly left the organization or simply transitioned into a new role.[58]
2009 February Project Eliezer Yudkowsky establishes LessWrong, a community blog dedicated to discussing topics related to rationality, decision theory, and the development of Friendly AI. The site serves as a spiritual successor to his posts on Overcoming Bias and quickly becomes a central hub for Singularity and Effective Altruism communities. It is described as instrumental in MIRI's recruitment efforts, with many participants of MIRI's Visiting Fellows Program having first encountered the organization through LessWrong.[59][58]
2009 February 16 Staff Michael Vassar announces his role as President of MIRI in a blog post titled "Introducing Myself." Vassar, who was a key figure in the organization’s outreach efforts, remains president until 2012, focusing on strategic vision and external partnerships.[60]
2009 April Publication Eliezer Yudkowsky completes the Sequences, a series of blog posts on LessWrong that cover topics ranging from epistemology and rationality to AI safety. These posts are later compiled into the book Rationality: From AI to Zombies.[61]
2009 August 13 Social Media MIRI establishes its official Twitter account under the handle @singinst. This move marks the beginning of MIRI's broader efforts to engage with the public through social media channels.[62]
2009 September Staff Amy Willey Labenz begins an internship at MIRI, focusing on administrative and operational tasks. During her time as an intern, she demonstrates keen attention to detail, particularly in financial oversight. In November, she discovers a significant case of embezzlement within the organization, identifying discrepancies that had gone unnoticed. Her findings lead to an internal investigation, and her role in resolving the issue is seen as critical in protecting MIRI's financial stability. Recognizing her diligence and competence, MIRI promotes Amy Willey Labenz to Chief Compliance Officer by the end of the year. In her new role, she is tasked with ensuring the organization's financial integrity and compliance with legal standards, playing a pivotal part in strengthening MIRI's internal operations.[63]
2009 October Project MIRI launches The Uncertain Future, a website that allows users to build mathematically rigorous models to predict the impact of future technologies. The project began development in 2008 and is seen as an innovative tool for those interested in exploring the potential trajectories of technological progress.[64][65]
2009 October 3–4 Conference The Singularity Summit 2009 takes place in New York, bringing together leading thinkers in technology, AI, and futurism. This annual event, hosted by MIRI, serves as a major platform for discussions about the Singularity and the implications of rapidly advancing technologies.[66][67]
2009 November Financial An embezzlement scandal involving a former contractor is uncovered, resulting in a reported theft of $118,803. The discovery leads to significant internal changes within MIRI and the eventual recovery of some funds through legal action.[68][69]
2009 December Staff Following the embezzlement case, Amy Willey Labenz, who uncovered the theft during her internship, is promoted to Chief Compliance Officer. Her role focuses on strengthening MIRI’s financial and operational compliance.[58][63]
2009 December 11 Influence The third edition of Artificial Intelligence: A Modern Approach, a seminal textbook by Stuart J. Russell and Peter Norvig, is published. In this edition, Friendly AI and Eliezer Yudkowsky are mentioned for the first time, marking an important moment for MIRI's ideas within mainstream AI literature.
2009 December 12 Project MIRI announces that The Uncertain Future has reached beta status. The tool, which allows users to explore scenarios of technological progress, is unveiled on the MIRI blog.[70]
2010 Mission The organization mission changes to: "To develop the theory and particulars of safe self-improving Artificial Intelligence; to support novel research and foster the creation of a research community focused on safe Artificial General Intelligence; and to otherwise improve the probability of humanity surviving future technological advances."[71] This mission is used in 2011 and 2012 as well.
2010 February 28 Publication The first chapter of Eliezer Yudkowsky's fan fiction Harry Potter and the Methods of Rationality is published. The book is published as a serial concluding on March 14, 2015.[72][73] The fan fiction becomes the initial contact with MIRI of several larger donors to MIRI.[74]
2010 April Staff Amy Willey Labenz is promoted to Chief Operating Officer; she was previously the Chief Compliance Officer. From 2010 to 2012 she also serves as the Executive Producer of the Singularity Summits.[63]
2010 June 17 Popular culture Zendegi, a science fiction book by Greg Egan, is published. The book includes a character called Nate Caplan (partly inspired by Eliezer Yudkowsky and Robin Hanson), a website called Overpowering Falsehood dot com (partly inspired by Overcoming Bias and LessWrong), and a Benign Superintelligence Bootstrap Project, inspired by the Singularity Institute's friendly AI project.[75][76][77]
2010 August 14–15 Conference The Singularity Summit 2010 takes place in San Francisco.[78]
2010 December 21 Social media The first post on the MIRI Facebook page is from this day.[79][80]
2010–2011 December 21, 2010 – January 20, 2011 Financial The Tallinn–Evans $125,000 Singularity Challenge takes place. The Challenge is a fundraiser in which Edwin Evans and Jaan Tallinn match each dollar donated to MIRI up to $125,000.[81][82]
2011 February 4 Project The Uncertain Future is open-sourced.[65]
2011 February Outside review Holden Karnofsky of GiveWell has a conversation with MIRI staff. The conversation reveals the existence of a "Persistent Problems Group" at MIRI, which will supposedly "assemble a blue-ribbon panel of recognizable experts to make sense of the academic literature on very applicable, popular, but poorly understood topics such as diet/nutrition".[83] On April 30, Karnofsky posts the conversation to the GiveWell mailing list.[84]
2011 April Staff Luke Muehlhauser begins as an intern at MIRI.[85]
2011 May 10 – June 24 Outside review Holden Karnofsky of GiveWell and Jaan Tallinn (with Dario Amodei being present in the initial phone conversation) correspond regarding MIRI's work. The correspondence is posted to the GiveWell mailing list on July 18.[86]
2011 June 24 Domain A Wayback Machine snapshot on this day shows that singularity.org has turned into a GoDaddy.com placeholder.[87] Before this, the domain is some blog, most likely unrelated to MIRI.[88]
2011 July 18 – October 20 Domain At least during this period, the singularity.org domain name redirects to singinst.org/singularityfaq.[88]
2011 September 6 Domain The first Wayback Machine capture of singularityvolunteers.org is from this day.[89] For a time, the site is used to coordinate volunteer efforts.
2011 October 15–16 Conference The Singularity Summit 2011 takes place in New York.[90]
2011 October 17 Social media The Singularity Summit YouTube account, SingularitySummits, is created.[91]
2011 November Staff Luke Muehlhauser is appointed executive director of MIRI.[92]
2011 December 12 Project Luke Muehlhauser announces the creation of Friendly-AI.com, a website introducing the idea of Friendly AI.[93]
2012 Staff Michael Vassar leaves MIRI to found MetaMed, a personalized medical advisory company. Vassar co-founds the company with Skype co-creator Jaan Tallinn and $500,000 in funding from Peter Thiel. MetaMed aims to revolutionize the medical system by applying rational decision-making and advanced data analysis to personalized health solutions. Despite its ambitious goals, MetaMed's services are initially aimed at wealthy clients, offering personalized medical literature reviews and customized health studies. The underlying mission is to signal the power of rationality in complex systems like medicine, even though it may initially serve the privileged.[94]
2011–2012 Opinion Q&A with Luke Muehlhauser: In a two-part Q&A series, Luke Muehlhauser, the newly appointed Executive Director of MIRI, shares his vision and priorities for the organization. He outlines MIRI’s evolving strategy to focus more intensively on AI alignment research and less on broader advocacy for the singularity. Muehlhauser discusses the challenges of making progress in AI safety research and the importance of recruiting highly talented researchers to tackle this problem. He highlights that MIRI’s goal is not just to build smarter AI, but to ensure its safety and alignment with human values.[95][96]
2012 Domain From February 4 to May 4, 2012, MIRI redirects its domain from singularity.org to singinst.org, reflecting its shift from broad discussions on the singularity to a focused emphasis on AI safety and technical research. This change consolidates its online presence under a new identity aligned with its mission of developing safe, human-aligned AI.[97]
2012 May 8 Progress Report MIRI publishes its April 2012 progress report, which includes the announcement of the new name for the Center for Applied Rationality (CFAR). Previously known as the “Rationality Group,” this rebranding signals a more structured approach to developing rationality training programs and fostering research in rational decision-making. CFAR’s establishment is an important step in formalizing the application of rationality tools, which would later become a key part of the broader Effective Altruism community.[98]
2012 May 11 Outside Review Holden Karnofsky, co-founder of GiveWell and later Open Philanthropy, publishes "Thoughts on the Singularity Institute (SI)" on LessWrong. In this post, Karnofsky outlines his reasons for not recommending the Singularity Institute (now MIRI) for GiveWell funding. His concerns revolve around the speculative nature of the institute’s research on AI safety, which he believes lacks the empirical grounding necessary for confident recommendations. This review contributes to MIRI’s reputation within the broader Effective Altruism and existential risk communities, prompting them to refine their research approach.[99]
2012 August 6 Newsletter MIRI begins publishing monthly newsletters as blog posts, starting with the July 2012 Newsletter. These newsletters provide regular updates on MIRI’s research, events, and organizational developments, and serve as a valuable resource for supporters and stakeholders interested in AI safety. The monthly cadence also marks a more structured communication approach for MIRI, enhancing its transparency and engagement with the community.[100]
2012 October 13–14 Conference The Singularity Summit 2012 takes place in San Francisco, attracting a wide array of speakers and attendees, including leaders in AI, neuroscience, and futurism. Speakers such as Eliezer Yudkowsky and Ray Kurzweil share their visions of the future, discussing topics from AI safety to human enhancement. The Summit is a key event for disseminating ideas about the singularity and fostering discussions about the long-term impact of artificial intelligence on humanity.[101]
2012 November 11–18 Workshop The 1st Workshop on Logic, Probability, and Reflection takes place, bringing together researchers to explore the intersections of these fields with AI alignment and decision theory. These workshops are critical for advancing MIRI’s foundational research on how to develop AI systems that can reason reliably under uncertainty, a key component of ensuring the safety and predictability of future AI systems.[102]
2012 December 6 Singularity Summit Acquisition Singularity University announces that it has acquired the Singularity Summit from MIRI. This acquisition marks the end of MIRI's direct involvement with the summit, a move praised by some, including Joshua Fox, for allowing MIRI to focus more directly on AI safety research. However, Singularity University does not continue the Summit tradition in its original form. The conference's ethos is eventually inherited by other events like EA Global, which carry forward similar themes of long-term thinking and technological foresight.[103][104]
2013 Mission Mission Statement Update MIRI's mission statement is revised to reflect its evolving focus on AI safety: "To ensure that the creation of smarter-than-human intelligence has a positive impact. Thus, the charitable purpose of the organization is to: a) perform research relevant to ensuring that smarter-than-human intelligence has a positive impact; b) raise awareness of this important issue; c) advise researchers, leaders, and laypeople around the world; d) as necessary, implement a smarter-than-human intelligence with humane, stable goals." This shift represents a more direct approach to developing safe AI systems, incorporating a broader outreach strategy, and addressing global challenges posed by advanced AI.[105]
2013–2014 Project Conversations Initiative During this period, MIRI engages in a large number of expert interviews. Out of 80 conversations listed as of July 2017, 75 occurred in this time frame (19 in 2013 and 56 in 2014). These conversations involve in-depth discussions on AI safety, strategy, and existential risk with leading thinkers in the field. By mid-2014, MIRI deprioritizes these interviews due to diminishing returns, as noted by executive director Luke Muehlhauser in MIRI’s 2014 review. However, the conversations contribute substantially to shaping AI safety dialogue during these years.[106][107]
2013 January Staff Michael Anissimov leaves MIRI following the acquisition of the Singularity Summit by Singularity University and a major shift in MIRI's public communication strategy. Although no longer employed at MIRI, Anissimov continues to support its mission and contributes as a volunteer. This departure reflects MIRI's pivot away from broader public outreach and its increased focus on research, particularly in AI alignment and decision theory.[108]
2013 January 30 Rebranding MIRI announces its renaming from the Singularity Institute for Artificial Intelligence (SIAI) to the Machine Intelligence Research Institute (MIRI). The name change reflects MIRI's growing focus on machine intelligence and the technical challenges of AI safety, rather than the broader singularity topics associated with its former title. This rebranding helps clarify MIRI's mission to external stakeholders and aligns with its shift toward more technical and research-focused projects.[109]
2013 February 1 Publication MIRI publishes "Facing the Intelligence Explosion" by executive director Luke Muehlhauser. This book provides an accessible introduction to the risks posed by artificial intelligence and highlights the urgent need for AI safety research. It underscores MIRI's mission to address the potentially existential risks that could arise from advanced AI systems, framing the conversation around the control and alignment of AI.[110]
2013 February 11 – February 28 Domain MIRI's new website, intelligence.org, begins operating during this period. The website’s launch marks a new digital presence for the organization, with a cleaner, more professional focus on machine intelligence research and AI safety. Executive director Luke Muehlhauser announces the new site in a blog post, emphasizing the transition away from the Singularity Institute’s prior domain and approach.[111][112]
2013 April 3 Publication "Singularity Hypotheses: A Scientific and Philosophical Assessment" is published by Springer. This collection, which includes contributions from MIRI researchers and research associates, examines the scientific and philosophical issues surrounding the concept of the singularity and the rise of advanced artificial intelligence. The book provides a detailed exploration of the potential trajectories of AI development and its impact on humanity.[113][114]
2013 April 3–24 Workshop MIRI hosts the 2nd Workshop on Logic, Probability, and Reflection, bringing together researchers to advance the development of decision theory, AI alignment, and formal methods for AI reasoning. These workshops form a critical part of MIRI’s strategy for improving foundational theoretical work on AI, which is key for creating safe, reliable AI systems.[102]
2013 April 13 Strategy MIRI publishes a strategic update, outlining plans to shift its focus more heavily toward Friendly AI mathematics and reducing its emphasis on public outreach. This transition is framed as a necessary step to concentrate resources on the technical challenges that will have the most direct impact on AI safety. The organization sees this as a way to prioritize high-value research areas that can contribute to controlling advanced AI.[115]
2014 January (approximate) Financial Jed McCaleb, the creator of Ripple and original founder of Mt. Gox, donates $500,000 worth of XRP to the Machine Intelligence Research Institute (MIRI). This marks a substantial financial contribution to support AI safety research, further emphasizing the growing interest in AI development from individuals in the cryptocurrency space. McCaleb's involvement highlights the intersection of cryptocurrency and AI safety, as both fields focus on technological innovation with significant societal impacts.[116]
2014 January 16 Outside Review MIRI staff meet with Holden Karnofsky, co-founder of GiveWell, for a strategic conversation about existential risks and AI safety. The discussion focuses on MIRI’s approach to managing existential risk, exploring potential avenues for collaboration between MIRI and other organizations involved in AI safety. This meeting is part of MIRI's ongoing effort to engage with influential figures in the effective altruism and philanthropic communities to advance AI safety research.[117]
2014 February 1 Publication MIRI publishes Stuart Armstrong's influential book "Smarter Than Us: The Rise of Machine Intelligence". The book explores the challenges humanity may face with the rise of intelligent machines and serves as an introduction to AI alignment issues for a broader audience. Armstrong, a research associate at MIRI, examines the potential risks of advanced AI systems, making this book a key piece of literature in the AI safety discourse.[118]
2014 March–May Influence The Future of Life Institute (FLI) is co-founded by Max Tegmark, Jaan Tallinn, Meia Chita-Tegmark, and Anthony Aguirre, with support from MIRI. FLI is an existential risk research and outreach organization focused on ensuring the benefits of AI are shared by humanity. MIRI's influence is notable, as Tallinn, a co-founder of FLI and the Cambridge Centre for the Study of Existential Risk (CSER), cites MIRI as a key source for his views on AI risk. This marks a major expansion in global efforts to address the long-term societal impacts of AI, with MIRI playing a pivotal role in the formation of FLI.[119]
2014 March 12–13 Staff MIRI announces the hiring of several new researchers, including Nate Soares, who would later become MIRI's executive director in 2015. This marks a key moment of growth for the institute as it expands its research team. MIRI also hosts an Expansion Party to introduce the new hires to local supporters, underscoring the organization’s increased visibility and capacity to take on more ambitious AI safety projects.[120]

[121][122]

2014 May 3–11 Workshop MIRI hosts the 7th Workshop on Logic, Probability, and Reflection. This workshop focuses on advancing decision theory and addressing problems related to AI's reasoning under uncertainty. Attendees include top researchers in AI safety and decision theory, working on foundational questions crucial for creating safe AI systems.[102]
2014 July–September Influence Nick Bostrom's seminal work "Superintelligence: Paths, Dangers, Strategies" is published. Bostrom, a research advisor to MIRI, draws heavily on AI safety concerns shared by MIRI researchers. MIRI played a significant role in shaping the discussions that led to the book's publication. "Superintelligence" becomes a widely recognized book in AI alignment, contributing to global discourse on managing the risks of powerful AI systems.[123]
2014 July 4 Project Earliest evidence of the existence of AI Impacts, an initiative focused on analyzing the future societal impacts of AI, appears. Katja Grace plays a key role in launching the project, which seeks to provide rigorous research on AI timelines and impact assessments.[124]
2014 August Project The AI Impacts website officially launches. This project, led by Paul Christiano and Katja Grace, provides detailed analyses and forecasts regarding the development of AI. The website becomes a hub for discussing the potential long-term future of AI and its impacts on society, solidifying AI Impacts as a key contributor to the existential risk community.[125]
2014 November 4 Project The Intelligent Agent Foundations Forum, run by MIRI, is launched. This forum serves as a space for discussing cutting-edge research on agent foundations and decision theory, crucial components in the development of safe AI systems. The forum attracts researchers from a variety of fields to contribute to the growing body of work on AI safety and alignment.[126]
2015 January Project AI Impacts, a project focused on assessing the potential long-term impacts of artificial intelligence, rolls out a redesigned website. This project aims to provide accessible, well-researched information on AI risks, timelines, and governance issues. The site’s overhaul, led by the Machine Intelligence Research Institute (MIRI), was part of a broader effort to improve public engagement and the dissemination of knowledge about AI’s potential dangers.[127]
2015 January 2–5 Conference The Future of AI: Opportunities and Challenges, an AI safety conference, takes place in Puerto Rico. Organized by the Future of Life Institute, the conference attracts influential researchers such as Luke Muehlhauser, Eliezer Yudkowsky, and Nate Soares from MIRI, as well as top AI academics. This event became pivotal in rallying attention to AI risks, leading Soares to describe it as a “turning point” where academia began seriously addressing AI existential risk. At this conference, leading thinkers discussed how AI, if left unchecked, could pose existential threats to humanity.[128][129]
2015 March 11 Influence Rationality: From AI to Zombies is published. This book, a compilation of Eliezer Yudkowsky's influential blog series "The Sequences" from the LessWrong community, explores rational thinking and decision-making, blending topics from AI development to human psychology. It became a key philosophical text within the Effective Altruism and rationality movements, widely regarded as a comprehensive introduction to AI alignment challenges and human cognitive biases.[61][130]
2015 May 4–6 Workshop The 1st Introductory Workshop on Logical Decision Theory takes place. This workshop is designed to educate researchers on decision theories that take into account AI's capacity to predict and influence decisions, aiming to tackle problems like Newcomb's paradox in AI alignment.[102]
2015 May 6 Staff Luke Muehlhauser announces his resignation as MIRI’s executive director, moving to the Open Philanthropy Project as a research analyst. In his farewell post, Muehlhauser expresses confidence in his successor, Nate Soares, who has been a key researcher at MIRI. Soares, known for his work on decision theory and AI safety, takes over as MIRI's executive director.[131]
2015 May 13–19 Conference In collaboration with the Centre for the Study of Existential Risk (CSER), MIRI co-organizes the Self-prediction in Decision Theory and Artificial Intelligence Conference. The event brings together experts to explore the implications of self-prediction in decision theory, which has major relevance to AI systems’ decision-making capabilities and how they predict their own actions.[132]
2015 May 29–31 Workshop The 1st Introductory Workshop on Logical Uncertainty is held, focusing on how AI systems deal with uncertainty in logic-based reasoning, a fundamental challenge in ensuring that AI systems can make reliable decisions in uncertain environments.[102]
2015 June 3–4 Staff Nate Soares officially begins as the executive director of MIRI. Soares, who previously worked on decision theory and AI alignment, steps into this leadership role with the goal of pushing MIRI’s research agenda towards solving AI’s long-term safety challenges.[133]
2015 June 11 AMA Nate Soares, MIRI's executive director, hosts an "ask me anything" (AMA) on the Effective Altruism Forum, engaging the community on topics ranging from AI alignment to his vision for MIRI’s future.[134]
2015 June 12–14 Workshop The 2nd Introductory Workshop on Logical Decision Theory takes place, building on the first workshop’s success by providing advanced tutorials on decision-making theories relevant to AI alignment.[102]
2015 June 26–28 Workshop The 1st Introductory Workshop on Vingean Reflection is held, focusing on how an AI system can reflect on and modify its own decision-making procedures in a safe and predictable manner.[102]
2015 July 7–26 Project The MIRI Summer Fellows Program 2015, run by the Center for Applied Rationality (CFAR), is held. This fellowship aims to cultivate new talent for MIRI’s AI safety research, and it is described as "relatively successful" at recruiting new staff members.[135][136]
2015 August 7–9 Workshop The 2nd Introductory Workshop on Logical Uncertainty takes place, continuing the discussion on how AI systems can make reliable decisions under uncertainty, which is critical to ensuring AI safety in complex, real-world environments.[102]
2015 August 28–30 Workshop The 3rd Introductory Workshop on Logical Decision Theory is held, focusing on refining decision-making frameworks for AI systems. Attendees delve deeper into logical decision theories, specifically how AI agents can navigate decision-making scenarios with incomplete information, ensuring robustness and safety.[102]
2015 September 26 External Review The Effective Altruism Wiki page on MIRI is created. This page provides an overview of the Machine Intelligence Research Institute's work and its mission to reduce existential risks associated with artificial intelligence, making its projects and goals more accessible to the Effective Altruism community.[137]
2016 Publication MIRI commissions Eliezer Yudkowsky to produce AI alignment content for Arbital, a platform that sought to explain complex technical concepts in a way accessible to a broader audience. The goal of this project was to provide more detailed educational materials on AI safety and alignment, addressing various AI risk topics. Arbital was envisioned as a solution for breaking down difficult technical topics related to AI risk for readers of all levels.[138][139]
2016 March 30 Staff MIRI announces the promotion of two key staff members. Malo Bourgon, who had been serving as a program management analyst, steps into the role of Chief Operating Officer (COO). Additionally, Rob Bensinger, previously an outreach coordinator, becomes the Research Communications Manager. This internal reshuffle signals a strengthening of MIRI’s operational and research communications capacities as it expands its AI alignment work.[140]
2016 April 1–3 Workshop The Self-Reference, Type Theory, and Formal Verification Workshop takes place. This workshop focused on advancing formal methods in AI, particularly in how self-referential AI systems can be verified to ensure they behave in alignment with human values. Type theory and formal verification are essential areas in AI safety, ensuring that AI systems can reason about their own decisions safely.[102]
2016 May 6 (talk), December 28 (transcript release) Publication In May 2016, Eliezer Yudkowsky gives a talk titled "AI Alignment: Why It’s Hard, and Where to Start" at Stanford University. Yudkowsky discusses the technical difficulties in aligning AI systems with human values, drawing attention to the challenges involved in controlling advanced AI systems. An edited version of this transcript is released on the MIRI blog in December 2016, where it becomes a key reference for researchers working on AI safety.[141][142]
2016 May 28–29 Workshop The Colloquium Series on Robust and Beneficial AI (CSRBAI) Workshop on Transparency takes place. This event focuses on the importance of transparency in AI systems, particularly how to ensure that advanced AI systems are interpretable and understandable by humans, which is critical to ensuring safe AI alignment.[102]
2016 June 4–5 Workshop The Colloquium Series on Robust and Beneficial AI (CSRBAI) Workshop on Robustness and Error-Tolerance takes place. The focus of this workshop is on developing AI systems that are robust to errors and can tolerate uncertainty, further contributing to safe deployment of AI systems in unpredictable real-world environments.[102]
2016 June 11–12 Workshop The Colloquium Series on Robust and Beneficial AI (CSRBAI) Workshop on Preference Specification is held. The workshop deals with the critical task of correctly specifying human preferences in AI systems, an essential aspect of AI alignment to ensure that the systems act in ways that reflect human values.[102]
2016 June 17 Workshop The Colloquium Series on Robust and Beneficial AI (CSRBAI) Workshop on Agent Models and Multi-Agent Dilemmas takes place, focusing on how AI systems can interact safely in multi-agent scenarios where the goals of different systems might conflict. This research is crucial for building AI systems that can cooperate or avoid harmful competition.[102]
2016 July 27 Publication MIRI announces its new technical agenda with the release of the paper "Alignment for Advanced Machine Learning Systems". The paper outlines the necessary steps for ensuring machine learning systems are aligned with human values as they become increasingly powerful. This agenda sets the course for MIRI’s future research efforts on machine learning and AI safety.[143]
2016 August Financial Open Philanthropy awards MIRI a $500,000 grant for general support. Despite reservations about MIRI’s technical research, the grant is awarded to support MIRI’s broader mission of reducing AI-related risks. This grant illustrates Open Philanthropy’s acknowledgment of the importance of MIRI’s work on AI alignment, despite differing opinions on technical approaches.[144]
2016 August 12–14 Workshop The 8th Workshop on Logic, Probability, and Reflection is held, continuing MIRI’s tradition of exploring how logic and probability can be used to reason about self-reflection in AI systems. This is a critical aspect of building AI systems capable of safely understanding their own behavior and decision-making processes.[102]
2016 August 26–28 Workshop The 1st Workshop on Machine Learning and AI Safety is held. This inaugural event focuses on the emerging field of AI safety in the context of machine learning, emphasizing the need for alignment in rapidly evolving machine learning models.[102]
2016 September 12 Publication MIRI releases a landmark paper titled "Logical Induction" by Scott Garrabrant, Tsvi Benson-Tilsen, Andrew Critch, Nate Soares, and Jessica Taylor. The paper presents a novel approach to reasoning under uncertainty, solving a long-standing problem in logic and opening new possibilities for ensuring safe AI reasoning. The paper is widely praised, with some calling it a "major breakthrough" in formal AI research.[145][146]
2016 October 12 AMA MIRI hosts an "Ask Me Anything" (AMA) session on the Effective Altruism Forum, giving the community an opportunity to ask questions about MIRI’s work, AI alignment, and related technical research. Rob Bensinger, Nate Soares, and other Mexecutives participate to discuss ongoing projects and research approaches in AI alignment and safety.[147]
2016 October 21–23 Workshop The 2nd Workshop on Machine Learning and AI Safety is held. The event continues from the first workshop earlier in the year, with a greater focus on understanding how to make machine learning systems safer as they grow in complexity. Topics discussed include adversarial training, model interpretability, and alignment techniques for machine learning models.[102]
2016 November 11–13 Workshop The 9th Workshop on Logic, Probability, and Reflection is held. This workshop delves further into how AI systems can use logical reasoning to improve decision-making under uncertainty. This remains a cornerstone of MIRI's approach to AI safety, where the focus is on creating systems that can handle complex real-world scenarios with logical consistency.[102]
2016 December Financial Open Philanthropy awards a $32,000 grant to AI Impacts, a project that aims to understand and evaluate the long-term risks of advanced artificial intelligence. The grant supports AI Impacts’ research and its efforts to provide clearer timelines and risk assessments of AI development.[148]
2016 December 1–3 Workshop The 3rd Workshop on Machine Learning and AI Safety is held, capping off a year of significant progress in AI safety research. This workshop provides an opportunity for researchers to reflect on the advancements made throughout the year and to identify new challenges for machine learning systems as AI capabilities expand.[102]
2017 March 25–26 Workshop The Workshop on Agent Foundations and AI Safety takes place. This workshop focuses on exploring foundational questions in AI safety, particularly the design of highly reliable agents that can reason under uncertainty and avoid catastrophic behaviors. Discussions center on robust agent design, decision theory, and safe AI deployment strategies.[102]
2017 April 1–2 Workshop The 4th Workshop on Machine Learning and AI Safety takes place, continuing to build upon previous workshops' discussions on ensuring machine learning models are aligned with human values. Topics include improving adversarial robustness, preventing unintended consequences from AI systems, and safe reinforcement learning. The goal is to ensure that as AI systems become more complex, they do not act unpredictably.[102]
2017 May 24 Publication The influential paper "When Will AI Exceed Human Performance? Evidence from AI Experts" is published on arXiv. This paper surveys AI experts to estimate when AI systems will outperform humans in various tasks. Two researchers from AI Impacts are co-authors. The paper gains widespread media attention, with over twenty news outlets discussing its implications for AI timelines and the potential risks associated with AI surpassing human intelligence.[149][150]
2017 July 4 Strategy MIRI announces a strategic shift, stating that it will be scaling back efforts on its "Alignment for Advanced Machine Learning Systems" agenda. This is due to the departure of key researchers Patrick LaVictoire and Jessica Taylor, and Andrew Critch taking leave. As a result, MIRI refocuses its research priorities while maintaining its commitment to AI safety.[151]
2017 July 7 Outside Review Daniel Dewey, a program officer at Open Philanthropy, publishes a post titled "My Current Thoughts on MIRI's Highly Reliable Agent Design Work". Dewey presents a critique of MIRI's approach to AI safety, arguing that while highly reliable agent design is an important area of research, other approaches such as learning to reason from humans may offer more promising paths to AI alignment. His post provides valuable insight into ongoing debates about AI safety strategies.[152]
2017 July 14 Outside Review The timelines wiki page on MIRI is publicly circulated. This wiki page documents the historical developments of MIRI's work, making it a valuable resource for understanding the evolution of AI safety research at the institute.
2017 October 13 Publication The paper "Functional Decision Theory: A New Theory of Instrumental Rationality" by Eliezer Yudkowsky and Nate Soares is posted on arXiv. This paper introduces Functional Decision Theory (FDT), a new framework for AI decision-making that differs from traditional decision theories. The authors argue that FDT offers better solutions to certain types of decision problems and could lead to safer AI systems. This paper marks a significant contribution to AI alignment research.[153][154]
2017 October 13 Publication Eliezer Yudkowsky publishes the blog post "There’s No Fire Alarm for Artificial General Intelligence" on the MIRI blog and on the newly relaunched LessWrong platform. In this post, Yudkowsky argues that there will be no clear "warning" or fire alarm to signal the arrival of AGI, making it crucial to prepare for AGI's development ahead of time. This post sparks significant discussion in the AI safety community.[155][156]
2017 October Financial Open Philanthropy awards MIRI a $3.75 million grant over three years ($1.25 million per year). The decision to award the grant is partly due to the positive reception of MIRI's "Logical Induction" paper, as well as the increased number of grants Open Philanthropy had made in the area of AI safety, allowing them to provide support to MIRI without it appearing as an outsized endorsement of one approach. This grant is a major financial boost for MIRI, enabling them to continue their work on AI safety and alignment.[157][158]
2017 November 16 Publication Eliezer Yudkowsky's sequence/book Inadequate Equilibria is fully published. The book was published chapter-by-chapter on LessWrong 2.0 and the Effective Altruism Forum starting October 28.[159][160][161] The book is reviewed on multiple blogs including Slate Star Codex (Scott Alexander),[162] Shtetl-Optimized (Scott Aaronson),[163] and Overcoming Bias (Robin Hanson).[164] The book outlines Yudkowsky's approach to epistemology, covering topics such as whether to trust expert consensus and whether one can expect to do better than average.
2017 November 25, November 26 Publication A two-part series "Security Mindset and Ordinary Paranoia" and "Security Mindset and the Logistic Success Curve" by Eliezer Yudkowsky is published. The series uses the analogy of "security mindset" to highlight the importance and non-intuitiveness of AI safety. This is based on Eliezer Yudkowsky's 2016 talk "AI Alignment: Why It’s Hard, and Where to Start."[165][166]
2017 December 1 Financial MIRI's 2017 fundraiser begins. The announcement post describes MIRI's fundraising targets, recent work at MIRI (including recent hires), and MIRI's strategic background (which gives a high-level overview of how MIRI's work relates to long-term outcomes).[167] The fundraiser would conclude with $2.5 million raised from over 300 distinct donors. The largest donation would be from Vitalik Buterin ($763,970 worth of Ethereum).[168]
2018 February Workshop MIRI and the Center for Applied Rationality (CFAR) conduct the first AI Risk for Computer Scientists (AIRCS) workshop. This would be the first of several AIRCS workshops, with seven more in 2018 and many more in 2019.[169] The page about AIRCS says: "The material at the workshop is a mixture of human rationality content that’s loosely similar to some CFAR material, and a variety of topics related to AI risk, including thinking about forecasting, different people’s ideas of where the technical problems are, and various potential paths for research."[170]
2018 October 29 Project The launch of the AI Alignment Forum (often abbreviated to just "Alignment Forum") is announced on the MIRI blog. The Alignment Forum is built and maintained by the LessWrong 2.0 team (which is distinct from MIRI), but with help from MIRI. The Alignment Forum replaces MIRI's existing Intelligent Agent Foundations Forum, and is intended as "a single online hub for alignment researchers to have conversations about all ideas in the field".[171][172] The Alignment Forum had previously launched in beta on July 10, 2018,[173] with the day of launch chosen as the first "AI Alignment Writing Day" for the MIRI Summer Fellows Program (beginning an annual tradition).[174]
2018 October 29 – November 15 Publication The Embedded Agency sequence, by MIRI researchers Abram Demski and Scott Garrabrant, is published on the MIRI blog (text version),[175] on LessWrong 2.0 (illustrated version),[176] and on the Alignment Forum (illustrated version)[177] in serialized installments from October 29 to November 8; on November 15 a full-text version containing the entire sequence is published.[178] The term "embedded agency" is a renaming of an existing concept researched at MIRI, called "naturalized agency".[179]
2018 November 22 Strategy Nate Soares, executive director of MIRI, publishes MIRI's 2018 update post (the post was not written exclusively by Soares; see footnote 1, which begins "This post is an amalgam put together by a variety of MIRI staff"). The post describes new research directions at MIRI (which are not explained in detail due to MIRI's nondisclosure policy); explains the concept of "deconfusion" and why MIRI values it; announces MIRI's "nondisclosed-by-default" policy for most of its research; and gives a recruitment pitch for people to join MIRI.[180]
2018 November 26 Financial MIRI's 2018 fundraiser begins.[169] The fundraiser would conclude on December 31 with $951,817 raised from 348 donors.[181]
2018 August (joining) November 28 (announcement), December 1 (AMA) Staff MIRI announces that prolific Haskell developer Edward Kmett has joined.[182] Kmett participates in an Ask Me Anything (AMA) on Reddit's Haskell subreddit on December 1, 2018. In reply to questions, he clarifies that MIRI's nondisclosure policy will not affect the openness of his work, but as the main researcher at MIRI who publishes openly, he will feel more pressure to produce higher-quality work as the whole organization may be judged by the quality of his work.[183]
2018 December 15 Publication MIRI announces a new edition of Eliezer Yudkowsky's Rationality: From AI to Zombies (i.e. the book version of "the Sequences"). At the time of the announcement, the new edition of only two sequences, Map and Territory and How to Actually Change Your Mind, are available.[184][185]
2019 February Financial Open Philanthropy grants MIRI $2,112,500 over two years. The grant amount is decided by the Committee for Effective Altruism Support, which also decides on amounts for grants to 80,000 Hours and the Centre for Effective Altruism at around the same time.[186] The Berkeley Existential Risk Initiative (BERI) grants $600,000 to MIRI at around the same time. MIRI discusses both grants in a blog post.[187]
2019 April 23 Financial The Long-Term Future Fund announces that it is donating $50,000 to MIRI as part of this grant round. Oliver Habryka, the main grant investigator, explains the reasoning in detail, including his general positive impression of MIRI and his thoughts on funding gaps.[188]
2019 December Financial MIRI's 2019 fundraiser raises $601,120 from over 259 donors. A retrospective blog post on the fundraiser, published February 2020, discusses possible reasons the fundraiser raised less money than fundraisers in previous years, particularly 2017. Reasons include: lower cryptocurrency prices causing fewer donations from cryptocurrency donors, nondisclosed-by-default policy making it harder for donors to evaluate research, US tax law changes in 2018 causing more donation-bunching across years, fewer counterfactual matching opportunities, donor perception of reduced marginal value of donations, skew in donations from a few big donors, previous donors moving from earning-to-give to direct work, and donors responding to MIRI's urgent need for funds in previous years by donating in those years and having less to donate now.[189]
2020 February Financial Open Philanthropy grants $7,703,750 to MIRI over two years, with the money amount determined by the Committee for Effective Altruism Support (CEAS). Of the funding, $6.24 million comes from Good Ventures (the usual funding source) and $1.46 milion comes from Ben Delo, co-founder of BitMEX and recent Giving Pledgee signatory, via a co-funding partnership. Other organizations receiving money based on CEAS recommendations at around the same time are Ought (also focused on AI safety), the Centre for Effective Altruism, and 80,000 Hours.[190] MIRI would blog about the grant in April 2020, calling the grant "our largest grant to date."[191]
2020 March 2 Financial The Berkeley Existential Risk Initiative (BERI) grants $300,000 to MIRI. Writing about the grant in April 2020, MIRI says: "at the time of our 2019 fundraiser, we expected to receive a grant from BERI in early 2020, and incorporated this into our reserves estimates. However, we predicted the grant size would be $600k; now that we know the final grant amount, that estimate should be $300k lower."[191]
2020 April 14 Financial The Long-Term Future Fund grants $100,000 to MIRI.[192][191]
2020 May Financial The Survival and Flourishing Fund publishes the outcome of its recommendation S-process for the first half of 2020. This includes three grant recommendations to MIRI: $20,000 from SFF, $280,000 from Jaan Tallinn, and $40,000 from Jed McCaleb.[193] The grant from SFF to MIRI would also be included in SFF's grant list with a grant date of May 2020.[194]
2020 October 9 A Facebook post by Rob Bensinger, MIRI's research communications manager, says that MIRI is considering moving its office from its current location in Berkeley, California (in the San Francisco Bay Area) to another location in the United States or Canada. Two areas under active consideration are the Northeastern US (New Hampshire in particular) and the area surrounding Toronto. In response to a question about reasons, Bensinger clarifies that he cannot disclose reasons yet, but that he wanted to announce preemptively so that people can factor this uncertainty into any plans to move or to start new rationalist hubs.[195]
2020 October 22 Publication Scott Garrabrant publishes (cross-posted to LessWrong and the Effective Altruism Forum) a blog post titled "Introduction to Cartesian Frames" that is is first post in a sequence about Cartesian frames, a new conceptual framework for thinking about agency.[196][197]
2020 November (announcement) Financial Jaan Tallinn grants $543,000 to MIRI as an outcome of the S-process carried out by the Survival and Flourishing Fund for the second half of 2020.[198]
2020 November 30 (announcement) Financial In the November newsletter, MIRI announces that it will not be running a formal fundraiser this year, but that it will continue participating in Giving Tuesday and other matching opportunities.[199]
2020 December 21 Strategy Malo Bourgon publishes MIRI's "2020 Updates and Strategy" blog post. The post talks about MIRI's efforts to relocate staff after the COVID-19 pandemic as well as the generally positive result of the changes, and possible future implications for MIRI itself moving out of the Bay Area. It also talks about slow progress on the research directions initiated in 2017, leading to MIRI feeling the need to change course. The post also talks about the public part of MIRI's progress in other research areas.[200]
2021 May 8 Rob Bensinger publishes a post on LessWrong providing an update on MIRI's ongoing considerations regarding relocation from the San Francisco Bay Area. The post opens a community discussion but lacks a definitive conclusion or subsequent actions, reflecting some internal uncertainty.[201]
2021 May 13 Financial MIRI announces two major donations: $15,592,829 in MakerDAO (MKR) from an anonymous donor with a restriction to spend a maximum of $2.5 million per year till 2024, and 1050 ETH from Vitalik Buterin, worth $4,378,159. While this was one of MIRI's largest donations, the restrictions on the use of funds limited their ability to make immediate strategic investments.[202]
2021 May 23 In a talk, MIRI researcher Scott Garrabrandt introduces "finite factored sets" as an alternative to the Pearlian paradigm. The concept generates some interest among the AI safety community, particularly on LessWrong, but does not significantly shift the broader landscape of causal inference research.[203]
2021 July 1 An update is added to Rob Bensinger's May 8 post about MIRI's potential relocation. The update links to a comment by board member Blake Borgeson, who had been tasked with coordinating MIRI's relocation decision. Ultimately, MIRI decides against relocating, citing uncertainty about long-term strategy. This decision indicates hesitation and a conservative approach amid organizational ambiguity.[201]
2021 November 15 Several private conversations between MIRI researchers (Eliezer Yudkowsky, Nate Soares, Rob Bensinger) and others in the AI safety community are published to the Alignment Forum and cross-posted to LessWrong and the Effective Altruism Forum. These conversations, titled "Late 2021 MIRI Conversations," attract moderate attention and foster some debate, particularly within niche AI safety circles, but do not significantly influence broader community consensus.[204]
2021 November 29 MIRI announces on the Alignment Forum that it is seeking assistance with its Visible Thoughts Project. Despite offering bounties for contributions, the project does not attract significant participation, indicating either a lack of interest or challenges in community engagement.[205]
2021 December Financial MIRI offers $200,000 to build an AI-dungeon-style writing dataset annotated with thoughts, and an additional $1,000,000 for scaling it 10x. The Visible Thoughts Project, while promising substantial incentives, struggles with engagement issues and fails to yield the expected contributions and outputs.[206]
2022 May 30 Publication Eliezer Yudkowsky publishes "Six Dimensions of Operational Adequacy in AGI Projects" on LessWrong. The post sparks some discussion among AI safety researchers but does not establish new standards or practices across broader AGI safety projects.[207]
2022 June 5 Publication Eliezer Yudkowsky's article "AGI Ruin: A List of Lethalities" is published on LessWrong. The post receives significant attention within the alignment community and reiterates Yudkowsky’s longstanding concerns about catastrophic AGI risks. It sparks debate, but the influence is largely confined to existing followers rather than drawing in broader public discourse.[208]
2022 April 25 Publication The article "Visible Thoughts Project and Bounty Announcement" is republished on LessWrong. Despite the sizable financial incentives offered, participation in the project remains low, and MIRI struggles to generate the expected level of interest and meaningful output.[209]
2022 July Strategy MIRI pauses its newsletter and public communications to refine internal strategies, an indication of both internal challenges and an effort to recalibrate its approach amid a rapidly evolving AI landscape.[210]
2022 December 1 Publication On behalf of his MIRI colleagues, Rob Bensinger publishes a blog post challenging organizations such as Anthropic and DeepMind to publicly write up their alignment plans. The challenge generates a mixed response, with some critiques of OpenAI’s plans emerging, but it does not spur any major public commitment from these organizations.[211]
2023 February 20 Publication Eliezer Yudkowsky appears on the Bankless podcast for an interview lasting a little under two hours, where he shares his relatively pessimistic views about the likelihood of catastrophic AGI with his hosts, neither of whom is deep into AI safety.[212] He also mentions he is taking a sabbatical due to burnout and the inevitable doom. He mentions considering potential ideas of working with other organizations such as Anthropic, Conjecture, or Redwood Research, noting that Redwood Research is "small" but that he trusts them and that they can also focus on one stream. A full transcript is published to LessWrong and the Alignment Forum a few days later.[213] The podcast gets a lot of traction, eliciting several reactions, and leads to a follow-up Q&A on Twitter Spaces.[214] A month later, a lengthy point-by-point response by alignment researcher Quintin Pope is published to LessWrong, attracting over 200 comments.[215]
2023 March 29 Publication An article by Eliezer Yudkowsky in Time Ideas, in response to the FLI Open Letter, argues that pausing AI for six months isn't enough. He says that what is needed won't happen in practice, but spells it out anyway: "The moratorium on new large training runs needs to be indefinite and worldwide. There can be no exceptions, including for governments or militaries. [...] Shut down all the large GPU clusters (the large computer farms where the most powerful AIs are refined). Shut down all the large training runs. Put a ceiling on how much computing power anyone is allowed to use in training an AI system, and move it downward over the coming years to compensate for more efficient training algorithms. No exceptions for governments and militaries. Make immediate multinational agreements to prevent the prohibited activities from moving elsewhere. [...] Frame nothing as a conflict between national interests, have it clear that anyone talking of arms races is a fool. [...] Shut it all down."[216] The post is shared to LessWrong where it receives over 250 comments.[217]
2023 April Leadership MIRI undergoes a significant leadership change, with Malo Bourgon appointed as CEO, Nate Soares transitioning to President, Alex Vermeer becoming COO, and Eliezer Yudkowsky assuming the role of Chair of the Board. This restructuring is seen by some as an attempt to address stagnation and operational challenges.[218]
2023 June 19 Publication Paul Christiano publishes an article titled "Where I Agree and Disagree with Eliezer" on the AI Alignment Forum, outlining areas of alignment and divergence with Eliezer Yudkowsky's perspectives. The article is well-received within AI alignment circles and generates a productive debate, but does not directly influence the wider public narrative around AI safety.[219]
2024 Jan 14 Strategy MIRI publishes a comprehensive update on its mission and strategy for 2024. The update reaffirms their approach to AI alignment research and emphasizes collaboration. While the update receives positive feedback within existing networks, it does not attract wider attention or lead to notable changes in AI safety practices.[220][221]
2024 March 9 Publication An article in Semafor titled "The Risks of Expanding the Definition of AI Safety" discusses concerns raised by Eliezer Yudkowsky about the broadening scope of AI safety. While the article garners attention within specialized AI safety and alignment circles, it does not significantly alter the public narrative around AI governance, reflecting its niche impact.[222]
2024 April Project MIRI launches a new research team dedicated to technical AI governance. The team, initially consisting of Lisa Thiergart and Peter Barnett, aims to expand by the end of the year. Early traction is limited, highlighting recruitment challenges and the evolving demands of governance work in a rapidly changing AI landscape.[223]
2024 May Project The Technical Governance Team at MIRI takes an active role in contributing to AI policy development by submitting responses to multiple key policy bodies. These submissions include the NTIA's request for comment on open-weight AI models, focusing on the implications of making AI model weights publicly available and the potential risks and benefits associated with open-access AI technology.[224] They also respond to the United Nations’ request for feedback on the "Governing AI for Humanity" interim report, offering insights on global AI governance frameworks and how they can be structured to prioritize safety, transparency, and ethical considerations.[225] Additionally, the team addresses the Office of Management and Budget’s request for information on AI procurement in government, providing recommendations on how AI technologies can be integrated responsibly within government infrastructures.[226] This proactive engagement highlights MIRI’s strategic involvement in shaping international AI governance and ensuring that safety and ethical standards are maintained in the development and use of AI technologies.[227]
2024 May 14 Project MIRI announces the shutdown of the Visible Thoughts Project, which was initiated in November 2021. The project faced several challenges, including evolving ML needs and limited community engagement, which ultimately led to its termination.[228]
2024 May 29 Publication MIRI publishes their 2024 Communications Strategy, focusing on halting the development of frontier AI systems worldwide. The strategy aims for direct, unvarnished communication with policymakers and the public. However, the approach avoids grassroots advocacy and receives mixed reactions, with limited evidence of a shift in AI policy or public sentiment.[229]
2024 June 7 Publication Rob Bensinger publishes a response to Daniel Kokotajlo's discussion of Aschenbrenner's views on situational awareness in AI. Bensinger critiques Kokotajlo’s interpretation, adding nuance to the debate on AI safety. While the discussion is valuable within the alignment community, it remains niche and does not lead to broader shifts in consensus.[230]
2024 June Research The Agent Foundations team, including Scott Garrabrant, departs MIRI to pursue independent work. This signals a shift in focus for MIRI, as they prioritize other areas in response to rapid AI advancements. The departure is seen as an outcome of MIRI reassessing its research priorities amid changing circumstances in the AI field.[231]

Numerical and visual data

Google Scholar

The following table summarizes per-year mentions on Google Scholar as of October 1, 2021.

Year "Machine Intelligence Research Institute"
2000 0
2001 2
2002 0
2003 0
2004 1
2005 0
2006 0
2007 1
2008 0
2009 5
2010 7
2011 6
2012 6
2013 29
2014 61
2015 72
2016 93
2017 128
2018 134
2019 127
2020 138
2021 120
2022 117
2023 122
MIRI GS.png

Google Trends

The comparative chart below shows Google Trends data Machine Intelligence Research Institute (Research institute) and Machine Intelligence Research Institute (Search term), from January 2004 to November 2024, when the screenshot was taken. Interest is also ranked by country and displayed on world map.[232]

Machine Intelligence Research Institute gt1.png

Google Ngram Viewer

The chart below shows Google Ngram Viewer data for Machine Intelligence Research Institute, from 2000 to 2024.[233]

Machine Intelligence Research Institute ngram1.png

Wikipedia desktop pageviews across the different names

The image below shows desktop pageviews of the page Machine Intelligence Research Institute and its predecessor pages, "Singularity Institute for Artificial Intelligence" and "Singularity Institute".[234] The change in names occurred on these dates:[235][236]

  • December 23, 2011: Two pages "Singularity Institute" and "Singularity Institute for Artificial Intelligence" merged into single page "Singularty Institute for Artificial Intelligence"
  • April 16, 2012: Page moved from "Singularity Institute for Artificial Intelligence" to "Singularity Institute" with the old name redirecting to the new name
  • February 1, 2013: Page moved from "Singularity Institute" to "Machine Intelligence Research Institute" with both old names redirecting to the new name

The red vertical line (for June 2015) represents a change in the method of estimating pageviews; specifically, pageviews by bots and spiders are excluded for months on the right of the line.

MIRI wv.jpeg

Meta information on the timeline

How the timeline was built

The initial version of the timeline was written by Issa Rice.

Issa likes to work locally and track changes with Git, so the revision history on this wiki only shows changes in bulk. To see more incremental changes, refer to the commit history.

Funding information for this timeline is available.

Feedback and comments

Feedback for the timeline can be provided at the following places:

What the timeline is still missing

  • TODO Figure out how to cover publications
  • TODO mention kurzweil
  • TODO maybe include some of the largest donations (e.g. the XRP/ETH ones, tallinn, thiel)
  • TODO maybe fundraisers
  • TODO look more closely through some AMAs: [1], [2]
  • TODO maybe more info in this SSC post [3]
  • TODO more links at EA Wikia page [4]
  • TODO lots of things from strategy updates, annual reviews, etc. [5]
  • TODO Ben Goertzel talks about his involvement with MIRI [6], also more on opencog
  • TODO giant thread on Ozy's blog [7]
  • NOTE From 2017-07-06: "years that have few events so far: 2003 (one event), 2007 (one event), 2008 (three events), 2010 (three events), 2017 (three events)"
  • TODO possibly include more from the old MIRI volunteers site. Some of the volunteering opportunities like proofreading and promoting MIRI by giving it good web of trust ratings seem to give a good flavor of what MIRI was like, the specific challenges in terms of switching domains, and so on.
  • TODO cover Berkeley Existential Risk Initiative (BERI), kinda a successor to MIRI volunteers?
  • TODO cover launch of Center for Human-Compatible AI
  • TODO not sure how exactly to include this in the timeline, but something about MIRI's changing approach to funding certain types of contract work. e.g. Vipul says "I believe the work I did with Luke would no longer be sponsored by MIRI as their research agenda is now much more narrowly focused on the mathematical parts."
  • TODO who is Tyler Emerson?
  • modal combat and some other domains: [8], [9], [10]
  • https://www.lesswrong.com/posts/yGZHQYqWkLMbXy3z7/video-q-and-a-with-singularity-institute-executive-director
  • https://ea.greaterwrong.com/posts/NBgpPaz5vYe3tH4ga/on-deference-and-yudkowsky-s-ai-risk-estimates

Timeline update strategy

Some places to look on the MIRI blog:

Also general stuff like big news coverage.

See also

External links

References

  1. ↑ Nate Soares (June 3, 2015). "Taking the reins at MIRI". LessWrong. Retrieved July 5, 2017. 
  2. ↑ "lukeprog comments on "Thoughts on the Singularity Institute"". LessWrong. May 10, 2012. Retrieved July 15, 2012. 
  3. ↑ "Halfwitz comments on "Breaking the vicious cycle"". LessWrong. November 23, 2014. Retrieved August 3, 2017. 
  4. ↑ Eliezer S. Yudkowsky (August 31, 2000). "Eliezer, the person". Archived from the original on February 5, 2001. 
  5. ↑ "Yudkowsky - Staring into the Singularity 1.2.5". Retrieved June 1, 2017. 
  6. ↑ Eliezer S. Yudkowsky. "Coding a Transhuman AI". Retrieved July 5, 2017. 
  7. ↑ Eliezer S. Yudkowsky. "Singularitarian mailing list". Retrieved July 5, 2017. The "Singularitarian" mailing list was first launched on Sunday, March 11th, 1999, to assist in the common goal of reaching the Singularity. It will do so by pooling the resources of time, brains, influence, and money available to Singularitarians; by enabling us to draw on the advice and experience of the whole; by bringing together individuals with compatible ideas and complementary resources; and by binding the Singularitarians into a community. 
  8. ↑ 8.0 8.1 8.2 Eliezer S. Yudkowsky. "PtS: Version History". Retrieved July 4, 2017. 
  9. ↑ "Yudkowsky's Coming of Age". LessWrong. Retrieved January 30, 2018. 
  10. ↑ "My Naturalistic Awakening". LessWrong. Retrieved January 30, 2018. 
  11. ↑ "jacob_cannell comments on FLI's recommended project grants for AI safety research announced". LessWrong. Retrieved January 30, 2018. 
  12. ↑ Eliezer S. Yudkowsky. "Singularitarian Principles 1.0". Retrieved July 5, 2017. 
  13. ↑ "SL4: By Date". Retrieved June 1, 2017. 
  14. ↑ Eliezer S. Yudkowsky. "SL4 Mailing List". Retrieved June 1, 2017. 
  15. ↑ 15.0 15.1 Eliezer S. Yudkowsky. "Coding a Transhuman AI § Version History". Retrieved July 5, 2017. 
  16. ↑ "Form 990-EZ 2000" (PDF). Retrieved June 1, 2017. Organization was incorporated in July 2000 and does not have a financial history for years 1996-1999. 
  17. ↑ "About the Singularity Institute for Artificial Intelligence". Retrieved July 1, 2017. The Singularity Institute for Artificial Intelligence, Inc. (SIAI) was incorporated on July 27th, 2000 by Brian Atkins, Sabine Atkins (then Sabine Stoeckel) and Eliezer Yudkowsky. The Singularity Institute is a nonprofit corporation governed by the Georgia Nonprofit Corporation Code, and is federally tax-exempt as a 501(c)(3) public charity. At this time, the Singularity Institute is funded solely by individual donors. 
  18. ↑ Eliezer S. Yudkowsky. "Singularity Institute for Artificial Intelligence, Inc.". Retrieved July 4, 2017. 
  19. ↑ Eliezer S. Yudkowsky. "Singularity Institute: News". Retrieved July 1, 2017. April 08, 2001: The Singularity Institute for Artificial Intelligence, Inc. announces that it has received tax-exempt status and is now accepting donations. 
  20. ↑ "Singularity Institute for Artificial Intelligence // News // Archive". Retrieved July 13, 2017. 
  21. ↑ Singularity Institute for Artificial Intelligence. "SIAI Guidelines on Friendly AI". Retrieved July 13, 2017. 
  22. ↑ Eliezer Yudkowsky (2001). "Creating Friendly AI 1.0: The Analysis and Design of Benevolent Goal Architectures" (PDF). The Singularity Institute. Retrieved July 5, 2017. 
  23. ↑ 23.0 23.1 23.2 23.3 23.4 Eliezer S. Yudkowsky. "Singularity Institute: News". Retrieved July 1, 2017. 
  24. ↑ "SL4: By Thread". Retrieved July 1, 2017. 
  25. ↑ Eliezer S. Yudkowsky (April 7, 2002). "SL4: PAPER: Levels of Organization in General Intelligence". Retrieved July 5, 2017. 
  26. ↑ Singularity Institute for Artificial Intelligence. "Levels of Organization in General Intelligence". Retrieved July 5, 2017. 
  27. ↑ "SL4: By Thread". Retrieved July 1, 2017. 
  28. ↑ "FlareProgrammingLanguage". SL4 Wiki. September 14, 2007. Retrieved July 13, 2017. 
  29. ↑ "Yudkowsky - Bayes' Theorem". Retrieved July 5, 2017. Eliezer Yudkowsky's work is supported by the Machine Intelligence Research Institute. If you've found Yudkowsky's pages on rationality useful, please consider donating to the Machine Intelligence Research Institute. 
  30. ↑ Yudkowsky, Eliezer (April 30, 2003). "Singularity Institute - update". SL4. 
  31. ↑ 31.0 31.1 31.2 31.3 31.4 31.5 31.6 "News of the Singularity Institute for Artificial Intelligence". Retrieved July 4, 2017. 
  32. ↑ "Singularity Institute for Artificial Intelligence // The SIAI Voice". Retrieved July 4, 2017. On March 4, 2004, the Singularity Institute announced Tyler Emerson as our Executive Director. Emerson will be responsible for guiding the Institute. His focus is in nonprofit management, marketing, relationship fundraising, leadership and planning. He will seek to cultivate a larger and more cohesive community that has the necessary resources to develop Friendly AI. 
  33. ↑ Tyler Emerson (April 7, 2004). "SL4: Michael Anissimov - SIAI Advocacy Director". Retrieved July 1, 2017. The Singularity Institute announces Michael Anissimov as our Advocacy Director. Michael has been an active volunteer for two years, and one of the more prominent voices in the singularity community. He is committed and thoughtful, and we feel very fortunate to have him help lead our advocacy. 
  34. ↑ "Machine Intelligence Research Institute: This is an old revision of this page, as edited by 63.201.36.156 (talk) at 19:28, 14 April 2004.". Retrieved July 15, 2017. 
  35. ↑ Eliezer Yudkowsky. "Coherent Extrapolated Volition" (PDF). Retrieved July 1, 2017. The information is current as of May 2004, and should not become dreadfully obsolete until late June, when I plant to have an unexpected insight. 
  36. ↑ "Yudkowsky - Technical Explanation". Retrieved July 5, 2017. Eliezer Yudkowsky's work is supported by the Machine Intelligence Research Institute. 
  37. ↑ Singularity Institute. "News of the Singularity Institute for Artificial Intelligence". Retrieved July 4, 2017. 
  38. ↑ 38.0 38.1 38.2 38.3 38.4 38.5 38.6 Brandon Reinhart. "SIAI - An Examination". LessWrong. Retrieved June 30, 2017. 
  39. ↑ "SL4: By Thread". Retrieved July 1, 2017. 
  40. ↑ "The Singularity Institute for Artificial Intelligence - 2006 $100,000 Singularity Challenge". Retrieved July 5, 2017. 
  41. ↑ "Twelve Virtues of Rationality". Retrieved July 5, 2017. 
  42. ↑ 42.0 42.1 "Singularity Summit". Machine Intelligence Research Institute. Retrieved June 30, 2017. 
  43. ↑ Dan Farber. "The great Singularity debate". ZDNet. Retrieved June 30, 2017. 
  44. ↑ Jerry Pournelle (May 20, 2006). "Chaos Manor Special Reports: The Stanford Singularity Summit". Retrieved June 30, 2017. 
  45. ↑ "Overcoming Bias : Bio". Retrieved June 1, 2017. 
  46. ↑ "Form 990 2007" (PDF). Retrieved July 8, 2017. 
  47. ↑ "Our History". Machine Intelligence Research Institute. 
  48. ↑ "Singularity Institute for Artificial Intelligence". YouTube. Retrieved July 8, 2017. 
  49. ↑ "The Power of Intelligence". Machine Intelligence Research Institute. July 10, 2007. Retrieved May 5, 2020. 
  50. ↑ "The Singularity Summit 2007". Retrieved June 30, 2017. 
  51. ↑ "Yudkowsky - The Simple Truth". Retrieved July 5, 2017. 
  52. ↑ "About". OpenCog Foundation. Retrieved July 6, 2017. 
  53. ↑ Goertzel, Ben (October 29, 2010). "The Singularity Institute's Scary Idea (and Why I Don't Buy It)". Retrieved September 15, 2019. 
  54. ↑ http://helldesign.net. "The Singularity Summit 2008: Opportunity, Risk, Leadership > Program". Retrieved June 30, 2017. 
  55. ↑ Elise Ackerman (October 23, 2008). "Annual A.I. conference to be held this Saturday in San Jose". The Mercury News. Retrieved July 5, 2017. 
  56. ↑ "The Hanson-Yudkowsky AI-Foom Debate". Lesswrongwiki. LessWrong. Retrieved July 1, 2017. 
  57. ↑ "Eliezer_Yudkowsky comments on Thoughts on the Singularity Institute (SI) - Less Wrong". LessWrong. Retrieved July 15, 2017. Nonetheless, it already has a warm place in my heart next to the debate with Robin Hanson as the second attempt to mount informed criticism of SIAI. 
  58. ↑ 58.0 58.1 58.2 58.3 "Recent Singularity Institute Accomplishments". Singularity Institute for Artificial Intelligence. Retrieved July 6, 2017. 
  59. ↑ "FAQ - LessWrong Wiki". LessWrong. Retrieved June 1, 2017. 
  60. ↑ Michael Vassar (February 16, 2009). "Introducing Myself". Machine Intelligence Research Institute. Retrieved July 1, 2017. 
  61. ↑ 61.0 61.1 RobbBB (March 13, 2015). "Rationality: From AI to Zombies". LessWrong. Retrieved July 1, 2017. 
  62. ↑ "Singularity Institute (@singinst)". Twitter. Retrieved July 4, 2017. 
  63. ↑ 63.0 63.1 63.2 Amy Willey Labenz. Personal communication. May 27, 2022.
  64. ↑ "Wayback Machine". Retrieved July 2, 2017. 
  65. ↑ 65.0 65.1 McCabe, Thomas (February 4, 2011). "The Uncertain Future Forecasting Project Goes Open-Source". H Plus Magazine. Archived from the original on April 13, 2012. Retrieved July 15, 2017. 
  66. ↑ "Singularity Summit 2009 Program". Singularity Institute. Retrieved June 30, 2017. 
  67. ↑ Stuart Fox (October 2, 2009). "Singularity Summit 2009: The Singularity Is Near". Popular Science. 
  68. ↑ "Form 990 2009" (PDF). Retrieved July 8, 2017. 
  69. ↑ "Reply to Holden on The Singularity Institute". July 10, 2012. Retrieved June 30, 2017. 
  70. ↑ Michael Anissimov (December 12, 2009). "The Uncertain Future". The Singularity Institute Blog. Retrieved July 5, 2017. 
  71. ↑ "Form 990 2010" (PDF). Retrieved July 8, 2017. 
  72. ↑ "Harry Potter and the Methods of Rationality Chapter 1: A Day of Very Low Probability, a harry potter fanfic". FanFiction. Retrieved July 1, 2017. Updated: 3/14/2015 - Published: 2/28/2010 
  73. ↑ David Whelan (March 2, 2015). "The Harry Potter Fan Fiction Author Who Wants to Make Everyone a Little More Rational". Vice. Retrieved July 1, 2017. 
  74. ↑ "2013 in Review: Fundraising - Machine Intelligence Research Institute". Machine Intelligence Research Institute. August 13, 2014. Retrieved July 1, 2017. Recently, we asked (nearly) every donor who gave more than $3,000 in 2013 about the source of their initial contact with MIRI, their reasons for donating in 2013, and their preferred methods for staying in contact with MIRI. […] Four came into contact with MIRI via HPMoR. 
  75. ↑ Rees, Gareth (August 17, 2010). "Zendegi - Gareth Rees". Retrieved July 15, 2017. 
  76. ↑ Sotala, Kaj (October 7, 2010). "Greg Egan disses stand-ins for Overcoming Bias, SIAI in new book". Retrieved July 15, 2017. 
  77. ↑ Hanson, Robin (March 25, 2012). "Egan's Zendegi". Retrieved July 15, 2017. 
  78. ↑ "Singularity Summit | Program". Retrieved June 30, 2017. 
  79. ↑ "Machine Intelligence Research Institute - Posts". Retrieved July 4, 2017. 
  80. ↑ "Machine Intelligence Research Institute - Posts". Retrieved July 4, 2017. 
  81. ↑ Louie Helm (December 21, 2010). "Announcing the Tallinn-Evans $125,000 Singularity Challenge". Machine Intelligence Research Institute. Retrieved July 7, 2017. 
  82. ↑ Kaj Sotala (December 26, 2010). "Tallinn-Evans $125,000 Singularity Challenge". LessWrong. Retrieved July 7, 2017. 
  83. ↑ "GiveWell conversation with SIAI". GiveWell. February 2011. Retrieved July 4, 2017. 
  84. ↑ Holden Karnofsky. "Singularity Institute for Artificial Intelligence". Yahoo! Groups. Retrieved July 4, 2017. 
  85. ↑ "lukeprog comments on Thoughts on the Singularity Institute (SI)". LessWrong. Retrieved June 30, 2017. When I began to intern with the Singularity Institute in April 2011, I felt uncomfortable suggesting that people donate to SingInst, because I could see it from the inside and it wasn't pretty. 
  86. ↑ Holden Karnofsky. "Re: [givewell] Singularity Institute for Artificial Intelligence". Yahoo! Groups. Retrieved July 4, 2017. 
  87. ↑ "singularity.org". Retrieved July 4, 2017. 
  88. ↑ 88.0 88.1 "Wayback Machine". Retrieved July 4, 2017. 
  89. ↑ "Singularity Institute Volunteering". Retrieved July 14, 2017. 
  90. ↑ "Singularity Summit | Program". Retrieved June 30, 2017. 
  91. ↑ "SingularitySummits". YouTube. Retrieved July 4, 2017. Joined Oct 17, 2011 
  92. ↑ Luke Muehlhauser (January 16, 2012). "Machine Intelligence Research Institute Progress Report, December 2011". Machine Intelligence Research Institute. Retrieved July 14, 2017. 
  93. ↑ lukeprog (December 12, 2011). "New 'landing page' website: Friendly-AI.com". LessWrong. Retrieved July 2, 2017. 
  94. ↑ Frank, Sam (January 1, 2015). "Come With Us If You Want to Live. Among the apocalyptic libertarians of Silicon Valley". Harper's Magazine. Retrieved July 15, 2017. 
  95. ↑ "Video Q&A with Singularity Institute Executive Director". LessWrong. December 10, 2011. Retrieved May 31, 2021. 
  96. ↑ "Q&A #2 with Luke Muehlhauser, Machine Intelligence Research Institute Executive Director". Machine Intelligence Research Institute. January 12, 2012. Retrieved May 31, 2021. 
  97. ↑ "Wayback Machine". Retrieved July 4, 2017. 
  98. ↑ Louie Helm (May 8, 2012). "Machine Intelligence Research Institute Progress Report, April 2012". Machine Intelligence Research Institute. Retrieved June 30, 2017. 
  99. ↑ Holden Karnofsky. "Thoughts on the Singularity Institute (SI)". LessWrong. Retrieved June 30, 2017. 
  100. ↑ Helm, Louie (August 6, 2012). "July 2012 Newsletter". Machine Intelligence Research Institute. Retrieved May 5, 2020. 
  101. ↑ David J. Hill (August 29, 2012). "Singularity Summit 2012 Is Coming To San Francisco October 13-14". Singularity Hub. Retrieved July 6, 2017. 
  102. ↑ 102.00 102.01 102.02 102.03 102.04 102.05 102.06 102.07 102.08 102.09 102.10 102.11 102.12 102.13 102.14 102.15 102.16 102.17 102.18 102.19 102.20 "Research Workshops - Machine Intelligence Research Institute". Machine Intelligence Research Institute. Retrieved July 1, 2017. 
  103. ↑ "Singularity University Acquires the Singularity Summit". Singularity University. December 9, 2012. Retrieved June 30, 2017. 
  104. ↑ Fox, Joshua (February 14, 2013). "The Singularity Wars". LessWrong. Retrieved July 15, 2017. 
  105. ↑ "Form 990 2013" (PDF). Retrieved July 8, 2017. 
  106. ↑ "Conversations Archives". Machine Intelligence Research Institute. Retrieved July 15, 2017. 
  107. ↑ Luke Muehlhauser (March 22, 2015). "2014 in review". Machine Intelligence Research Institute. Retrieved July 15, 2017. 
  108. ↑ "March Newsletter". Machine Intelligence Research Institute. March 7, 2013. Retrieved July 1, 2017. 
  109. ↑ "We are now the "Machine Intelligence Research Institute" (MIRI)". Machine Intelligence Research Institute. January 30, 2013. Retrieved June 30, 2017. 
  110. ↑ "Facing the Intelligence Explosion, Luke Muehlhauser". Amazon.com. Retrieved July 1, 2017. 
  111. ↑ "Machine Intelligence Research Institute - Coming soon...". Retrieved July 4, 2017. 
  112. ↑ Luke Muehlhauser (February 28, 2013). "Welcome to Intelligence.org". Machine Intelligence Research Institute. Retrieved May 5, 2020. 
  113. ↑ Luke Muehlhauser (April 25, 2013). ""Singularity Hypotheses" Published". Machine Intelligence Research Institute. Retrieved July 14, 2017. 
  114. ↑ "Singularity Hypotheses: A Scientific and Philosophical Assessment (The Frontiers Collection): 9783642325595: Medicine & Health Science Books". Amazon.com. Retrieved July 14, 2017. 
  115. ↑ Luke Muehlhauser (December 11, 2013). "MIRI's Strategy for 2013". Machine Intelligence Research Institute. Retrieved July 6, 2017. 
  116. ↑ Jon Southurst (January 19, 2014). "Ripple Creator Donates $500k in XRP to Artificial Intelligence Research Charity". CoinDesk. Retrieved July 6, 2017. 
  117. ↑ Luke Muehlhauser (January 27, 2014). "Existential Risk Strategy Conversation with Holden Karnofsky". Machine Intelligence Research Institute. Retrieved July 7, 2017. 
  118. ↑ "Smarter Than Us: The Rise of Machine Intelligence, Stuart Armstrong". Amazon.com. Retrieved July 1, 2017. Publisher: Machine Intelligence Research Institute (February 1, 2014) 
  119. ↑ Rob Bensinger (August 10, 2015). "Assessing Our Past and Potential Impact". Machine Intelligence Research Institute. Retrieved July 6, 2017. 
  120. ↑ "Recent Hires at MIRI". Machine Intelligence Research Institute. March 13, 2014. Retrieved July 13, 2017. 
  121. ↑ "MIRI's March 2014 Newsletter". Machine Intelligence Research Institute. March 18, 2014. Retrieved May 27, 2018. 
  122. ↑ "Machine Intelligence Research Institute - Photos". Facebook. Retrieved May 27, 2018. 
  123. ↑ "Carl_Shulman comments on My Cause Selection: Michael Dickens". Effective Altruism Forum. September 17, 2015. Retrieved July 6, 2017. 
  124. ↑ "Recent Site Activity - AI Impacts". Retrieved June 30, 2017. Jul 4, 2014, 10:39 AM Katja Grace edited Predictions of human-level AI timelines 
  125. ↑ "MIRI's September Newsletter". Machine Intelligence Research Institute. September 1, 2014. Retrieved July 15, 2017. 
  126. ↑ Benja Fallenstein. "Welcome!". Intelligent Agent Foundations Forum. Retrieved June 30, 2017. Post by Benja Fallenstein 969 days ago 
  127. ↑ Luke Muehlhauser (January 11, 2015). "An improved "AI Impacts" website". Machine Intelligence Research Institute. Retrieved June 30, 2017. 
  128. ↑ "AI safety conference in Puerto Rico". Future of Life Institute. October 12, 2015. Retrieved July 13, 2017. 
  129. ↑ Nate Soares (July 16, 2015). "An Astounding Year". Machine Intelligence Research Institute. Retrieved July 13, 2017. 
  130. ↑ Ryan Carey. "Rationality: From AI to Zombies was released today!". Effective Altruism Forum. Retrieved July 1, 2017. 
  131. ↑ Luke Muehlhauser (May 6, 2015). "A fond farewell and a new Executive Director". Machine Intelligence Research Institute. Retrieved June 30, 2017. 
  132. ↑ "Self-prediction in Decision Theory and Artificial Intelligence — Faculty of Philosophy". Retrieved February 24, 2018. 
  133. ↑ Nate Soares (June 3, 2015). "Taking the reins at MIRI". LessWrong. Retrieved July 5, 2017. 
  134. ↑ "I am Nate Soares, AMA!". Effective Altruism Forum. Retrieved July 5, 2017. 
  135. ↑ "MIRI Summer Fellows 2015". CFAR. June 21, 2015. Retrieved July 8, 2017. 
  136. ↑ "Center for Applied Rationality — General Support". Open Philanthropy. Retrieved July 8, 2017. We have some doubts about CFAR's management and operations, and we see CFAR as having made only limited improvements over the last two years, with the possible exception of running the MIRI Summer Fellows Program in 2015, which we understand to have been relatively successful at recruiting staff for MIRI. 
  137. ↑ "Library/Machine Intelligence Research Institute". Effective Altruism Wikia. September 26, 2015. Retrieved July 15, 2017. 
  138. ↑ Larks (December 13, 2016). "2017 AI Risk Literature Review and Charity Comparison". Effective Altruism Forum. Retrieved July 8, 2017. 
  139. ↑ "Arbital AI Alignment Exploration". Retrieved January 30, 2018. 
  140. ↑ Soares, Nate (March 30, 2016). "MIRI has a new COO: Malo Bourgon". Machine Intelligence Research Institute. Retrieved September 15, 2019. 
  141. ↑ "The AI Alignment Problem: Why It's Hard, and Where to Start". May 6, 2016. Retrieved May 7, 2020. 
  142. ↑ Yudkowsky, Eliezer (December 28, 2016). "AI Alignment: Why It's Hard, and Where to Start". Retrieved May 7, 2020. 
  143. ↑ Rob Bensinger (July 27, 2016). "New paper: "Alignment for advanced machine learning systems"". Machine Intelligence Research Institute. Retrieved July 1, 2017. 
  144. ↑ "Machine Intelligence Research Institute — General Support". Open Philanthropy. Retrieved June 30, 2017. 
  145. ↑ "New paper: "Logical induction"". Machine Intelligence Research Institute. September 12, 2016. Retrieved July 1, 2017. 
  146. ↑ Scott Aaronson (October 9, 2016). "Shtetl-Optimized » Blog Archive » Stuff That's Happened". Retrieved July 1, 2017. Some of you will also have seen that folks from the Machine Intelligence Research Institute (MIRI)—Scott Garrabrant, Tsvi Benson-Tilsen, Andrew Critch, Nate Soares, and Jessica Taylor—recently put out a major 130-page paper entitled "Logical Induction". 
  147. ↑ Rob Bensinger (October 11, 2016). "Ask MIRI Anything (AMA)". Effective Altruism Forum. Retrieved July 5, 2017. 
  148. ↑ "AI Impacts — General Support". Open Philanthropy. Retrieved June 30, 2017. 
  149. ↑ "[1705.08807] When Will AI Exceed Human Performance? Evidence from AI Experts". Retrieved July 13, 2017. 
  150. ↑ "Media discussion of 2016 ESPAI". AI Impacts. June 14, 2017. Retrieved July 13, 2017. 
  151. ↑ "Updates to the research team, and a major donation - Machine Intelligence Research Institute". Machine Intelligence Research Institute. July 4, 2017. Retrieved July 4, 2017. 
  152. ↑ Daniel Dewey (July 7, 2017). "My Current Thoughts on MIRI's "Highly Reliable Agent Design" Work". Effective Altruism Forum. Retrieved July 7, 2017. 
  153. ↑ Yudkowsky, Eliezer; Soares, Nate. "[1710.05060] Functional Decision Theory: A New Theory of Instrumental Rationality". Retrieved October 22, 2017. Submitted on 13 Oct 2017 
  154. ↑ Matthew Graves (October 22, 2017). "New Paper: "Functional Decision Theory"". Machine Intelligence Research Institute. Retrieved October 22, 2017. 
  155. ↑ "There's No Fire Alarm for Artificial General Intelligence". Machine Intelligence Research Institute. October 13, 2017. Retrieved April 19, 2020. 
  156. ↑ "There's No Fire Alarm for Artificial General Intelligence". LessWrong. October 13, 2017. Retrieved April 19, 2020. 
  157. ↑ Malo Bourgon (November 8, 2017). "A Major Grant from Open Philanthropy". Machine Intelligence Research Institute. Retrieved November 11, 2017. 
  158. ↑ "Machine Intelligence Research Institute — General Support (2017)". Open Philanthropy. November 8, 2017. Retrieved November 11, 2017. 
  159. ↑ "Inadequacy and Modesty". Retrieved October 29, 2017. 
  160. ↑ "Inadequacy and Modesty". Effective Altruism Forum. Retrieved October 29, 2017. 
  161. ↑ "Discussion - Inadequate Equilibria". Inadequate Equilibria. Retrieved December 12, 2017. 
  162. ↑ "Book Review: Inadequate Equilibria". Slate Star Codex. December 9, 2017. Retrieved December 12, 2017. 
  163. ↑ "Shtetl-Optimized » Blog Archive » Review of "Inadequate Equilibria," by Eliezer Yudkowsky". Retrieved December 12, 2017. 
  164. ↑ Robin Hanson (November 25, 2017). "Overcoming Bias : Why Be Contrarian?". Retrieved December 12, 2017. 
  165. ↑ Yudkowsky, Eliezer (November 25, 2017). "Security Mindset and Ordinary Paranoia". Machine Intelligence Research Institute. Retrieved May 7, 2020. 
  166. ↑ Yudkowsky, Eliezer (November 26, 2017). "Security Mindset and the Logistic Success Curve". Machine Intelligence Research Institute. Retrieved May 7, 2020. 
  167. ↑ Malo Bourgon (December 1, 2017). "MIRI's 2017 Fundraiser". Machine Intelligence Research Institute. Retrieved December 12, 2017. 
  168. ↑ Malo Bourgon (January 10, 2018). "Fundraising success!". Machine Intelligence Research Institute. Retrieved January 30, 2018. 
  169. ↑ 169.0 169.1 "MIRI's 2018 Fundraiser". Machine Intelligence Research Institute. November 26, 2018. Retrieved February 14, 2019. 
  170. ↑ "AI Risk for Computer Scientists. Join us for four days of leveling up thinking on AI risk.". Machine Intelligence Research Institute. Retrieved September 14, 2019. 
  171. ↑ Oliver Habryka (October 29, 2018). "Announcing the new AI Alignment Forum". Machine Intelligence Research Institute. Retrieved February 14, 2019. 
  172. ↑ "Introducing the AI Alignment Forum (FAQ)". AI Alignment Forum. Retrieved February 14, 2019. 
  173. ↑ Arnold, Raymond (July 10, 2018). "Announcing AlignmentForum.org Beta". LessWroong. Retrieved April 18, 2020. 
  174. ↑ "AI Alignment Writing Day 2018". Retrieved April 19, 2020. 
  175. ↑ "Embedded Agency". Machine Intelligence Research Institute. Retrieved February 14, 2019. 
  176. ↑ "Embedded Agency". LessWrong 2.0. October 29, 2018. Retrieved February 14, 2019. 
  177. ↑ "Embedded Agency". AI Alignment Forum. October 29, 2018. Retrieved February 14, 2019. 
  178. ↑ "MIRI on Twitter". Twitter. Retrieved February 14, 2019. "Embedded Agency" in finished form, with new material on self-reference and logical uncertainty 
  179. ↑ Rob Bensinger. "Rob Bensinger comments on Embedded Agents". LessWrong 2.0 viewer. Retrieved February 14, 2019. 
  180. ↑ "2018 Update: Our New Research Directions - Machine Intelligence Research Institute". Machine Intelligence Research Institute. November 22, 2018. Retrieved February 14, 2019. 
  181. ↑ "Our 2018 Fundraiser Review - Machine Intelligence Research Institute". Machine Intelligence Research Institute. February 11, 2019. Retrieved February 14, 2019. 
  182. ↑ Bensinger, Rob (November 28, 2018). "MIRI's newest recruit: Edward Kmett!". Machine Intelligence Research Institute. Retrieved September 14, 2019. 
  183. ↑ "MIRI's newest recruit: Edward Kmett!". Reddit. December 1, 2018. Retrieved September 14, 2019. 
  184. ↑ "Announcing a new edition of "Rationality: From AI to Zombies"". Machine Intelligence Research Institute. December 16, 2018. Retrieved February 14, 2019. 
  185. ↑ Rob Bensinger. "New edition of "Rationality: From AI to Zombies"". LessWrong 2.0. Retrieved February 14, 2019. 
  186. ↑ "Machine Intelligence Research Institute — General Support (2019)". Open Philanthropy. April 1, 2019. Retrieved September 14, 2019. 
  187. ↑ Bensinger, Rob (April 1, 2019). "New grants from the Open Philanthropy Project and BERI". Machine Intelligence Research Institute. Retrieved September 14, 2019. 
  188. ↑ Habryka, Oliver (April 23, 2019). "MIRI ($50,000)". Effective Altruism Forum. Retrieved September 15, 2019. 
  189. ↑ Colm Ó Riain (February 13, 2020). "Our 2019 Fundraiser Review". Machine Intelligence Research Institute. Retrieved April 19, 2020. 
  190. ↑ "Open Philanthropy donations made (filtered to cause areas matching AI risk)". Retrieved July 27, 2017. 
  191. ↑ 191.0 191.1 191.2 Bensinger, Rob (April 27, 2020). "MIRI's largest grant to date!". Machine Intelligence Research Institute. Retrieved May 2, 2020. 
  192. ↑ "Fund Payout Report: April 2020 – Long-Term Future Fund Grant Recommendations". Effective Altruism Funds. April 14, 2020. Retrieved May 5, 2020. 
  193. ↑ "SFF-2020-H1 S-process Recommendations Announcement". Survival and Flourishing Fund. May 29, 2020. Retrieved October 10, 2020. 
  194. ↑ "Survival and Flourishing Fund". Retrieved October 10, 2020. 
  195. ↑ Bensinger, Rob (October 9, 2020). "MIRI, the place where I work, is very seriously considering moving to a different country soon (most likely Canada), or moving to elsewhere in the US.". Facebook. Retrieved October 10, 2020. 
  196. ↑ Garrabrant, Scott (October 22, 2020). "Introduction to Cartesian Frames". LessWrong. Retrieved December 20, 2020. 
  197. ↑ Bensinger, Rob (October 23, 2020). "October 2020 Newsletter". Machine Intelligence Research Institute. Retrieved December 20, 2020. 
  198. ↑ "SFF-2020-H2 S-process Recommendations Announcement". Survival and Flourishing Fund. Retrieved December 10, 2020. 
  199. ↑ Bensinger, Rob (November 30, 2020). "November 2020 Newsletter". Machine Intelligence Research Institute. Retrieved December 20, 2020. 
  200. ↑ Bourgon, Malo (December 21, 2020). "2020 Updates and Strategy". Machine Intelligence Research Institute. Retrieved December 22, 2020. 
  201. ↑ 201.0 201.1 Bensinger, Rob (May 8, 2021). "MIRI location optimization (and related topics) discussion". LessWrong. Retrieved May 31, 2021. 
  202. ↑ Colm Ó Riain (May 13, 2021). "Our all-time largest donation, and major crypto support from Vitalik Buterin". Retrieved May 31, 2021. 
  203. ↑ Garrabrandt, Scott (May 23, 2021). "Finite Factored Sets". Machine Intelligence Research Institute. Retrieved May 31, 2021. 
  204. ↑ Bensinger, Rob (November 15, 2021). "Late 2021 MIRI Conversations". Alignment Forum. Retrieved December 1, 2021. 
  205. ↑ Soares, Nate (November 29, 2021). "Visible Thoughts Project and Bounty Announcement". Retrieved December 2, 2021. 
  206. ↑ "December 2021 Newsletter". Machine Intelligence Research Institute. December 31, 2021. Retrieved September 2, 2024. 
  207. ↑ "Six Dimensions of Operational Adequacy in AGI Projects". LessWrong. May 30, 2022. Retrieved September 5, 2024. 
  208. ↑ "AGI Ruin: A List of Lethalities". LessWrong. June 5, 2022. Retrieved September 5, 2024. 
  209. ↑ "Visible Thoughts Project and Bounty Announcement". LessWrong. April 25, 2023. Retrieved September 6, 2024. 
  210. ↑ "July 2022 Newsletter". Machine Intelligence Research Institute. July 30, 2022. Retrieved September 2, 2024. 
  211. ↑ Eliezer Yudkowsky (January 1, 2020). "Eliezer Yudkowsky's AI Plan Challenge". Retrieved October 8, 2024. 
  212. ↑ "159 - We're All Gonna Die with Eliezer Yudkowsky". YouTube. February 20, 2023. Retrieved April 14, 2024. 
  213. ↑ "Full Transcript: Eliezer Yudkowsky on the Bankless podcast". LessWrong. February 23, 2023. Retrieved April 14, 2024. 
  214. ↑ "Transcript: Yudkowsky on Bankless follow-up Q&A". LessWrong. February 27, 2023. Retrieved April 14, 2024. 
  215. ↑ Pope, Quintin (March 20, 2023). "My Objections to "We're All Gonna Die with Eliezer Yudkowsky"". LessWrong. Retrieved May 17, 2024. 
  216. ↑ Yudkowsky, Eliezer (March 29, 2023). "Pausing AI Developments Isn't Enough. We Need to Shut it All Down". Time Magazine. Retrieved May 17, 2024. 
  217. ↑ "Pausing AI Developments Isn't Enough. We Need to Shut it All Down by Eliezer Yudkowsky". March 29, 2023. Retrieved May 17, 2024. 
  218. ↑ "Announcing MIRI's New CEO and Leadership Team". Machine Intelligence Research Institute. October 10, 2023. Retrieved September 2, 2024. 
  219. ↑ "Where I Agree and Disagree with Eliezer". LessWrong. August 10, 2023. Retrieved September 5, 2024. 
  220. ↑ "MIRI 2024 Mission and Strategy Update". LessWrong. May 14, 2024. Retrieved September 5, 2024. 
  221. ↑ "April 2024 Newsletter". Machine Intelligence Research Institute. April 12, 2024. Retrieved September 2, 2024. 
  222. ↑ "The Risks of Expanding the Definition of AI Safety". Semafor. March 9, 2024. Retrieved September 5, 2024. 
  223. ↑ "April 2024 Newsletter". Machine Intelligence Research Institute. April 12, 2024. Retrieved September 2, 2024. 
  224. ↑ "NTIA Request for Comment on Open-Weight AI Models". Regulations.gov. Retrieved September 10, 2024. 
  225. ↑ "Governing AI for Humanity Interim Report" (PDF). United Nations. Retrieved September 10, 2024. 
  226. ↑ "OMB Request for Information on AI Procurement in Government". Federal Register. Retrieved September 10, 2024. 
  227. ↑ "May 2024 Newsletter". Machine Intelligence Research Institute. May 14, 2024. Retrieved September 5, 2024. 
  228. ↑ "May 2024 Newsletter". Machine Intelligence Research Institute. May 14, 2024. Retrieved September 5, 2024. 
  229. ↑ "MIRI 2024 Communications Strategy". Machine Intelligence Research Institute. May 29, 2024. Retrieved September 5, 2024. 
  230. ↑ "Response to Aschenbrenner's Situational Awareness". Effective Altruism Forum. May 15, 2023. Retrieved September 5, 2024. 
  231. ↑ "June 2024 Newsletter". Machine Intelligence Research Institute. June 14, 2024. Retrieved September 2, 2024. 
  232. ↑ "Machine Intelligence Research Institute". Google Trends. Retrieved 11 March 2021. 
  233. ↑ "Machine Intelligence Research Institute". books.google.com. Retrieved 11 March 2021. 
  234. ↑ "Singularity Institute". wikipediaviews.org. Retrieved 17 March 2021. 
  235. ↑ "Singularity Institute for Artificial Intelligence: Revision history". Retrieved July 15, 2017. 
  236. ↑ "All public logs: search Singularity Institute". Retrieved July 15, 2017. 
Retrieved from "http://timelines.issarice.com/index.php?title=Timeline_of_Machine_Intelligence_Research_Institute&oldid=83108"