ABAB Pedia

Christopher Olah

Christopher Olah (commonly known as Chris Olah) is a machine learning researcher and co-founder of Anthropic who works on neural network interpretability. He was previously a research scientist at Google BrainGoogle BrainGoogle Brain was Google’s AI research team, co-founded by Jeff Dean and others in 2011. It open-sourced the TensorFlow machine learning system in 2015 and was merged with DeepMind into Google DeepMind in April 2023.Open the full entry, led the Clarity team at OpenAIOpenAIOpenAI is an artificial intelligence research and product organization founded in 2015 that develops the GPT model series, ChatGPT, Codex and a developer API, and conducts research on image, speech, video and intelligent agents. After its 2025 restructuring, the nonprofit OpenAI Foundation continues to control the for-profit OpenAI Group PBC. Its growth has involved large-scale fundraising and compute partnerships, and has also been accompanied by controversies over governance, copyright, privacy and AI safety.Open the full entry, and founded the scientific journal Distill.

Contents14 sections
Key facts

Joining hacklab.to in Toronto

From 2009, Olah was a member of hacklab.to, a hackerspace in Toronto, and served as a director of the organization from February 2012 to February 2014. From 2011 to 2013, he founded the open-source project ImplicitCAD, a programming language written in Haskell that compiles into 3D objects. In May 2011, he disclosed a distributed denial-of-service (DDoS) vulnerability in the Unbound DNS resolver (CVE-2011-1922).⁠[1][1][1]

Thiel Fellowship

In July 2012, at age 19, Olah received a Thiel FellowshipThiel FellowshipThe Thiel Fellowship is a two-year grant program started by investor Peter Thiel and run by the Thiel Foundation that funds young people who skip or stop out of college to build new things instead. The program has been criticized by figures in education for encouraging young people to leave college.Open the full entry. The program provides $100,000 to people under 20 pursuing research or entrepreneurship, encouraging them to go straight into research or entrepreneurship rather than attend university. Olah did not attend college and holds no university degree.⁠[1][2][3]

Internship at Google Brain

From July to October 2014, Olah interned at Google Brain under host Jeff DeanJeff DeanJeff Dean is a computer scientist and software engineer, the co-founder and CEO of Discovery Loop and a former chief scientist at Google. He joined Google in 1999 and worked on MapReduce, Bigtable, Spanner, TensorFlow and machine learning infrastructure. He left in August 2026 to start Discovery Loop. In July of that year he discussed AI startups, engineering optimization and automated research at Y Combinator’s Startup School.Open the full entry, exploring the use of interactive media to present the internal representations of neural networks, and created the meta-SNE algorithm for visualizing the space of neural networks. He interned again from May to October 2015 under host Greg Corrado; in October 2015 he became a research associate, and from October 2016 he was a research scientist, until 2018.⁠[1][1][1]

While at Google Brain, he published “Understanding LSTM Networks,” an article explaining long short-term memory networks, on his personal blog in August 2015, and was listed as one of the authors of the 2015 TensorFlow systems paper.⁠[1][1]

Co-Launching DeepDream

In June 2015, the Google Research blog published “Inceptionism: Going Deeper into Neural Networks,” by Alexander Mordvintsev, Olah (then a software engineering intern), and Mike Tyka. This launch became the widely known DeepDream neural network image visualization, with Olah as second author.⁠[4][3]

Founding Distill

In 2016, Olah founded the scientific journal Distill and served as an editor, alongside fellow editor Shan Carter. The journal published articles explaining machine learning research with web-native interactive diagrams; it operated from 2016 to 2021 and has since been on an indefinite hiatus.⁠[1][5][5][5]

Co-Authoring “Concrete Problems in AI Safety”

Joining OpenAI to Lead the Clarity Team

In October 2018, Olah joined OpenAI as a member of technical staff, leading the newly formed Clarity team working on neural network interpretability. In March 2019, he published “Activation Atlases” in Distill with Shan Carter, Zan Armstrong, Ludwig Schubert, and Ian Johnson. He led OpenAI’s interpretability team until December 2020.⁠[1][1][3]

Co-Founding Anthropic

Identifying Millions of “Features” in Claude

On May 21, 2024, Anthropic’s mechanistic interpretability team, led by Olah, released a paper. Using a method called “dictionary learning,” the 18-person team identified millions of “features” in the Claude Sonnet model, meaning combinations of neurons that correspond to a concept when activated together, including a feature associated with the Golden Gate Bridge and features related to safety topics such as biological warfare.⁠[7][7][7][7][7][7]

Remarks at the Vatican Presentation of an Encyclical

On May 25, 2026, Pope Leo XIV released “Magnifica humanitas,” an encyclical on the topic of AI. Olah was invited to speak at the presentation of the encyclical in the Vatican, as part of Anthropic’s initiative to widen the conversation on AI. In his remarks, he said that every frontier AI lab, including Anthropic, operates inside a set of incentives and constraints that can sometimes conflict with “doing the right thing,” and that people outside those incentives are therefore needed to watch and criticize the technology; he also said that AI models are not engineered the way a bridge or an airplane is, but are “grown.”⁠[8][8][8][8]

Ownership and governance in the amended IPO filing

In its second amended S-1/A filed with the SECU.S. Securities and Exchange CommissionThe U.S. Securities and Exchange Commission is the federal securities regulator of the United States. Created during the Great Depression by the Securities Exchange Act of 1934 and led by five commissioners, it has a three-part mission of protecting investors, maintaining fair, orderly and efficient markets, and facilitating capital formation.Open the full entry on November 10, 2026, Anthropic listed Christopher Olah with beneficial ownership of 242,882,420 Class A shares before the offering, using September 30, 2026 as the ownership reference date. The filing proposed selling 3,378,378 shares in the IPO, leaving 239,504,042 shares after the offering, assuming no exercise of the underwriters’ additional-share option. These were proposed transactions in a registration statement that was not yet effective.⁠[9]

The footnote broke that ownership figure into 240,057,520 shares held of record by him, 2,185,400 shares underlying options exercisable within 60 days of September 30, 2026, and 639,500 shares issuable upon settlement of RSUs whose service-based vesting condition would be satisfied within that period. The total therefore combined record holdings with equity awards included under the filing’s beneficial-ownership rules.⁠[9]

The filing also named Christopher Olah among the seven initial members of Founder LLC. Immediately before completion of the offering, the company planned to issue that collective entity one Class F share without economic rights, carrying 50.1% of voting power on matters submitted to stockholders generally. The Long-Term Benefit Trust would separately retain the right to elect a board majority through Class T shares. Each Continuing Founder would have one vote within Founder LLC; its collective voting power was not an individual founder’s holding.⁠[9]

The prospectus also disclosed each co-founder’s pledge to devote 80% of their personal Anthropic equity to philanthropy. The pledges were voluntary and revocable, with no binding legal obligation or fixed timing, amount, recipients or structure. They concerned personal economic interests and did not change the Founder LLC’s Class F voting arrangement.⁠[9]

Research Focus, Citizenship, and Net Worth

Olah describes his work as reverse engineering artificial neural networks into human-understandable algorithms; he is currently the interpretability research lead at Anthropic, and his citizenship is Canadian.⁠[6][2][2]

Related Organizations, Websites, and Public Accounts

Sources11 sources

  1. Chris Olah CV (opens in a new window)colah.github.io↑
  2. Christopher Olah (opens in a new window)forbes.com↑
  3. Chris Olah on what the hell is going on inside neural networks (opens in a new window)80000hours.org↑
  4. Inceptionism: Going Deeper into Neural Networks (opens in a new window)research.google↑
  5. Distill is dedicated to making machine learning clear and dynamic (opens in a new window)distill.pub↑
  6. About Me - colah's blog (opens in a new window)colah.github.io↑
  7. AI Is a Black Box. Anthropic Figured Out a Way to Look Inside (opens in a new window)wired.com↑
  8. Anthropic co-founder Chris Olah's remarks on Pope Leo XIV's encyclical "Magnifica humanitas" (opens in a new window)anthropic.com↑
  9. Anthropic, PBC — Amendment No. 2 to Form S-1 (opens in a new window)sec.gov↑
  10. Chris Olah (@ch402) / X (opens in a new window)x.com↑
  11. Christopher Olah | LinkedIn (opens in a new window)linkedin.com↑