Nicholas Carlini
Machine-learning security researcher who announced in March 2025 that he was leaving Google DeepMind to join Anthropic.
Description from Selekoa’s record. Sources: Career Update: Google DeepMind -> Anthropic
Computed from stored claims by fixed rules. The page adds no connecting text and infers no moves between roles.
Download research starter
The documented path, undated records, works, credits, governance records and every other claim on this page, with evidence locations, qualifications, the computed limits, stored paraphrases and permitted short quotes. Original documents are not included.
Original sources for a reference manager. Every source this export cites, for Zotero, EndNote and similar tools. Cite the original documents, not Selekoa, and check each entry against them.
Appears in
No collection holds a record about this person yet.
Documented path
Roles, founding, launches and departures in date order. A snapshot is one dated report, not a tenure. An interval keeps only the start and end its source states, and a missing end is unknown. One row following another does not mean one role led to the next.
- 2025-03-05 · snapshotSnapshot · one dated reportHeld role at Anthropic · not stated (research on adversarial machine learning, for a stated period of one year)Nicholas Carlini announced on 5 March 2025 that he would be joining Anthropic for a year to continue his research on adversarial machine learning.Source published 2025-03-05 · Stated in source · Supported AnnouncedInspect claim & evidence ↗
- 2025-03-05 · eventEventHad departure announced from Google DeepMind · Not stated (researcher; he joined Google Brain in 2018)Nicholas Carlini announced on 5 March 2025 that he had decided to leave Google DeepMind after seven years; he stated that he no longer believed DeepMind (and DeepMind leadership) supported the type of high-impact security and privacy research he likes to do.Source published 2025-03-05 · Stated in source · Supported AnnouncedInspect claim & evidence ↗
Stated reason, as given by Nicholas Carlini (2025-03-05)
Carlini wrote that he no longer believed DeepMind (and DeepMind leadership) supported the type of high-impact security and privacy research he likes to do. Attributed to Nicholas Carlini, in “Career Update: Google DeepMind -> Anthropic”; not a finding about why it happened.
What this record cannot tell you
1 record reports a role or state at a single date. A snapshot establishes neither when the role began nor when it ended.
No end date or departure announcement is recorded for Anthropic. That is a coverage gap: it does not mean the person remained.
Nothing here records why a person moved, or that one role led to another. Authorship and a printed affiliation do not show employment. A record that has not been collected is missing, and its absence disproves nothing.
Works authored
Authorship as credited in each document. A credit does not imply sole authorship.
No authorship record.
Explicit credits
Documents that expressly credit this person, within the scope each record states. Credit that this person gives in their own work appears on that work’s document page.
No explicit credit is recorded.
Governance roles
Records classified under governance, such as board actions and executive appointments announced as governance events. They also appear in the documented path where they are role or departure records.
No governance record.