{"id":"d3bcb6489f3d94f01e5fa0e6c263d3f3250bc938eeaf392565ff6bc4cc5a7933","title":"METR seeks ML engineers for AI R&D risk evaluations","argument":"Published May 16, 2024, METR says it is developing evaluations for AI R&D capabilities and hiring ML research engineers and scientists to advance them. METR warns that systems able to automate AI engineering and research might rapidly improve AI capabilities, with dangerous capabilities emerging faster than humanity could respond with protective measures. It says the risks are hard to bound and argues that robust tests are critical to predict if or when this might occur.\n\nMETR says it has started developing threshold evaluations to help determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks. It hopes eventually to build evaluations that track progress smoothly, so evaluators are not caught by surprise. METR identifies researchers and engineers with substantial ML R&D experience as the main bottleneck. The post presents hiring them as a way to move the evaluations forward; it does not report that the evaluations already predict dangerous capabilities or that protective measures have been triggered.","topicIds":["artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 24, 2026; originally published May 16, 2024. This is a hiring post describing METR’s goals and plans; it reports no evaluation results or evidence of predictive performance.","speaker":{"name":"METR","roleAtSource":"Publisher"},"source":{"title":"ML Engineers Needed for New AI R&D Evals Project","publisher":"METR","format":"Article","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","publishedAt":"2024-05-16T07:00:00.000Z","dateBasis":"Date supplied by the publisher’s feed."},"evidence":[{"text":"METR says its goal is to provide early warning about AI agents that might dramatically improve themselves and trigger an explosion of dangerous capabilities.","excerpt":"Our goal is to provide an early warning before AI agents might dramatically improve themselves and kick off an ‘explosion’ of dangerous capabilities.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"},{"text":"METR argues that AI systems able to automate AI engineering and research might cause dangerous capabilities to emerge faster than humanity could respond with protective measures.","excerpt":"new dangerous capabilities emerge far more quickly than humanity could respond with protective measures.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"},{"text":"METR says it has started developing threshold evaluations to determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks.","excerpt":"METR has recently started developing threshold evaluations that can be run to determine whether AI R&D capabilities warrant protective measures such as information security that is resilient to state-actor attacks.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"},{"text":"METR says experienced researchers and engineers are the main bottleneck to progress on its evaluations.","excerpt":"Having researchers and engineers with substantial ML R&D experience themselves is the main bottleneck to progress on these evaluations.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How well will METR’s evaluations predict when AI R&D capabilities warrant protective measures?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"METR seeks ML engineers for AI R&D risk evaluations","summary":"Published May 16, 2024, METR says it is developing evaluations for AI R&D capabilities and hiring ML research engineers and scientists to advance them. METR warns that systems able to automate AI engineering and research might rapidly improve AI capabilities, with dangerous capabilities emerging faster than humanity could respond with protective measures. It says the risks are hard to bound and argues that robust tests are critical to predict if or when this might occur.\n\nMETR says it has started developing threshold evaluations to help determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks. It hopes eventually to build evaluations that track progress smoothly, so evaluators are not caught by surprise. METR identifies researchers and engineers with substantial ML R&D experience as the main bottleneck. The post presents hiring them as a way to move the evaluations forward; it does not report that the evaluations already predict dangerous capabilities or that protective measures have been triggered.","question":"How well will METR’s evaluations predict when AI R&D capabilities warrant protective measures?","claims":[{"text":"METR says its goal is to provide early warning about AI agents that might dramatically improve themselves and trigger an explosion of dangerous capabilities.","excerpt":"Our goal is to provide an early warning before AI agents might dramatically improve themselves and kick off an ‘explosion’ of dangerous capabilities.","support":"source-stated"},{"text":"METR argues that AI systems able to automate AI engineering and research might cause dangerous capabilities to emerge faster than humanity could respond with protective measures.","excerpt":"new dangerous capabilities emerge far more quickly than humanity could respond with protective measures.","support":"source-stated"},{"text":"METR says it has started developing threshold evaluations to determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks.","excerpt":"METR has recently started developing threshold evaluations that can be run to determine whether AI R&D capabilities warrant protective measures such as information security that is resilient to state-actor attacks.","support":"source-stated"},{"text":"METR says experienced researchers and engineers are the main bottleneck to progress on its evaluations.","excerpt":"Having researchers and engineers with substantial ML R&D experience themselves is the main bottleneck to progress on these evaluations.","support":"source-stated"}],"limitations":["Text as retrieved on October 24, 2026; originally published May 16, 2024.","This is a hiring post describing METR’s goals and plans; it reports no evaluation results or evidence of predictive performance."],"mediaId":null,"correctionReason":null,"id":"d3bcb6489f3d94f01e5fa0e6c263d3f3250bc938eeaf392565ff6bc4cc5a7933","publicationId":"75536446-16c7-bb29-3431-f8eabd2c5c3d","version":1,"sourceUrl":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","sourceTitle":"ML Engineers Needed for New AI R&D Evals Project","publisher":"METR","publishedAt":"2024-05-16T07:00:00.000Z","originalPublished":{"value":"2024-05-16T07:00:00.000Z","precision":"instant","basis":"backfill:feed_published"},"firstObservedAt":"2026-10-18T03:49:29.878Z","createdAt":"2026-11-18T16:33:09.767Z","historical":true,"entityIds":["metr"],"topicIds":["artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"b5dabf1ee015ed08c02c91ee953240869f8a5b8f16f91ca403d48a3796f8c54c","rubricVersion":"unrecorded","provider":"unrecorded","model":"anthropic/claude-opus-5.5","reasonCodes":["bridge_admitted_by_model","second_opinion_categorical"],"confidence":1,"decidedAt":"2026-10-24T09:45:40.444Z"},"provenance":{"textHash":"4962450b33b392b9f294f5b9bd3a34d981a33e2496b6bcd8bb2e86831d071505","observationId":"9f33981f7cc6c9969ab1ef03090b836aa2fabc9e2b5487acc94f9a582420ec9e","retrievedAt":"2026-10-24T09:47:38.929Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"1c284e92-6904-c5f3-1872-647e92d11105"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"46fe2cdc-bf09-02d8-4bcc-5079c44e9ebc"}},"audit":{"status":"passed","reviewedAt":"2026-11-18T16:29:12.433Z","model":"anthropic/claude-opus-5.5"}},"canonicalUrl":"https://escapemeridian.com/readings/metr-seeks-ml-engineers-for-ai-r-and-d-risk-evaluations-d3bcb648","formats":{"html":"/readings/metr-seeks-ml-engineers-for-ai-r-and-d-risk-evaluations-d3bcb648","markdown":"/readings/metr-seeks-ml-engineers-for-ai-r-and-d-risk-evaluations-d3bcb648.md","json":"/readings/metr-seeks-ml-engineers-for-ai-r-and-d-risk-evaluations-d3bcb648.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"conditions":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[{"id":"term-machine-learning","name":"Machine learning","explanation":"Methods that let computers learn patterns from examples instead of following rules written by hand.","referenceUrl":"https://developers.google.com/machine-learning/glossary#machine-learning"}],"entities":[{"id":"metr","name":"METR","kind":"organization","context":"METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.","referenceUrls":["https://metr.org/"],"path":"/entities/metr","json":"/entities/metr.json","markdown":"/entities/metr.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"d3bcb6489f3d94f01e5fa0e6c263d3f3250bc938eeaf392565ff6bc4cc5a7933","revision":"75536446-16c7-bb29-3431-f8eabd2c5c3d","revisionKind":"publication-id","title":"METR seeks ML engineers for AI R&D risk evaluations","url":"https://escapemeridian.com/readings/metr-seeks-ml-engineers-for-ai-r-and-d-risk-evaluations-d3bcb648"},"source":{"type":"source","id":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","revision":"4962450b33b392b9f294f5b9bd3a34d981a33e2496b6bcd8bb2e86831d071505","revisionKind":"sha256-source-text","title":"ML Engineers Needed for New AI R&D Evals Project","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings to this reading. Shared topics are not citations. Other record types need an explicit reference adapter."},"revisions":[{"id":"d3bcb6489f3d94f01e5fa0e6c263d3f3250bc938eeaf392565ff6bc4cc5a7933","title":"METR seeks ML engineers for AI R&D risk evaluations","argument":"Published May 16, 2024, METR says it is developing evaluations for AI R&D capabilities and hiring ML research engineers and scientists to advance them. METR warns that systems able to automate AI engineering and research might rapidly improve AI capabilities, with dangerous capabilities emerging faster than humanity could respond with protective measures. It says the risks are hard to bound and argues that robust tests are critical to predict if or when this might occur.\n\nMETR says it has started developing threshold evaluations to help determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks. It hopes eventually to build evaluations that track progress smoothly, so evaluators are not caught by surprise. METR identifies researchers and engineers with substantial ML R&D experience as the main bottleneck. The post presents hiring them as a way to move the evaluations forward; it does not report that the evaluations already predict dangerous capabilities or that protective measures have been triggered.","topicIds":["artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 24, 2026; originally published May 16, 2024. This is a hiring post describing METR’s goals and plans; it reports no evaluation results or evidence of predictive performance.","speaker":{"name":"METR","roleAtSource":"Publisher"},"source":{"title":"ML Engineers Needed for New AI R&D Evals Project","publisher":"METR","format":"Article","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","publishedAt":"2024-05-16T07:00:00.000Z","dateBasis":"Date supplied by the publisher’s feed."},"evidence":[{"text":"METR says its goal is to provide early warning about AI agents that might dramatically improve themselves and trigger an explosion of dangerous capabilities.","excerpt":"Our goal is to provide an early warning before AI agents might dramatically improve themselves and kick off an ‘explosion’ of dangerous capabilities.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"},{"text":"METR argues that AI systems able to automate AI engineering and research might cause dangerous capabilities to emerge faster than humanity could respond with protective measures.","excerpt":"new dangerous capabilities emerge far more quickly than humanity could respond with protective measures.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"},{"text":"METR says it has started developing threshold evaluations to determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks.","excerpt":"METR has recently started developing threshold evaluations that can be run to determine whether AI R&D capabilities warrant protective measures such as information security that is resilient to state-actor attacks.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"},{"text":"METR says experienced researchers and engineers are the main bottleneck to progress on its evaluations.","excerpt":"Having researchers and engineers with substantial ML R&D experience themselves is the main bottleneck to progress on these evaluations.","url":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"How well will METR’s evaluations predict when AI R&D capabilities warrant protective measures?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"METR seeks ML engineers for AI R&D risk evaluations","summary":"Published May 16, 2024, METR says it is developing evaluations for AI R&D capabilities and hiring ML research engineers and scientists to advance them. METR warns that systems able to automate AI engineering and research might rapidly improve AI capabilities, with dangerous capabilities emerging faster than humanity could respond with protective measures. It says the risks are hard to bound and argues that robust tests are critical to predict if or when this might occur.\n\nMETR says it has started developing threshold evaluations to help determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks. It hopes eventually to build evaluations that track progress smoothly, so evaluators are not caught by surprise. METR identifies researchers and engineers with substantial ML R&D experience as the main bottleneck. The post presents hiring them as a way to move the evaluations forward; it does not report that the evaluations already predict dangerous capabilities or that protective measures have been triggered.","question":"How well will METR’s evaluations predict when AI R&D capabilities warrant protective measures?","claims":[{"text":"METR says its goal is to provide early warning about AI agents that might dramatically improve themselves and trigger an explosion of dangerous capabilities.","excerpt":"Our goal is to provide an early warning before AI agents might dramatically improve themselves and kick off an ‘explosion’ of dangerous capabilities.","support":"source-stated"},{"text":"METR argues that AI systems able to automate AI engineering and research might cause dangerous capabilities to emerge faster than humanity could respond with protective measures.","excerpt":"new dangerous capabilities emerge far more quickly than humanity could respond with protective measures.","support":"source-stated"},{"text":"METR says it has started developing threshold evaluations to determine whether AI R&D capabilities warrant protective measures, including information security resilient to state-actor attacks.","excerpt":"METR has recently started developing threshold evaluations that can be run to determine whether AI R&D capabilities warrant protective measures such as information security that is resilient to state-actor attacks.","support":"source-stated"},{"text":"METR says experienced researchers and engineers are the main bottleneck to progress on its evaluations.","excerpt":"Having researchers and engineers with substantial ML R&D experience themselves is the main bottleneck to progress on these evaluations.","support":"source-stated"}],"limitations":["Text as retrieved on October 24, 2026; originally published May 16, 2024.","This is a hiring post describing METR’s goals and plans; it reports no evaluation results or evidence of predictive performance."],"mediaId":null,"correctionReason":null,"id":"d3bcb6489f3d94f01e5fa0e6c263d3f3250bc938eeaf392565ff6bc4cc5a7933","publicationId":"75536446-16c7-bb29-3431-f8eabd2c5c3d","version":1,"sourceUrl":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","sourceTitle":"ML Engineers Needed for New AI R&D Evals Project","publisher":"METR","publishedAt":"2024-05-16T07:00:00.000Z","originalPublished":{"value":"2024-05-16T07:00:00.000Z","precision":"instant","basis":"backfill:feed_published"},"firstObservedAt":"2026-10-18T03:49:29.878Z","createdAt":"2026-11-18T16:33:09.767Z","historical":true,"entityIds":["metr"],"topicIds":["artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://metr.org/blog/2024-05-16-ml-engineers-needed/","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"b5dabf1ee015ed08c02c91ee953240869f8a5b8f16f91ca403d48a3796f8c54c","rubricVersion":"unrecorded","provider":"unrecorded","model":"anthropic/claude-opus-5.5","reasonCodes":["bridge_admitted_by_model","second_opinion_categorical"],"confidence":1,"decidedAt":"2026-10-24T09:45:40.444Z"},"provenance":{"textHash":"4962450b33b392b9f294f5b9bd3a34d981a33e2496b6bcd8bb2e86831d071505","observationId":"9f33981f7cc6c9969ab1ef03090b836aa2fabc9e2b5487acc94f9a582420ec9e","retrievedAt":"2026-10-24T09:47:38.929Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"1c284e92-6904-c5f3-1872-647e92d11105"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"46fe2cdc-bf09-02d8-4bcc-5079c44e9ebc"}},"audit":{"status":"passed","reviewedAt":"2026-11-18T16:29:12.433Z","model":"anthropic/claude-opus-5.5"}}}],"audits":[{"publicationId":"75536446-16c7-bb29-3431-f8eabd2c5c3d","verdict":"passed","findings":[{"dimension":"attribution","status":"pass","explanation":"This is a single METR blog post on METR's own site. The reading consistently attributes the goals, warnings and plans to METR as its own statements and does not present them as established facts.","excerpt":"METR is hiring ML research engineers/scientists to drive the development of these AI R&D evaluations forward."},{"dimension":"summary","status":"pass","explanation":"The reading leads with the actor (METR) and the development: it is hiring ML engineers to build evaluations of AI R&D capabilities. It explains METR's rationale, a possible capability explosion outpacing protective measures, and makes clear the evaluations are planned work rather than results. ML is defined on the page, and threshold evaluations are explained by what they are meant to determine. The question follows from the evidence. Omitting METR's non-profit status and its third-party role advising regulators is a minor gap, not a material one.","excerpt":"METR is developing evaluations for AI R&D capabilities."},{"dimension":"date","status":"pass","explanation":"The reading gives May 16, 2024, which matches the UTC calendar day of evidence.originalPublished and the URL path. The evidence is a publisher page, not an archive copy. The plans are reported as intentions, and there is no hindsight framing or later event.","excerpt":null},{"dimension":"support","status":"pass","explanation":"Each claim matches a verbatim excerpt. The summary's statements are supported: hard-to-bound risk, the need for robust tests, the hope to track progress smoothly, and the bottleneck. The note that the post reports no predictive results accurately reflects a hiring post. The question does not assume an outcome.","excerpt":"Over time, we’d like to build AI R&D evaluations that smoothly track progress, so evaluators aren’t caught by surprise."},{"dimension":"media","status":"pass","explanation":"No media was selected and no verified media candidates were supplied, so leaving the media empty is appropriate.","excerpt":null}],"createdAt":"2026-11-18T16:29:12.433Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}