· Apr 10, 2026
David Rein’s METR linkpost points readers to MirrorCode
Markdown · JSON · Original source
David Rein’s page was published on April 10, 2026. It is a linkpost about MirrorCode, a project that METR says it funded and co-developed with . The page directs readers to Epoch AI’s blog post for more detail; it does not describe the project’s methods or results. Its title frames MirrorCode as evidence that AI can already do some weeks-long coding tasks, but this page alone does not provide the evidence behind that claim. What the project found, and what tasks it tested, remain unknown from this linkpost.
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.
Original source
· Article
Source details
Original source date: Apr 10, 2026 (2026-04-10T07:00:00.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-11-07T12:53:23.142Z. First observed: 2026-10-18T02:57:23.259Z. Retrieved: 2026-10-23T18:29:39.933Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on October 23, 2026; originally published April 10, 2026. This is a linkpost, not a detailed account of MirrorCode’s methods or results; the linked post is not included in the supplied evidence. The page title presents a claim about AI’s ability to do some weeks-long coding tasks, but this page supplies no supporting results.
Open question
What methods and results does ’s linked MirrorCode post report?
Source-stated
METR says it funded and co-developed MirrorCode with Epoch AI.
a project that METR funded and co-developed with Epoch AI.
Source-stated
METR directs readers to Epoch AI’s blog post for more detail.
See Epoch AI’s blog post for more detail: https://epoch.ai/publications/mirrorcode-preliminary-results/
Source-stated
The METR page URL includes the date April 10, 2026.
https://metr.org/blog/2026-04-10-mirrorcode-preliminary-results/
Source and contributor context
METR · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
METR
METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Epoch AI
Epoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Editorial provenance
Version 1. Publication ID: 76d2468a-42be-d454-ee59-563dfb8711b7.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-07T12:49:10.993Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: 5918ae8c6690364e9ca9c66917f3c44dea832f1bb3d6818f5650f0459a53101a. Source text hash: 7e10c23ee3386023a9b78bc87138681ffedbde6dc8f946dc513ddc108b1cfb91.
Revision and source-check history
Version 1 · 2026-11-07T12:53:23.142Z
David Rein’s METR linkpost points readers to MirrorCode
David Rein’s page was published on April 10, 2026. It is a linkpost about MirrorCode, a project that METR says it funded and co-developed with . The page directs readers to Epoch AI’s blog post for more detail; it does not describe the project’s methods or results. Its title frames MirrorCode as evidence that AI can already do some weeks-long coding tasks, but this page alone does not provide the evidence behind that claim. What the project found, and what tasks it tested, remain unknown from this linkpost.
Original source
· Article
Source details
Original source date: Apr 10, 2026 (2026-04-10T07:00:00.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-11-07T12:53:23.142Z. First observed: 2026-10-18T02:57:23.259Z. Retrieved: 2026-10-23T18:29:39.933Z.
Limits
Text as retrieved on October 23, 2026; originally published April 10, 2026. This is a linkpost, not a detailed account of MirrorCode’s methods or results; the linked post is not included in the supplied evidence. The page title presents a claim about AI’s ability to do some weeks-long coding tasks, but this page supplies no supporting results.
Open question
What methods and results does ’s linked MirrorCode post report?
Source-stated
METR says it funded and co-developed MirrorCode with Epoch AI.
a project that METR funded and co-developed with Epoch AI.
Source-stated
METR directs readers to Epoch AI’s blog post for more detail.
See Epoch AI’s blog post for more detail: https://epoch.ai/publications/mirrorcode-preliminary-results/
Source-stated
The METR page URL includes the date April 10, 2026.
https://metr.org/blog/2026-04-10-mirrorcode-preliminary-results/
Source and contributor context
METR · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
METR
METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Epoch AI
Epoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Editorial provenance
Version 1. Publication ID: 76d2468a-42be-d454-ee59-563dfb8711b7.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-07T12:49:10.993Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: 5918ae8c6690364e9ca9c66917f3c44dea832f1bb3d6818f5650f0459a53101a. Source text hash: 7e10c23ee3386023a9b78bc87138681ffedbde6dc8f946dc513ddc108b1cfb91.
passed · 2026-11-07T12:49:10.993Z · anthropic/claude-opus-5.5
Publication: 76d2468a-42be-d454-ee59-563dfb8711b7
attribution — pass: The page is a single METR linkpost. Its citation block names David Rein as the author, so the reading's attribution to Rein and METR fits the evidence. The 'Featured research' blurbs are page furniture, and the reading neither attributes nor uses them. The funding and co-development statement is correctly framed as what METR says.
author = {David Rein},
summary — pass: The summary accurately describes a short linkpost. It names MirrorCode, says METR funded and co-developed it with Epoch AI, and notes the page sends readers to Epoch AI's post for detail. It presents the title's capability claim as framing, not established fact, and states that the methods and results are unknown from this page. The term linkpost is made clear by context. The question is a genuine open issue that follows from the evidence.
This is a linkpost for MirrorCode, a project that METR funded and co-developed with Epoch AI.
date — pass: evidence.originalPublished is 2026-04-10T07:00Z, so its UTC calendar day is April 10, 2026. This matches the reading, the URL path and the citation's year and month. The reading uses no later sidebar content and has no hindsight framing. Evidence kind is publisher-page, so no archive disclosure is needed.
https://metr.org/blog/2026-04-10-mirrorcode-preliminary-results/
support — pass: Each claim is directly supported by a verbatim excerpt: METR funded and co-developed MirrorCode with Epoch AI, it directs readers to Epoch AI's post, and the URL date is correct. The reading adds no benefits, costs or results beyond the source. It correctly notes the title's claim is unsupported on this page.
See Epoch AI’s blog post for more detail: https://epoch.ai/publications/mirrorcode-preliminary-results/
media — pass: No media was selected, and no verified media candidates were supplied. Selecting none is appropriate.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METREpoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Explore Epoch AIMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METREpoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Explore Epoch AIEpoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Explore Epoch AIMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METREpoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Explore Epoch AIMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METREpoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Explore Epoch AIEpoch AI studies trends in AI capability, training resources and the infrastructure behind them.
Explore Epoch AI