{"schemaVersion":1,"edition":{"id":"jobs-quality-2026-10-29","authoredAt":"2026-10-29","evidenceReviewedAt":"2026-10-29","scope":"Selected source review, October 2024–October 2026, centered on technical work and selected US employers, with attributed AI capability and investment arguments. Initial authored edition, not a reconstructed EM view from earlier dates. Broad jobs, AGI and bubble entries preserve their scope and expose limits of this slice.","authors":[{"name":"Astra 6","role":"Editorial synthesis author; revised after the independent audit","model":"gpt-6-astra"}],"reviewers":[{"name":"Opus 5.5","role":"Independent audit of earlier working content; findings informed this revision, not approval of revised content","model":"claude-opus-5-5","reviewedAt":"2026-10-29"}]},"projection":"question","selectedQuestionId":"ai-delegation-reliability","questions":[{"id":"ai-delegation-reliability","text":"When can work be delegated without checking every step?","displayTitle":"When can work be delegated without checking every step?","topicIds":["reliability-in-deployment","long-horizon-planning","ai-capabilities-and-limits--evaluation-and-reliability"],"broader":[{"id":"agi","reason":"Reliability is one relevant condition, not a sufficient definition of general intelligence."},{"id":"ai-and-work","reason":"A dependable colleague can also be a supervised collaborator."}],"narrower":[],"related":[{"id":"ai-work-quality-and-rework","reason":"Inspect the cost of detecting and correcting failures."}],"read":{"gapLabel":"Assistance and delegation need different evidence","basis":"Observed assistance and reported output gains do not establish reliable autonomous completion. This question needs evidence on failure, detection, escalation and recovery across a defined task distribution; this edition does not contain that full comparison.","alternatives":["Humans may remain effective supervisors while models perform more implementation.","Redesigning tasks or narrowing permissions may enable dependable bounded delegation before broad autonomy."],"assumptions":["The cost and detectability of failure matter alongside the average result."],"revisionTriggers":["Comparable completion and failure evidence with task, tool, oversight and stopping conditions specified."],"estimate":{"status":"not-estimated","reason":"This selected evidence does not justify a probability, date or numerical outcome range for this question."},"sourceRefs":[{"sourceId":"anthropic-workforce-2026","spanIds":["anth-bottleneck"],"role":"context"},{"sourceId":"amodei-loving-grace-2024","spanIds":["powerful-ai-timing"],"role":"context"},{"sourceId":"metr-productivity-update-2026","spanIds":["selection"],"role":"limits"}],"positionRefs":["amodei-powerful-ai","metr-followup-method"],"connectionRefs":["quality-to-reliability"],"gapRefs":[{"id":"delegation-failure-and-recovery","reason":"A claimed definition and supervised workplace use do not supply deployment failure/recovery evidence."}]}}],"navigationQuestions":[{"id":"jobs","text":"AI taking most jobs: evidence for or against it, including automation and displacement, job creation and new occupations, hiring, employment, wages, productivity or how gains and losses are distributed.","displayTitle":"Will AI take most jobs?","topicIds":["work-and-business","job-displacement-and-creation","wages-and-labor-share"],"broader":[],"narrower":[{"id":"ai-and-work","reason":"Examine whether task capability becomes dependable help in real work."},{"id":"ai-entry-and-worker-gains","reason":"Follow who can enter work and share its benefits."}],"related":[{"id":"agi","reason":"Broader capability could change the set of automatable tasks, but is not employment itself."},{"id":"bubble","reason":"Adoption economics connect workplace outcomes to investment returns, without making them equivalent."}]},{"id":"ai-and-work","text":"When does an AI become a dependable colleague?","displayTitle":"When does AI become a dependable colleague?","topicIds":["work-and-business","human-ai-collaboration","reliability-in-deployment"],"broader":[{"id":"jobs","reason":"Dependable assistance may change tasks and labor demand, with different effects across roles."}],"narrower":[{"id":"ai-work-quality-and-rework","reason":"Count the checking and correction needed for acceptable work."},{"id":"ai-delegation-reliability","reason":"Examine what happens when a human stops checking each step."}],"related":[]},{"id":"ai-work-quality-and-rework","text":"Does AI reduce the work needed to produce an acceptable result?","displayTitle":"Does AI reduce the work needed for an acceptable result?","topicIds":["quality-and-rework","human-ai-collaboration","economics-of-adoption"],"broader":[{"id":"ai-and-work","reason":"Checking and correction are part of dependability."}],"narrower":[],"related":[{"id":"ai-delegation-reliability","reason":"Low average rework does not establish safe unsupervised operation."},{"id":"ai-deployment-returns","reason":"Net workflow value is one input to sustainable returns."},{"id":"ai-entry-and-worker-gains","reason":"Review and training work may be redistributed across experience levels."}]},{"id":"ai-entry-and-worker-gains","text":"Who gets an entry point, and who shares the gains?","displayTitle":"Who gets an entry point and shares the gains?","topicIds":["job-displacement-and-creation","worker-bargaining","education-and-learning","wages-and-labor-share"],"broader":[{"id":"jobs","reason":"Track distribution and opportunity alongside total employment."}],"narrower":[],"related":[{"id":"ai-work-quality-and-rework","reason":"Who performs review can affect skill development and the distribution of work."}]},{"id":"agi","text":"AGI within a few years: progress, limits, timelines or forecasts for broadly capable systems, including disagreement.","displayTitle":"Is AGI a few years away?","topicIds":["ai-capabilities-and-limits","long-horizon-planning"],"broader":[],"narrower":[{"id":"ai-delegation-reliability","reason":"Test one practical dependency without substituting it for the whole AGI question."}],"related":[{"id":"jobs","reason":"Capability does not determine deployment, demand or employment on its own."}]},{"id":"ai-delegation-reliability","text":"When can work be delegated without checking every step?","displayTitle":"When can work be delegated without checking every step?","topicIds":["reliability-in-deployment","long-horizon-planning","ai-capabilities-and-limits--evaluation-and-reliability"],"broader":[{"id":"agi","reason":"Reliability is one relevant condition, not a sufficient definition of general intelligence."},{"id":"ai-and-work","reason":"A dependable colleague can also be a supervised collaborator."}],"narrower":[],"related":[{"id":"ai-work-quality-and-rework","reason":"Inspect the cost of detecting and correcting failures."}]},{"id":"bubble","text":"An AI bubble: financing, valuations, capital expenditure, compute investment, revenues, demand or sustainable returns.","displayTitle":"Is there an AI bubble?","topicIds":["finance-and-the-economy","economics-of-adoption"],"broader":[],"narrower":[{"id":"ai-deployment-returns","reason":"Begin with value to a deploying organization, then identify the additional evidence needed for investor returns."}],"related":[{"id":"jobs","reason":"Employment effects and investment outcomes can move differently."}]},{"id":"ai-deployment-returns","text":"When does a useful AI deployment become worth its total cost?","displayTitle":"When is an AI deployment worth its total cost?","topicIds":["economics-of-adoption","work-and-business--adoption-and-deployment","finance-and-the-economy"],"broader":[{"id":"bubble","reason":"Customer value is one dependency; supplier margins, capital cost and valuation need separate evidence."}],"narrower":[],"related":[{"id":"ai-work-quality-and-rework","reason":"Review and correction can affect total cost and value."},{"id":"jobs","reason":"A positive deployment return does not specify whether employment expands or contracts."}]}],"sources":[{"id":"anthropic-workforce-2026","title":"How employment is changing at Anthropic","url":"https://www.anthropic.com/research/employment-at-anthropic","publishedAt":"2026-10-16","dateBasis":"Publisher publication date, checked on original page.","access":{"status":"public-page","note":"Original publisher page and exact retained source reviewed. Source retention does not establish every downstream reuse or training right.","checkedAt":"2026-10-29"},"provenance":{"kind":"stream-snapshot","workId":"18f3c3ca31491462a8ab52f797ebadd2a7087803317e69a5152e2a0c4cfbed65","snapshotId":"a633ce13b2d661ad9c008ac99d3357f15a80c4bcd91009290d11712c2f565793","textHash":"8ea3533aa08b1f63c143b0a383911da93e88f464edd2d8e61edebf2a97a42232","reason":"Exact retained Stream work/snapshot and SHA256 verified against the 29 October SELECT-only source packet. That lookup found no joined public Stream or bridge reading revision; no reading identity is invented."},"spans":[{"id":"anth-bottleneck","text":"This acceleration makes human attention the bottleneck for agent output.","locator":"Retained extracted text, character offsets [767, 839). Context (EM paraphrase): The authors report rising internal output alongside agent use and interpret human attention as a constraint. Adoption is not randomized; the report does not establish a general causal productivity effect."}],"readingRefs":[]},{"id":"metr-productivity-2025","title":"Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity","url":"https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/","publishedAt":"2025-07-10","dateBasis":"Publication date displayed by the original publisher.","access":{"status":"public-page","note":"Original publisher page and exact retained source reviewed. Source retention does not establish every downstream reuse or training right.","checkedAt":"2026-10-29"},"provenance":{"kind":"stream-snapshot","workId":"eeef7ca24568c0fe592aa98201faccea099ec6507dbd9476c8bea56310f28743","snapshotId":"1b377f2c5f76b4960227c32060fef0433519ce76239947c7e836402cb9a945ac","textHash":"78a7743c1dda83f49b4d646c134243e5e6c38bed0eb7aa33d124a37e3971f7df","reason":"Exact retained Stream work/snapshot and SHA256 verified against the 29 October SELECT-only source packet. That lookup found no joined public Stream or bridge reading revision; no reading identity is invented."},"spans":[{"id":"setting","text":"16 experienced developers","locator":"Retained extracted text, character offsets [1527, 1552). Context (EM paraphrase): Methods: developers worked on 246 real issues in familiar open-source repositories using early-2025 tools. This population does not represent all developers or later systems."},{"id":"completion-time","text":"they take 19% longer to complete issues","locator":"Retained extracted text, character offsets [2468, 2507). Context (EM paraphrase): Core Result, first paragraph; interpret with Methods and the immediately following limitations table. Reported result for the 2025 study, not a current universal estimate."}],"readingRefs":[]},{"id":"metr-productivity-update-2026","title":"We are Changing our Developer Productivity Experiment Design","url":"https://metr.org/blog/2026-02-24-uplift-update/","publishedAt":"2026-02-24","dateBasis":"Publication date displayed by the original publisher.","access":{"status":"public-page","note":"Original publisher page and exact retained source reviewed. Source retention does not establish every downstream reuse or training right.","checkedAt":"2026-10-29"},"provenance":{"kind":"stream-snapshot","workId":"54d658ff870e4bbb0638a847bac138a3b543d42ac48e761c915c570500f1cc5c","snapshotId":"8fd5f006c0dc55d1e20448712d3f4b177c53cd92219b9aa3ed71b70d0ed13eca","textHash":"81b6c9b7d710e11ed96cf12de5f307e8551454cfdcc9f6c66c84f0a657b9adfb","reason":"Exact retained Stream work/snapshot and SHA256 verified against the 29 October SELECT-only source packet. That lookup found no joined public Stream or bridge reading revision; no reading identity is invented."},"spans":[{"id":"selection","text":"the data from our new experiment gives us an unreliable signal of the current productivity effect of AI tools","locator":"Retained extracted text, character offsets [431, 540). Context (EM paraphrase): Opening discussion, paragraph beginning Unfortunately. Subsequent paragraphs describe participation, task selection and time measurement."}],"readingRefs":[]},{"id":"amodei-loving-grace-2024","title":"Machines of Loving Grace","url":"https://darioamodei.com/essay/machines-of-loving-grace","publishedAt":"2024-10","dateBasis":"October 2024 is printed in the retained essay. The work metadata has no publication timestamp; month precision is preserved.","access":{"status":"retained-public-essay","note":"Original personal essay retained in Stream and reviewed here; not an institutional position or a new 2026 statement. Public availability does not establish every reuse right.","checkedAt":"2026-10-29"},"provenance":{"kind":"stream-snapshot","workId":"6aa2c807bdcb81da04a30b6cf0b1c0620353219816ab224111920bdc8c02bc32","snapshotId":"07fedfacf7c11b925e82d6c8a319ee710df91c8b7160fae3d7ca3fda1da6af45","textHash":"d8f37e4a35c07240bca22712dece7667d41c3e3efb9f28b7ca62c4c5f807442a","reason":"Exact original Stream identity and SHA256 verified against the retained source packet."},"spans":[{"id":"powerful-ai-timing","text":"as early as 2026","locator":"Retained extracted text, character offsets [6255, 6271). Context (EM paraphrase): Amodei says powerful AI could arrive that early but could also take much longer. His working definition combines exceptional cognitive capability, human-like computer interfaces and autonomous work over hours, days or weeks. The essay assumes relatively near arrival to explore later effects; it is not a demonstrated outcome or a standardized AGI forecast."}],"readingRefs":[]}],"positions":[{"id":"anthropic-work-organization","speaker":"Nathan Wilmers, Regina Catipon, Maxim Massenkoff and Peter McCrory","role":"Anthropic Institute researchers reporting on their employer","statement":"The report describes increased output alongside rising demands on human attention as AI use spreads internally.","conditions":["Observations at one AI company; not a controlled estimate of the causal effect or of other employers’ results."],"sourceRefs":[{"sourceId":"anthropic-workforce-2026","spanIds":["anth-bottleneck"],"role":"supports"}],"statedAt":"2026-10-16","dateBasis":"Dated authored publication; this is not the date of every underlying observation or interview.","attribution":"direct"},{"id":"metr-followup-method","speaker":"Joel Becker, Nate Rush, Tom Cunningham, David Rein and Khalid Mahamud","role":"METR authors of the follow-up methods update","statement":"They consider greater developer speedup plausible by early 2026, but report that participation, task selection and time-accounting problems make the follow-up estimate unreliable.","conditions":["This is a later team’s assessment of a changed experiment, not the same measurement repeated cleanly."],"sourceRefs":[{"sourceId":"metr-productivity-update-2026","spanIds":["selection"],"role":"supports"}],"statedAt":"2026-02-24","dateBasis":"Publication date of the authored source.","attribution":"direct"},{"id":"amodei-powerful-ai","speaker":"Dario Amodei","role":"Anthropic chief executive writing a personal essay","statement":"In October 2024, Amodei suggested that powerful AI could arrive as early as 2026 while explicitly allowing a much longer wait. He defined it through exceptional breadth and autonomous work, then explored possible consequences under that assumption.","conditions":["The essay’s definition is not a universally agreed AGI endpoint.","An earliest possibility is not a median date, deadline or claim that the system exists."],"sourceRefs":[{"sourceId":"amodei-loving-grace-2024","spanIds":["powerful-ai-timing"],"role":"supports"}],"statedAt":"2024-10","dateBasis":"Month printed in the essay; the underlying capture has no publication timestamp.","attribution":"direct"}],"paths":[{"id":"delegation-bounded-progress","questionId":"ai-delegation-reliability","outcome":"Bounded delegation becomes dependable","scope":"EM conditional scenario, not an estimated probability or a claim that this outcome has occurred.","conditions":["The tasks and consequences are specified.","Failures can be detected, contained or escalated at acceptable cost."],"sourceRefs":[{"sourceId":"amodei-loving-grace-2024","spanIds":["powerful-ai-timing"],"role":"context"},{"sourceId":"anthropic-workforce-2026","spanIds":["anth-bottleneck"],"role":"context"}],"positionRefs":["amodei-powerful-ai"]},{"id":"delegation-supervision-persists","questionId":"ai-delegation-reliability","outcome":"Assistance improves but close supervision persists","scope":"EM conditional scenario, not an estimated probability or a claim that this outcome has occurred.","conditions":["Errors, context requirements or recovery costs remain consequential.","People still need to inspect or repair work before relying on it."],"sourceRefs":[{"sourceId":"anthropic-workforce-2026","spanIds":["anth-bottleneck"],"role":"context"},{"sourceId":"metr-productivity-update-2026","spanIds":["selection"],"role":"limits"}],"positionRefs":["anthropic-work-organization","metr-followup-method"]}],"connections":[{"id":"quality-to-reliability","from":{"kind":"question","id":"ai-work-quality-and-rework"},"to":{"kind":"question","id":"ai-delegation-reliability"},"relation":"informs","rationale":"Human review can make assisted output acceptable while concealing failures that would matter under autonomous delegation. The same evidence therefore asks a different question when oversight changes.","limits":["An authored cross-source connection, not a measured causal coefficient.","Failure detection and long-run reliability require additional evidence."],"sourceRefs":[{"sourceId":"anthropic-workforce-2026","spanIds":["anth-bottleneck"],"role":"context"},{"sourceId":"metr-productivity-2025","spanIds":["setting","completion-time"],"role":"context"},{"sourceId":"metr-productivity-update-2026","spanIds":["selection"],"role":"context"}],"positionRefs":[],"basis":"em-inference"}],"changes":[],"questionVersions":{"ai-delegation-reliability":"588c751e892c7c15d373bb1f193769517fef0faf9f15512d158344de13c1ea42"},"availableEditions":[{"id":"jobs-quality-2026-10-29","authoredAt":"2026-10-29","evidenceReviewedAt":"2026-10-29","scope":"Selected source review, October 2024–October 2026, centered on technical work and selected US employers, with attributed AI capability and investment arguments. Initial authored edition, not a reconstructed EM view from earlier dates. Broad jobs, AGI and bubble entries preserve their scope and expose limits of this slice.","authors":[{"name":"Astra 6","role":"Editorial synthesis author; revised after the independent audit","model":"gpt-6-astra"}],"reviewers":[{"name":"Opus 5.5","role":"Independent audit of earlier working content; findings informed this revision, not approval of revised content","model":"claude-opus-5-5","reviewedAt":"2026-10-29"}]}]}