{"id":"cc93bdfa1ed4365375823d6b33950f0fe5515afe7744bf06a9cccd7b273d8615","title":"Ethan Mollick argues in One Useful Thing that AI’s value depends on how people use it","argument":"Ethan Mollick’s piece was published September 29, 2025. He argues that AI can now perform economically valuable tasks, including reproducing research from existing data, but that doing tasks is not the same as replacing whole jobs. In his account, Claude Sonnet 4.5 worked through replication files and results for several papers; Mollick says spot checks and a second AI model’s reproduction checked out, while some papers could not be accessed. He sees a possible way to scale checks on published research, while noting barriers including benchmarking for accuracy and fairness.\n\nMollick also warns that agents can multiply work people do not need: he asked Claude to turn a memo into presentations and received 17 versions. He argues that the important distinction is whether people use their judgment to decide what is worth doing, rather than simply producing more. He relays an OpenAI paper’s suggested workflow: delegate a first pass, review it, try corrections if needed, and do the work oneself if that fails. The paper estimates this could make work faster and cheaper while retaining human control. Whether AI can reproduce research accurately and fairly at scale, and how organizations will choose to use agents, remain open questions.","topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation","artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 23, 2026; originally published September 29, 2025. Mollick’s account of reproducing research describes his tests and spot checks; the piece does not report a systematic evaluation of accuracy or fairness across published papers.","speaker":{"name":"One Useful Thing","roleAtSource":"Publisher"},"source":{"title":"Real AI Agents and Real Work","publisher":"One Useful Thing","format":"Article","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","publishedAt":"2025-09-29T18:52:42.000Z","dateBasis":"Date supplied by the publisher’s feed."},"evidence":[{"text":"Ethan Mollick argues that AI can perform economically valuable tasks, but that automating tasks does not by itself replace whole jobs.","excerpt":"AIs have quietly crossed a threshold: they can now perform real, economically relevant work.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick reports that Claude Sonnet 4.5 worked through replication data for an economics paper and that his spot checks and a second AI model’s reproduction checked out.","excerpt":"I spot checked the results and had another AI model, GPT-5 Pro, reproduce the reproduction. It all checked out.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick says some papers could not be tested because of file-size limits or problems with their replication data, and identifies accuracy and fairness benchmarking as barriers to broader use.","excerpt":"There are still barriers to doing this, including benchmarking for accuracy and fairness","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick warns that agents can produce unnecessary work, describing how Claude generated 17 PowerPoints from a corporate memo.","excerpt":"Until I got 17 different PowerPoints. That is too many PowerPoints.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick argues that people’s judgment about what work is worth doing will shape whether AI makes them more capable or merely more productive.","excerpt":"By using our judgement in deciding what’s worth doing, not just what can be done, we can ensure these tools make us more capable, not just more productive.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"The source is dated September 29, 2025.","excerpt":"Ethan Mollick\nSep 29, 2025","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What accuracy and fairness standards would be needed before AI could reproduce published research at scale?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Ethan Mollick argues in One Useful Thing that AI’s value depends on how people use it","summary":"Ethan Mollick’s piece was published September 29, 2025. He argues that AI can now perform economically valuable tasks, including reproducing research from existing data, but that doing tasks is not the same as replacing whole jobs. In his account, Claude Sonnet 4.5 worked through replication files and results for several papers; Mollick says spot checks and a second AI model’s reproduction checked out, while some papers could not be accessed. He sees a possible way to scale checks on published research, while noting barriers including benchmarking for accuracy and fairness.\n\nMollick also warns that agents can multiply work people do not need: he asked Claude to turn a memo into presentations and received 17 versions. He argues that the important distinction is whether people use their judgment to decide what is worth doing, rather than simply producing more. He relays an OpenAI paper’s suggested workflow: delegate a first pass, review it, try corrections if needed, and do the work oneself if that fails. The paper estimates this could make work faster and cheaper while retaining human control. Whether AI can reproduce research accurately and fairly at scale, and how organizations will choose to use agents, remain open questions.","question":"What accuracy and fairness standards would be needed before AI could reproduce published research at scale?","claims":[{"text":"Ethan Mollick argues that AI can perform economically valuable tasks, but that automating tasks does not by itself replace whole jobs.","excerpt":"AIs have quietly crossed a threshold: they can now perform real, economically relevant work.","support":"source-stated"},{"text":"Mollick reports that Claude Sonnet 4.5 worked through replication data for an economics paper and that his spot checks and a second AI model’s reproduction checked out.","excerpt":"I spot checked the results and had another AI model, GPT-5 Pro, reproduce the reproduction. It all checked out.","support":"source-stated"},{"text":"Mollick says some papers could not be tested because of file-size limits or problems with their replication data, and identifies accuracy and fairness benchmarking as barriers to broader use.","excerpt":"There are still barriers to doing this, including benchmarking for accuracy and fairness","support":"source-stated"},{"text":"Mollick warns that agents can produce unnecessary work, describing how Claude generated 17 PowerPoints from a corporate memo.","excerpt":"Until I got 17 different PowerPoints. That is too many PowerPoints.","support":"source-stated"},{"text":"Mollick argues that people’s judgment about what work is worth doing will shape whether AI makes them more capable or merely more productive.","excerpt":"By using our judgement in deciding what’s worth doing, not just what can be done, we can ensure these tools make us more capable, not just more productive.","support":"source-stated"},{"text":"The source is dated September 29, 2025.","excerpt":"Ethan Mollick\nSep 29, 2025","support":"source-stated"}],"limitations":["Text as retrieved on October 23, 2026; originally published September 29, 2025.","Mollick’s account of reproducing research describes his tests and spot checks; the piece does not report a systematic evaluation of accuracy or fairness across published papers."],"mediaId":null,"correctionReason":null,"id":"cc93bdfa1ed4365375823d6b33950f0fe5515afe7744bf06a9cccd7b273d8615","publicationId":"fe9c690e-c71c-ed07-4a0e-690d05193565","version":1,"sourceUrl":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","sourceTitle":"Real AI Agents and Real Work","publisher":"One Useful Thing","publishedAt":"2025-09-29T18:52:42.000Z","originalPublished":{"value":"2025-09-29T18:52:42.000Z","precision":"instant","basis":"backfill:feed_published"},"firstObservedAt":"2026-09-23T05:45:06.331Z","createdAt":"2026-11-12T01:03:23.472Z","historical":true,"entityIds":["one-useful-thing"],"topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation","artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"68fd06d5776ecdb976ea0b8ce998910220b3e7ed887409fe2c6712f74af9c473","rubricVersion":"unrecorded","provider":"vercel","model":"typesafe-ai/jev","reasonCodes":["bridge_admitted_by_model"],"confidence":0.65,"decidedAt":"2026-10-23T20:13:47.623Z"},"provenance":{"textHash":"0347a5d6dd15753c9220a5e56f8a8da9052cd955497cffbb9b27900b4f6ad425","observationId":"9d12fbda725b166a1b68975fcb938e9fd268763b6877a3caf9baafea26c6f91c","retrievedAt":"2026-10-23T20:15:35.062Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"7352b4ed-9b9a-728e-3d2a-2479c05d6da8"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"94fca478-3a21-50f1-fa43-cc47e803d8ae"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-12T00:58:42.374Z","model":"anthropic/claude-opus-5.5"}},"canonicalUrl":"https://escapemeridian.com/readings/ethan-mollick-argues-in-one-useful-thing-that-ai-s-value-cc93bdfa","formats":{"html":"/readings/ethan-mollick-argues-in-one-useful-thing-that-ai-s-value-cc93bdfa","markdown":"/readings/ethan-mollick-argues-in-one-useful-thing-that-ai-s-value-cc93bdfa.md","json":"/readings/ethan-mollick-argues-in-one-useful-thing-that-ai-s-value-cc93bdfa.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":[],"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-benchmark","name":"Benchmark","explanation":"A defined set of tests used to compare systems. Results depend on the tasks and measures selected.","referenceUrl":"https://arxiv.org/abs/2211.09110"}],"entities":[{"id":"ethan-mollick","name":"Ethan Mollick","kind":"person","context":"Ethan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.","referenceUrls":["https://mgmt.wharton.upenn.edu/profile/emollick/","https://www.oneusefulthing.org/"],"path":"/entities/ethan-mollick","json":"/entities/ethan-mollick.json","markdown":"/entities/ethan-mollick.md"},{"id":"one-useful-thing","name":"One Useful Thing","kind":"organization","context":"One Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.","referenceUrls":["https://www.oneusefulthing.org/"],"path":"/entities/one-useful-thing","json":"/entities/one-useful-thing.json","markdown":"/entities/one-useful-thing.md"},{"id":"claude","name":"Claude","kind":"model","context":"Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.","referenceUrls":["https://claude.com/product/overview","https://www.anthropic.com/news","https://www.anthropic.com/research/team/economics"],"path":"/entities/claude","json":"/entities/claude.json","markdown":"/entities/claude.md"},{"id":"openai","name":"OpenAI","kind":"organization","context":"OpenAI develops AI models and products and publishes research, release announcements and evaluations.","referenceUrls":["https://openai.com/news/"],"path":"/entities/openai","json":"/entities/openai.json","markdown":"/entities/openai.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"cc93bdfa1ed4365375823d6b33950f0fe5515afe7744bf06a9cccd7b273d8615","revision":"fe9c690e-c71c-ed07-4a0e-690d05193565","revisionKind":"publication-id","title":"Ethan Mollick argues in One Useful Thing that AI’s value depends on how people use it","url":"https://escapemeridian.com/readings/ethan-mollick-argues-in-one-useful-thing-that-ai-s-value-cc93bdfa"},"source":{"type":"source","id":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","revision":"0347a5d6dd15753c9220a5e56f8a8da9052cd955497cffbb9b27900b4f6ad425","revisionKind":"sha256-source-text","title":"Real AI Agents and Real Work","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"cc93bdfa1ed4365375823d6b33950f0fe5515afe7744bf06a9cccd7b273d8615","title":"Ethan Mollick argues in One Useful Thing that AI’s value depends on how people use it","argument":"Ethan Mollick’s piece was published September 29, 2025. He argues that AI can now perform economically valuable tasks, including reproducing research from existing data, but that doing tasks is not the same as replacing whole jobs. In his account, Claude Sonnet 4.5 worked through replication files and results for several papers; Mollick says spot checks and a second AI model’s reproduction checked out, while some papers could not be accessed. He sees a possible way to scale checks on published research, while noting barriers including benchmarking for accuracy and fairness.\n\nMollick also warns that agents can multiply work people do not need: he asked Claude to turn a memo into presentations and received 17 versions. He argues that the important distinction is whether people use their judgment to decide what is worth doing, rather than simply producing more. He relays an OpenAI paper’s suggested workflow: delegate a first pass, review it, try corrections if needed, and do the work oneself if that fails. The paper estimates this could make work faster and cheaper while retaining human control. Whether AI can reproduce research accurately and fairly at scale, and how organizations will choose to use agents, remain open questions.","topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation","artificial-general-intelligence"],"domain":"other","claimType":"attributed reading","limits":"Text as retrieved on October 23, 2026; originally published September 29, 2025. Mollick’s account of reproducing research describes his tests and spot checks; the piece does not report a systematic evaluation of accuracy or fairness across published papers.","speaker":{"name":"One Useful Thing","roleAtSource":"Publisher"},"source":{"title":"Real AI Agents and Real Work","publisher":"One Useful Thing","format":"Article","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","publishedAt":"2025-09-29T18:52:42.000Z","dateBasis":"Date supplied by the publisher’s feed."},"evidence":[{"text":"Ethan Mollick argues that AI can perform economically valuable tasks, but that automating tasks does not by itself replace whole jobs.","excerpt":"AIs have quietly crossed a threshold: they can now perform real, economically relevant work.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick reports that Claude Sonnet 4.5 worked through replication data for an economics paper and that his spot checks and a second AI model’s reproduction checked out.","excerpt":"I spot checked the results and had another AI model, GPT-5 Pro, reproduce the reproduction. It all checked out.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick says some papers could not be tested because of file-size limits or problems with their replication data, and identifies accuracy and fairness benchmarking as barriers to broader use.","excerpt":"There are still barriers to doing this, including benchmarking for accuracy and fairness","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick warns that agents can produce unnecessary work, describing how Claude generated 17 PowerPoints from a corporate memo.","excerpt":"Until I got 17 different PowerPoints. That is too many PowerPoints.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"Mollick argues that people’s judgment about what work is worth doing will shape whether AI makes them more capable or merely more productive.","excerpt":"By using our judgement in deciding what’s worth doing, not just what can be done, we can ensure these tools make us more capable, not just more productive.","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"},{"text":"The source is dated September 29, 2025.","excerpt":"Ethan Mollick\nSep 29, 2025","url":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","relationship":"Source-stated"}],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","checkedAt":null,"items":[]},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"What accuracy and fairness standards would be needed before AI could reproduce published research at scale?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Ethan Mollick argues in One Useful Thing that AI’s value depends on how people use it","summary":"Ethan Mollick’s piece was published September 29, 2025. He argues that AI can now perform economically valuable tasks, including reproducing research from existing data, but that doing tasks is not the same as replacing whole jobs. In his account, Claude Sonnet 4.5 worked through replication files and results for several papers; Mollick says spot checks and a second AI model’s reproduction checked out, while some papers could not be accessed. He sees a possible way to scale checks on published research, while noting barriers including benchmarking for accuracy and fairness.\n\nMollick also warns that agents can multiply work people do not need: he asked Claude to turn a memo into presentations and received 17 versions. He argues that the important distinction is whether people use their judgment to decide what is worth doing, rather than simply producing more. He relays an OpenAI paper’s suggested workflow: delegate a first pass, review it, try corrections if needed, and do the work oneself if that fails. The paper estimates this could make work faster and cheaper while retaining human control. Whether AI can reproduce research accurately and fairly at scale, and how organizations will choose to use agents, remain open questions.","question":"What accuracy and fairness standards would be needed before AI could reproduce published research at scale?","claims":[{"text":"Ethan Mollick argues that AI can perform economically valuable tasks, but that automating tasks does not by itself replace whole jobs.","excerpt":"AIs have quietly crossed a threshold: they can now perform real, economically relevant work.","support":"source-stated"},{"text":"Mollick reports that Claude Sonnet 4.5 worked through replication data for an economics paper and that his spot checks and a second AI model’s reproduction checked out.","excerpt":"I spot checked the results and had another AI model, GPT-5 Pro, reproduce the reproduction. It all checked out.","support":"source-stated"},{"text":"Mollick says some papers could not be tested because of file-size limits or problems with their replication data, and identifies accuracy and fairness benchmarking as barriers to broader use.","excerpt":"There are still barriers to doing this, including benchmarking for accuracy and fairness","support":"source-stated"},{"text":"Mollick warns that agents can produce unnecessary work, describing how Claude generated 17 PowerPoints from a corporate memo.","excerpt":"Until I got 17 different PowerPoints. That is too many PowerPoints.","support":"source-stated"},{"text":"Mollick argues that people’s judgment about what work is worth doing will shape whether AI makes them more capable or merely more productive.","excerpt":"By using our judgement in deciding what’s worth doing, not just what can be done, we can ensure these tools make us more capable, not just more productive.","support":"source-stated"},{"text":"The source is dated September 29, 2025.","excerpt":"Ethan Mollick\nSep 29, 2025","support":"source-stated"}],"limitations":["Text as retrieved on October 23, 2026; originally published September 29, 2025.","Mollick’s account of reproducing research describes his tests and spot checks; the piece does not report a systematic evaluation of accuracy or fairness across published papers."],"mediaId":null,"correctionReason":null,"id":"cc93bdfa1ed4365375823d6b33950f0fe5515afe7744bf06a9cccd7b273d8615","publicationId":"fe9c690e-c71c-ed07-4a0e-690d05193565","version":1,"sourceUrl":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","sourceTitle":"Real AI Agents and Real Work","publisher":"One Useful Thing","publishedAt":"2025-09-29T18:52:42.000Z","originalPublished":{"value":"2025-09-29T18:52:42.000Z","precision":"instant","basis":"backfill:feed_published"},"firstObservedAt":"2026-09-23T05:45:06.331Z","createdAt":"2026-11-12T01:03:23.472Z","historical":true,"entityIds":["one-useful-thing"],"topicIds":["work-and-business--jobs-and-occupations","job-displacement-and-creation","artificial-general-intelligence"],"media":[],"sourceMedia":{"status":"unchecked","sourceUrl":"https://www.oneusefulthing.org/p/real-ai-agents-and-real-work","checkedAt":null,"items":[]},"editorialDecision":{"outcome":"accept","mediaId":null,"inputHash":"68fd06d5776ecdb976ea0b8ce998910220b3e7ed887409fe2c6712f74af9c473","rubricVersion":"unrecorded","provider":"vercel","model":"typesafe-ai/jev","reasonCodes":["bridge_admitted_by_model"],"confidence":0.65,"decidedAt":"2026-10-23T20:13:47.623Z"},"provenance":{"textHash":"0347a5d6dd15753c9220a5e56f8a8da9052cd955497cffbb9b27900b4f6ad425","observationId":"9d12fbda725b166a1b68975fcb938e9fd268763b6877a3caf9baafea26c6f91c","retrievedAt":"2026-10-23T20:15:35.062Z","kind":"publisher-page","preparer":{"model":"openai/gpt-6-luna","runId":"7352b4ed-9b9a-728e-3d2a-2479c05d6da8"},"reviewer":{"model":"anthropic/claude-opus-5.5","runId":"94fca478-3a21-50f1-fa43-cc47e803d8ae"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-12T00:58:42.374Z","model":"anthropic/claude-opus-5.5"}}}],"audits":[{"publicationId":"fe9c690e-c71c-ed07-4a0e-690d05193565","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading correctly attributes the essay to Ethan Mollick in One Useful Thing. It keeps his arguments and reported tests framed as his claims, and attributes the suggested workflow to the OpenAI paper he relays.","excerpt":"Real AI Agents and Real Work\nThe race between human-centered work and infinite PowerPoints\nEthan Mollick"},{"dimension":"summary","status":"uncertain","explanation":"The reading is largely faithful, but 'agents' is central to it and never explained. The definitions list covers only Benchmark. Smallest repair: add the source's own plain description, AI that is given a task and can plan and use tools such as coding or web search to complete it, or add a term-library entry. The reading also omits Mollick's opening evidence: the OpenAI expert-task test, where human experts won only narrowly, and the METR task-length trend. That trims context for his threshold claim but does not change his position.","excerpt":"The dream of autonomous AI agents, which, when given a task, can plan and use tools (coding, web search) to accomplish it, seemed far away."},{"dimension":"date","status":"pass","explanation":"The instant 2025-09-29T18:52:42Z falls on the UTC day September 29, 2025, which matches the source dateline. There is no hindsight framing, and the kind is publisher-page, so no archive disclosure is needed. The modification date equals the publication date.","excerpt":"Ethan Mollick\nSep 29, 2025"},{"dimension":"support","status":"pass","explanation":"Each claim is supported by the source text: the threshold and task-versus-job argument, the Claude replication with spot checks and a GPT-5 Pro cross-check, results on other papers, file-size and data limits, barriers around accuracy and fairness benchmarking, the 17 PowerPoints, and the judgment conclusion. The OpenAI workflow is presented as the paper's estimate. The question follows from the barriers the source names.","excerpt":"I tried this on several other papers with similarly good results, though some were inaccessible due to file size limitations or issues with the replication data provided."},{"dimension":"media","status":"pass","explanation":"No media was selected, and no verified candidates were supplied, so no media is the appropriate choice.","excerpt":null}],"createdAt":"2026-11-12T00:58:42.374Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}