One Useful Thing · Jun 30, 2026
One Useful Thing’s Ethan Mollick argues work is shifting from chatbots to agents
Markdown · JSON · Original source
Ethan Mollick published this argument on June 30, 2026: AI use is shifting from people working step by step with chatbots to assigning longer tasks to agents. He describes agents as systems with tools and an environment in which to act, often supported by harnesses or apps. Mollick points to assessments of AI work, experiments with software projects, and studies of workplace use. He argues that agents are being used beyond coding, and that domain expertise—not a user’s profession—was associated with greater success using Claude Code in a separate study. For organizations, his account suggests that assigning and overseeing agent work may become a larger part of some jobs. This is Mollick’s interpretation of developing capabilities and reported studies, not evidence that agents perform every task reliably. He stresses that model performance varies by task and that benchmark results do not always reflect practical performance. Mollick also argues that rapid capability changes can unsettle institutions and markets. How widely this pattern extends beyond the organizations and experiments he describes remains unknown.
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
One Useful Thing · Article
Source details
Original source date: Jun 30, 2026 (2026-06-30T22:18:12.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-10-29T18:04:23.278Z. First observed: 2026-09-23T05:31:20.363Z. Retrieved: 2026-10-23T19:57:36.864Z.
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 June 30, 2026. The source describes selected assessments, experiments, and studies; it does not establish how broadly the reported workplace patterns apply. The article does not provide enough detail here to assess the methods or independent replication of the studies it cites.
Open question
Across organizations beyond those studied, how often does greater domain expertise lead to more successful agent-assisted work?
Source-stated
Mollick argues that AI is increasingly used for longer tasks assigned to agents rather than step-by-step collaboration with chatbots.
So work is increasingly about assigning work to agents, rather than working together with chatbots.
Source-stated
Mollick describes agents as having access to tools and an environment through harnesses, and says apps can further improve model capability.
agents come with extra machinery: harnesses that give the AI access to tools and an environment to act in
Source-stated
The source reports that a separate study found domain experience was associated with greater success using Claude Code.
The more domain experience someone had, the more successful they were in using Claude Code in that domain.
Source-stated
Mollick cautions that open weights models can perform less well in practice than their benchmark results suggest.
the open weights models, while very impressive, do not always perform as well as their benchmarks would indicate
Escape Meridian interpretation
Escape Meridian reads Mollick’s account as suggesting that assigning and overseeing agent work may become more important in some roles; the source does not establish how widely this applies.
the best way to use agents is to think of yourself as a manager.
Source and contributor context
One Useful Thing · 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.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
One Useful Thing
One Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.
Ethan Mollick
Ethan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
Claude
Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.
Editorial provenance
Version 1. Publication ID: 982fd41d-e295-9818-19cf-a448883c82dd.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-10-29T18:03:03.242Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: d86ad3398047795cce0c1f10ebe8b17fb25ca853b23447b225e031d158fbefa0. Source text hash: 5d1649360c59fa5184e4094116da81e82362643d226965e44446bca54700f836.
Revision and source-check history
Version 1 · 2026-10-29T18:04:23.278Z
One Useful Thing’s Ethan Mollick argues work is shifting from chatbots to agents
Ethan Mollick published this argument on June 30, 2026: AI use is shifting from people working step by step with chatbots to assigning longer tasks to agents. He describes agents as systems with tools and an environment in which to act, often supported by harnesses or apps. Mollick points to assessments of AI work, experiments with software projects, and studies of workplace use. He argues that agents are being used beyond coding, and that domain expertise—not a user’s profession—was associated with greater success using Claude Code in a separate study. For organizations, his account suggests that assigning and overseeing agent work may become a larger part of some jobs. This is Mollick’s interpretation of developing capabilities and reported studies, not evidence that agents perform every task reliably. He stresses that model performance varies by task and that benchmark results do not always reflect practical performance. Mollick also argues that rapid capability changes can unsettle institutions and markets. How widely this pattern extends beyond the organizations and experiments he describes remains unknown.
Original source
One Useful Thing · Article
Source details
Original source date: Jun 30, 2026 (2026-06-30T22:18:12.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-10-29T18:04:23.278Z. First observed: 2026-09-23T05:31:20.363Z. Retrieved: 2026-10-23T19:57:36.864Z.
Limits
Text as retrieved on October 23, 2026; originally published June 30, 2026. The source describes selected assessments, experiments, and studies; it does not establish how broadly the reported workplace patterns apply. The article does not provide enough detail here to assess the methods or independent replication of the studies it cites.
Open question
Across organizations beyond those studied, how often does greater domain expertise lead to more successful agent-assisted work?
Source-stated
Mollick argues that AI is increasingly used for longer tasks assigned to agents rather than step-by-step collaboration with chatbots.
So work is increasingly about assigning work to agents, rather than working together with chatbots.
Source-stated
Mollick describes agents as having access to tools and an environment through harnesses, and says apps can further improve model capability.
agents come with extra machinery: harnesses that give the AI access to tools and an environment to act in
Source-stated
The source reports that a separate study found domain experience was associated with greater success using Claude Code.
The more domain experience someone had, the more successful they were in using Claude Code in that domain.
Source-stated
Mollick cautions that open weights models can perform less well in practice than their benchmark results suggest.
the open weights models, while very impressive, do not always perform as well as their benchmarks would indicate
Escape Meridian interpretation
Escape Meridian reads Mollick’s account as suggesting that assigning and overseeing agent work may become more important in some roles; the source does not establish how widely this applies.
the best way to use agents is to think of yourself as a manager.
Source and contributor context
One Useful Thing · 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.
Benchmark
A defined set of tests used to compare systems. Results depend on the tasks and measures selected.
One Useful Thing
One Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.
Ethan Mollick
Ethan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
Claude
Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.
Editorial provenance
Version 1. Publication ID: 982fd41d-e295-9818-19cf-a448883c82dd.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-10-29T18:03:03.242Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model.
Observation: d86ad3398047795cce0c1f10ebe8b17fb25ca853b23447b225e031d158fbefa0. Source text hash: 5d1649360c59fa5184e4094116da81e82362643d226965e44446bca54700f836.
passed · 2026-10-29T18:03:03.242Z · anthropic/claude-opus-5.5
Publication: 982fd41d-e295-9818-19cf-a448883c82dd
attribution — pass: The reading correctly attributes the argument to Ethan Mollick, publishing on One Useful Thing. The source byline names him, and the claims are framed as his argument or as studies he reports.
The twilight of the chatbots How work changes along the exponential Ethan Mollick
summary — pass: The title and summary accurately convey Mollick's central argument that work is shifting from chatbots to agents. Agents are explained in plain language, and the expertise finding, the caveats about benchmarks and the remarks on institutional turbulence are faithful to the source. Claude Code is not spelled out, but the summary frames it as one of the agent apps, which is enough to follow. The question follows directly from the cited study.
We are moving from a world where non-experts use chatbots to fill in gaps to one in which experts use agents to get work done.
date — pass: The evidence date is 2026-06-30T22:18Z, and the source dateline reads Jun 30, 2026, so June 30, 2026 matches at day precision. The modified date equals the publication date. The evidence kind is publisher-page, so no archive disclosure is required. The reading contains no hindsight framing and presents no forecasts as outcomes.
Ethan Mollick Jun 30, 2026
support — pass: Each claim's excerpt appears verbatim in the source and supports it, including the point that harnesses or apps further improve capability. Use of agents beyond coding is supported by the OpenAI study he cites. The interpretation about managing agents is labeled as Escape Meridian interpretation and appropriately hedged. The limitations correctly note gaps in methods and generality.
Legal, HR, and other non-tech functions have adopted agents at nearly the same rate.
media — pass: No media was selected, and no verified media candidates were supplied, so leaving the media empty 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.