· Nov 9, 2026
Ethan Mollick calls Fable 5.2 “almost” AGI, with a jagged edge
Markdown · JSON · Original source
argues that Fable 5.2 is “almost” : across filmmaking, hardware, business, literature, research and 3D world-building, it can produce work he says would once have seemed like expert-level performance. But the gaps he describes are less about obvious errors than about judgment. The model made an accurate but joyless course, produced a sprawling paper, repeated creative ideas and built a rule-bound process that slowed its work. Mollick says critics and fact-checkers caught errors, but did not notice when work was too long, repetitive or no fun; he had to identify those problems and steer revisions. He also stopped the business experiment before learning whether the plan could work in practice, and says he would not trust Fable to run a business alone. His account therefore distinguishes impressive performance on digital tasks from reliably judging whether its own work is good enough. Mollick notes that vague instructions and differences in taste may also explain some failures, and says he could not always tell whether a problem came from the AI or how he managed it. He describes this phase as exciting because it let him direct projects he would not otherwise have attempted, while wondering how long that will last.
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: Nov 9, 2026 (2026-11-09T19:04:26+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-10T18:48:06.055Z. First observed: 2026-11-09T19:12:25.558Z. Retrieved: 2026-11-09T19:15:56.697Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
This is Mollick’s account of experiments with early access to Fable 5.2; he says he did not fully check the research results himself. The business experiment was stopped before Fable’s ability to run the business could be tested. Mollick says he could not always distinguish model weakness from the effects of his instructions or differences in taste.
Open question
When a project goes wrong, how can Mollick tell whether the cause is a weakness in the AI, his instructions or a difference in taste?
Source-stated
Ethan Mollick defines AGI for this post as an AI as good as most human experts at most intellectual tasks done with a computer, and says Fable 5.2 is “almost” AGI by that standard.
For this post, I’ll go with: “an AI as good as most human experts at most intellectual tasks you can do with a computer.”
Source-stated
Mollick describes Fable producing impressive work across several digital domains, including filmmaking, hardware, business, literary scholarship, social science and 3D world-building.
one model worked as a filmmaker, a hardware hacker, an entrepreneur, a literary scholar, a social scientist and a builder of 3D worlds
Source-stated
In Mollick’s account, Fable made work that was accurate but joyless, produced an unwieldy research paper, repeated ideas and built rules that slowed a creative project.
it made a course that was accurate and joyless, it wrote a 182-page paper, it kept coming back to its favorite ideas, and it buried a creative project under rules and approvals.
Source-stated
Mollick says the model did not notice important quality problems itself, despite using critics, fact-checkers or judges on nearly every project.
Fable put critics, fact-checkers or judges on nearly every project, but those checks found errors and added fixes. None of them asked whether the work was too long, too repetitive, or no fun.
Source-stated
Mollick stopped the business experiment before testing whether Fable could actually run the proposed business, and says he would not trust it to run one alone.
I stopped the experiment before finding out whether it could actually run the business
Source-stated
Mollick says some problems may have come from his vague instructions or differences in taste, and that he could not always tell the cause.
I could not tell. That is not a question I ever had to ask about software.
Source-stated
Mollick says directing and steering the projects made him feel empowered, while he remained uncertain how long this phase would last.
I felt empowered by Fable, able to command things to be made that I never would have even considered doing before.
Source and contributor context
Ethan Mollick · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Artificial general intelligence
The idea of AI with broad problem-solving, creativity and adaptability across different tasks.
Ethan Mollick
Ethan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
One Useful Thing
One Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.
Editorial provenance
Version 1. Publication ID: 5ef0e93c-e9d4-41aa-8ffd-5ac9683b5382.
Prepared by server-excerpt-repair/stream-excerpt-repair-v1 (original draft: gpt-6-luna); grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T18:48:06.055Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: e18602e96bcda92821bf24b72dc5b1e3d0963b22b46a0bb915464b24134dc309. Source text hash: ade0e511c2b1acb40ab3a9e87a623e215c608788034b06706c68b71a6c241d55.
Revision and source-check history
Version 1 · 2026-11-10T18:48:06.055Z
Ethan Mollick calls Fable 5.2 “almost” AGI, with a jagged edge
argues that Fable 5.2 is “almost” : across filmmaking, hardware, business, literature, research and 3D world-building, it can produce work he says would once have seemed like expert-level performance. But the gaps he describes are less about obvious errors than about judgment. The model made an accurate but joyless course, produced a sprawling paper, repeated creative ideas and built a rule-bound process that slowed its work. Mollick says critics and fact-checkers caught errors, but did not notice when work was too long, repetitive or no fun; he had to identify those problems and steer revisions. He also stopped the business experiment before learning whether the plan could work in practice, and says he would not trust Fable to run a business alone. His account therefore distinguishes impressive performance on digital tasks from reliably judging whether its own work is good enough. Mollick notes that vague instructions and differences in taste may also explain some failures, and says he could not always tell whether a problem came from the AI or how he managed it. He describes this phase as exciting because it let him direct projects he would not otherwise have attempted, while wondering how long that will last.
Original source
· Article
Source details
Original source date: Nov 9, 2026 (2026-11-09T19:04:26+00:00).
Date basis: Date stated by the publisher.
Added to Escape Meridian: 2026-11-10T18:48:06.055Z. First observed: 2026-11-09T19:12:25.558Z. Retrieved: 2026-11-09T19:15:56.697Z.
Limits
This is Mollick’s account of experiments with early access to Fable 5.2; he says he did not fully check the research results himself. The business experiment was stopped before Fable’s ability to run the business could be tested. Mollick says he could not always distinguish model weakness from the effects of his instructions or differences in taste.
Open question
When a project goes wrong, how can Mollick tell whether the cause is a weakness in the AI, his instructions or a difference in taste?
Source-stated
Ethan Mollick defines AGI for this post as an AI as good as most human experts at most intellectual tasks done with a computer, and says Fable 5.2 is “almost” AGI by that standard.
For this post, I’ll go with: “an AI as good as most human experts at most intellectual tasks you can do with a computer.”
Source-stated
Mollick describes Fable producing impressive work across several digital domains, including filmmaking, hardware, business, literary scholarship, social science and 3D world-building.
one model worked as a filmmaker, a hardware hacker, an entrepreneur, a literary scholar, a social scientist and a builder of 3D worlds
Source-stated
In Mollick’s account, Fable made work that was accurate but joyless, produced an unwieldy research paper, repeated ideas and built rules that slowed a creative project.
it made a course that was accurate and joyless, it wrote a 182-page paper, it kept coming back to its favorite ideas, and it buried a creative project under rules and approvals.
Source-stated
Mollick says the model did not notice important quality problems itself, despite using critics, fact-checkers or judges on nearly every project.
Fable put critics, fact-checkers or judges on nearly every project, but those checks found errors and added fixes. None of them asked whether the work was too long, too repetitive, or no fun.
Source-stated
Mollick stopped the business experiment before testing whether Fable could actually run the proposed business, and says he would not trust it to run one alone.
I stopped the experiment before finding out whether it could actually run the business
Source-stated
Mollick says some problems may have come from his vague instructions or differences in taste, and that he could not always tell the cause.
I could not tell. That is not a question I ever had to ask about software.
Source-stated
Mollick says directing and steering the projects made him feel empowered, while he remained uncertain how long this phase would last.
I felt empowered by Fable, able to command things to be made that I never would have even considered doing before.
Source and contributor context
Ethan Mollick · Article author
These are credits stated by the source; they do not establish a profile match or authorship of every passage.
Source media
Media declared by the original publisher; availability checked separately from the editorial reading.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Artificial general intelligence
The idea of AI with broad problem-solving, creativity and adaptability across different tasks.
Ethan Mollick
Ethan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
One Useful Thing
One Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.
Editorial provenance
Version 1. Publication ID: 5ef0e93c-e9d4-41aa-8ffd-5ac9683b5382.
Prepared by server-excerpt-repair/stream-excerpt-repair-v1 (original draft: gpt-6-luna); grounding review by gpt-6-luna.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-10T18:48:06.055Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: useful_perspective.
Observation: e18602e96bcda92821bf24b72dc5b1e3d0963b22b46a0bb915464b24134dc309. Source text hash: ade0e511c2b1acb40ab3a9e87a623e215c608788034b06706c68b71a6c241d55.
uncertain · 2026-11-10T18:48:06.055Z · anthropic/claude-opus-5.5
Publication: 5ef0e93c-e9d4-41aa-8ffd-5ac9683b5382
attribution — pass: The draft attributes the argument, experiments and judgments to Ethan Mollick, the author of the One Useful Thing post. His evaluations, such as 'almost' AGI and his claim that the model did not notice its own problems, stay framed as his views rather than established fact.
Fable 5.2 and Jagged AGI General, often expert, and sometimes jagged Ethan Mollick
summary — uncertain: The summary tracks Mollick's argument and keeps his uncertainty, but a reader without background lacks key context. It never says Fable 5.2 is Anthropic's Claude model, given to Mollick before public launch, or what AGI means. It also omits Mollick's deliberately limited definition, which excludes physical-world work, interaction with people and most actual jobs. That scope sets the bar for 'almost' AGI and appears only in the claims. Smallest repair: add one sentence naming Fable as Anthropic's new Claude model and stating his limited AGI definition and its exclusions.
I have had the chance to use the new Claude Fable 5.2 quite a bit before its public launch
date — pass: evidence.originalPublished is 2026-11-09, taken from the publisher page. It matches the byline date in the source text, and the feed timestamp is the same.
Ethan Mollick Nov 09, 2026
support — pass: Each claim excerpt is verbatim and supports its claim. The source supports the summary's statements: the joyless course, the 182-page paper, repeated themes, the rule book, checks that missed length and fun, the stopped business test, his refusal to trust it alone, the unchecked research results and his uncertainty about causes. The question follows from his stated difficulty telling team problems from management problems. No unsupported benefits or costs are implied.
None of them asked whether the work was too long, too repetitive, or no fun. That all fell on me
media — pass: The selected media is the verified link to the original One Useful Thing article. It is the most direct way to read the source.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Ethan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
Explore Ethan MollickEthan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
Explore Ethan MollickThe idea of AI with broad problem-solving, creativity and adaptability across different tasks.
Definition referenceOne Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.
Explore One Useful ThingEthan Mollick is a Wharton researcher and author exploring how AI changes work, learning and decision-making.
Explore Ethan MollickThe idea of AI with broad problem-solving, creativity and adaptability across different tasks.
Definition referenceOne Useful Thing is Ethan Mollick’s publication about AI in work, education and everyday practice.
Explore One Useful Thing