MIT Technology Review · Jan 13, 2025
MIT Technology Review’s The Download on embryos, AI agents and smaller models
Markdown · JSON · Original source
MIT Technology Review published this edition of The Download on January 13, 2025. Jessica Hamzelou describes frozen IVF embryos as caught in a moral and practical limbo: people disagree over their moral status, leaving patients, clinicians, embryologists and legislators to face questions about what to do and who should be responsible. The newsletter then reports that cofounder and chief scientist Jared Kaplan discussed how —systems that carry out tasks—might improve in 2025. The example it gives is ’s computer use, an extension that lets users direct the model to move a , click buttons and type. The edition also argues that smaller language models can match or outperform larger ones on certain tasks when trained on more focused data sets, as gains from newer high-end models trail off. These items point to different unsettled consequences: how people should govern embryos with contested moral status, what agents will be able to do, and where smaller models can perform as well as larger ones. The newsletter does not resolve those questions; its embryo section asks what the embryos mean to us and who should be responsible for them.
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
MIT Technology Review · Article
Source details
Original source date: Jan 13, 2025 (2025-01-13T13:10:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-16T01:08:23.171Z. First observed: 2026-11-02T08:44:01.032Z. Retrieved: 2026-11-02T08:45:25.372Z.
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 November 2, 2026; originally published January 13, 2025. The edition links to full stories on embryos and agents but does not include those stories’ complete arguments or evidence.
Open question
What principles should determine who is responsible for frozen IVF embryos when people disagree about their moral status?
Source-stated
Jessica Hamzelou’s section describes frozen IVF embryos as facing disagreement over their moral status and responsibility for them.
To some, they’re human cells and nothing else. To others, they’re morally equivalent to children. Many feel they exist somewhere between those two extremes.
Source-stated
MIT Technology Review says Anthropic cofounder and chief scientist Jared Kaplan discussed five ways agents would improve in 2025, and describes Claude’s computer use as an extension that can carry out on-screen tasks.
To learn what’s coming next, MIT Technology Review talked to Anthropic’s cofounder and chief scientist Jared Kaplan. Here are five ways that agents are going to get even better in 2025.
Source-stated
Will Douglas Heaven’s section says smaller models can perform as well as or better than larger ones on certain tasks when trained on more focused data sets.
For certain tasks, smaller models that are trained on more focused data sets can now perform just as well as larger ones—if not better.
Source and contributor context
MIT Technology Review · 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.
AI agent
An AI system that chooses and carries out actions toward a goal, often using software tools.
Anthropic
Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.
Claude
Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.
Cursor
Cursor makes an AI coding environment and publishes evaluations of agents working on software tasks.
Editorial provenance
Version 1. Publication ID: 48baf73d-1765-6733-1f6c-d209b89c2eed.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T01:04:04.892Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.
Observation: 9809211b80e1fe43e3e0d4137e63d4d420c91e758e5a09dbcd0c8929ae902f30. Source text hash: af13a4ba516013533003235053279eab0697ad773b41abb6fe2cf77add62cd44.
Revision and source-check history
Version 1 · 2026-11-16T01:08:23.171Z
MIT Technology Review’s The Download on embryos, AI agents and smaller models
MIT Technology Review published this edition of The Download on January 13, 2025. Jessica Hamzelou describes frozen IVF embryos as caught in a moral and practical limbo: people disagree over their moral status, leaving patients, clinicians, embryologists and legislators to face questions about what to do and who should be responsible. The newsletter then reports that cofounder and chief scientist Jared Kaplan discussed how —systems that carry out tasks—might improve in 2025. The example it gives is ’s computer use, an extension that lets users direct the model to move a , click buttons and type. The edition also argues that smaller language models can match or outperform larger ones on certain tasks when trained on more focused data sets, as gains from newer high-end models trail off. These items point to different unsettled consequences: how people should govern embryos with contested moral status, what agents will be able to do, and where smaller models can perform as well as larger ones. The newsletter does not resolve those questions; its embryo section asks what the embryos mean to us and who should be responsible for them.
Original source
MIT Technology Review · Article
Source details
Original source date: Jan 13, 2025 (2025-01-13T13:10:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-16T01:08:23.171Z. First observed: 2026-11-02T08:44:01.032Z. Retrieved: 2026-11-02T08:45:25.372Z.
Limits
Text as retrieved on November 2, 2026; originally published January 13, 2025. The edition links to full stories on embryos and agents but does not include those stories’ complete arguments or evidence.
Open question
What principles should determine who is responsible for frozen IVF embryos when people disagree about their moral status?
Source-stated
Jessica Hamzelou’s section describes frozen IVF embryos as facing disagreement over their moral status and responsibility for them.
To some, they’re human cells and nothing else. To others, they’re morally equivalent to children. Many feel they exist somewhere between those two extremes.
Source-stated
MIT Technology Review says Anthropic cofounder and chief scientist Jared Kaplan discussed five ways agents would improve in 2025, and describes Claude’s computer use as an extension that can carry out on-screen tasks.
To learn what’s coming next, MIT Technology Review talked to Anthropic’s cofounder and chief scientist Jared Kaplan. Here are five ways that agents are going to get even better in 2025.
Source-stated
Will Douglas Heaven’s section says smaller models can perform as well as or better than larger ones on certain tasks when trained on more focused data sets.
For certain tasks, smaller models that are trained on more focused data sets can now perform just as well as larger ones—if not better.
Source and contributor context
MIT Technology Review · 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.
AI agent
An AI system that chooses and carries out actions toward a goal, often using software tools.
Anthropic
Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.
Claude
Claude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.
Cursor
Cursor makes an AI coding environment and publishes evaluations of agents working on software tasks.
Editorial provenance
Version 1. Publication ID: 48baf73d-1765-6733-1f6c-d209b89c2eed.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-16T01:04:04.892Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_admitted_by_model, second_opinion_categorical.
Observation: 9809211b80e1fe43e3e0d4137e63d4d420c91e758e5a09dbcd0c8929ae902f30. Source text hash: af13a4ba516013533003235053279eab0697ad773b41abb6fe2cf77add62cd44.
passed · 2026-11-16T01:04:04.892Z · anthropic/claude-opus-5.5
Publication: 48baf73d-1765-6733-1f6c-d209b89c2eed
attribution — pass: This is a single newsletter issue that comments on several stories in its own words, so it counts as one piece. The reading correctly credits MIT Technology Review's The Download. It attributes the embryo section to Jessica Hamzelou and the small-models section to Will Douglas Heaven. It presents Kaplan's views as reported by the publication, not as established fact.
This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology.
summary — pass: A non-specialist can follow the reading. It explains AI agents as systems that carry out tasks, and the page's definitions list also covers the term. It describes computer use as moving a cursor, clicking buttons and typing. It sets out the moral-status disagreement over embryos and the small-model claim. The title and question match the content. The question follows directly from the embryo section's own questions rather than leading the reader.
While these embryos persist in suspended animation, patients, clinicians, embryologists, and legislators must grapple with the essential question of what we should do with them.
date — pass: The reading's January 13, 2025 matches the UTC day of originalPublished, the instant 2025-01-13T13:10:00Z. Its limitations note the later retrieval date. The Deep Dive links describe apparently later items, such as OpenAI's math findings, and are likely later page additions. The reading neither uses nor relies on them. Kaplan's 2025 expectations are reported as forecasts ('might improve'), not outcomes. The page is a publisher page, so no archive disclosure is needed.
support — pass: Each claim matches the source text. The embryo moral-status disagreement and the responsibility questions are stated in the source. The Kaplan interview is described as covering five ways agents will get better in 2025. The small-model claim follows the source wording, including the 'certain tasks' condition and the trailing-off marginal gains. The reading implies no benefits or costs beyond what the source states.
But as the marginal gains for new high-end models trail off, researchers are figuring out how to do more with less.
media — pass: No media was selected. No verified media candidates were supplied, so 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.
Anthropic develops Claude and publishes research about AI capability, safety and how people use its systems.
Explore AnthropicAn AI system that chooses and carries out actions toward a goal, often using software tools.
Definition referenceClaude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.
Explore ClaudeCursor makes an AI coding environment and publishes evaluations of agents working on software tasks.
Explore CursorAnthropic develops Claude and publishes research about AI capability, safety and how people use its systems.
Explore AnthropicAn AI system that chooses and carries out actions toward a goal, often using software tools.
Definition referenceClaude is Anthropic’s family of AI models and assistants for tasks such as writing, analysis and coding.
Explore ClaudeCursor makes an AI coding environment and publishes evaluations of agents working on software tasks.
Explore Cursor