MIT Technology Review · Sep 26, 2024
MIT Technology Review reports on CRISPR patents and tiny AI models
Markdown · JSON · Original source
MIT Technology Review’s September 26, 2024, edition of The Download reports that lawyers for Emmanuelle Charpentier and Jennifer Doudna are seeking to withdraw two European CRISPR patents after an appeals board ruled their earliest filing did not explain CRISPR well enough for other scientists to use. The newsletter says the ruling could affect who collects licensing fees, but gives no details of the dispute’s outcome. The edition also reports that is releasing Molmo, a family of open-source language models. Ai2 says its largest model outperforms GPT-4o on tests involving images, charts, and documents; it credits its smaller model’s performance to more efficient data collection and training. The newsletter suggests that using less, higher-quality data could lower computing costs and help govern AI training data. These are reported claims and implications, not evidence of real-world performance. The newsletter gives no test results or comparison methods. It also reports Humane Intelligence’s competition for a computer vision model to track hateful image-based propaganda. Whether Molmo’s test performance translates to practical use, and what licensing consequences follow from the patent decision, remain 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
MIT Technology Review · Article
Source details
Original source date: Sep 26, 2024 (2024-09-26T12:10:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-17T01:08:23.295Z. First observed: 2026-10-30T08:00:43.325Z. Retrieved: 2026-10-30T08:01:59.408Z.
Referenced by
Recorded citations in public readings. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on October 30, 2026; originally published September 26, 2024. The newsletter links to full stories but provides only brief summaries here; it gives no detailed patent record or account of the appeals board’s reasoning. The newsletter provides no test results or comparison methodology for ’s reported Molmo performance, so the claims do not establish real-world performance.
Open question
What evidence or comparison methods underlie ’s reported Molmo test results?
Source-stated
MIT Technology Review reports that lawyers for Emmanuelle Charpentier and Jennifer Doudna are asking to withdraw two of their European CRISPR patents after an appeals board ruling about their earliest filing.
the team that earned the Nobel Prize in chemistry for developing CRISPR is asking to cancel two of their own seminal patents
Source-stated
The newsletter says the patent decision could affect who receives licensing fees for CRISPR.
The decision could have major ramifications regarding who gets to collect the lucrative licensing fees on using the technology.
Source-stated
Ai2 says its largest Molmo model outperforms GPT-4o on tests involving images, charts, and documents.
its biggest Molmo model outperforms OpenAI’s GPT-4o in tests that measure things like understanding images, charts, and documents.
Source-stated
The newsletter suggests that training on less, higher-quality data could lower computing costs and help govern AI development data.
training models on less, but higher-quality, data can lower computing costs.
Source-stated
Humane Intelligence is launching a competition challenging developers to create a computer vision model to track hateful image-based propaganda online.
is launching a competition that challenges developers to create a computer vision model that can track hateful image-based propaganda online.
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.
Multimodal model
A model that works with more than one kind of input or output, such as text, images and audio.
Ai2
Ai2 develops AI research tools and models, including the openly documented Olmo family.
Editorial provenance
Version 1. Publication ID: a70c4c05-d6c5-d0f1-ff95-a6134f83cd58.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-17T01:03:44.409Z. 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: 9acaf7a2281ca28b6cac5ec98cb8ebf0a5e8cf3f4c80d0331aa3b4d02697b258. Source text hash: ccc892c71e0581e94f2792c9167f6c4616be5753ed59f26f79450bed687e2ef1.
Revision and source-check history
Version 1 · 2026-11-17T01:08:23.295Z
MIT Technology Review reports on CRISPR patents and tiny AI models
MIT Technology Review’s September 26, 2024, edition of The Download reports that lawyers for Emmanuelle Charpentier and Jennifer Doudna are seeking to withdraw two European CRISPR patents after an appeals board ruled their earliest filing did not explain CRISPR well enough for other scientists to use. The newsletter says the ruling could affect who collects licensing fees, but gives no details of the dispute’s outcome. The edition also reports that is releasing Molmo, a family of open-source language models. Ai2 says its largest model outperforms GPT-4o on tests involving images, charts, and documents; it credits its smaller model’s performance to more efficient data collection and training. The newsletter suggests that using less, higher-quality data could lower computing costs and help govern AI training data. These are reported claims and implications, not evidence of real-world performance. The newsletter gives no test results or comparison methods. It also reports Humane Intelligence’s competition for a computer vision model to track hateful image-based propaganda. Whether Molmo’s test performance translates to practical use, and what licensing consequences follow from the patent decision, remain unknown.
Original source
MIT Technology Review · Article
Source details
Original source date: Sep 26, 2024 (2024-09-26T12:10:00Z).
Date basis: Date from the publisher’s own records, through its API.
Added to Escape Meridian: 2026-11-17T01:08:23.295Z. First observed: 2026-10-30T08:00:43.325Z. Retrieved: 2026-10-30T08:01:59.408Z.
Limits
Text as retrieved on October 30, 2026; originally published September 26, 2024. The newsletter links to full stories but provides only brief summaries here; it gives no detailed patent record or account of the appeals board’s reasoning. The newsletter provides no test results or comparison methodology for ’s reported Molmo performance, so the claims do not establish real-world performance.
Open question
What evidence or comparison methods underlie ’s reported Molmo test results?
Source-stated
MIT Technology Review reports that lawyers for Emmanuelle Charpentier and Jennifer Doudna are asking to withdraw two of their European CRISPR patents after an appeals board ruling about their earliest filing.
the team that earned the Nobel Prize in chemistry for developing CRISPR is asking to cancel two of their own seminal patents
Source-stated
The newsletter says the patent decision could affect who receives licensing fees for CRISPR.
The decision could have major ramifications regarding who gets to collect the lucrative licensing fees on using the technology.
Source-stated
Ai2 says its largest Molmo model outperforms GPT-4o on tests involving images, charts, and documents.
its biggest Molmo model outperforms OpenAI’s GPT-4o in tests that measure things like understanding images, charts, and documents.
Source-stated
The newsletter suggests that training on less, higher-quality data could lower computing costs and help govern AI development data.
training models on less, but higher-quality, data can lower computing costs.
Source-stated
Humane Intelligence is launching a competition challenging developers to create a computer vision model to track hateful image-based propaganda online.
is launching a competition that challenges developers to create a computer vision model that can track hateful image-based propaganda online.
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.
Multimodal model
A model that works with more than one kind of input or output, such as text, images and audio.
Ai2
Ai2 develops AI research tools and models, including the openly documented Olmo family.
Editorial provenance
Version 1. Publication ID: a70c4c05-d6c5-d0f1-ff95-a6134f83cd58.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-17T01:03:44.409Z. 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: 9acaf7a2281ca28b6cac5ec98cb8ebf0a5e8cf3f4c80d0331aa3b4d02697b258. Source text hash: ccc892c71e0581e94f2792c9167f6c4616be5753ed59f26f79450bed687e2ef1.
uncertain · 2026-11-17T01:03:44.409Z · anthropic/claude-opus-5.5
Publication: a70c4c05-d6c5-d0f1-ff95-a6134f83cd58
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 attributes the content to MIT Technology Review's The Download. It credits Molmo's performance claims to Ai2 and the patent withdrawal to lawyers for Charpentier and Doudna, matching the source.
This is today's edition of The Download, our weekday newsletter
summary — uncertain: The central term CRISPR is never explained, and the definitions list does not cover it. The source supplies a plain gloss, a super-tool for modifying DNA. Smallest repair: add that clause after the first mention of CRISPR. GPT-4o also appears without noting it is an OpenAI model, which the source states. Otherwise the title, the question and the account of both items are faithful.
In the decade-long fight to control CRISPR, the super-tool for modifying DNA
date — pass: evidence.originalPublished is 2024-09-26T12:10Z, and the reading states September 26, 2024, which matches the UTC day. The reading does not use the later Deep Dive items, such as the OpenAI math findings. It uses no hindsight framing and presents no forecast as an outcome. This is a publisher page, so no archive disclosure is required.
support — pass: Every claim is source-stated and correctly attributed. The reading keeps Ai2's benchmark claims as claims. Its point about less but higher-quality data and data governance follows the newsletter's own suggestion. Its stated limits, including that no test methodology is given and real-world performance is unknown, accurately reflect what the evidence lacks.
These techniques could prove really useful if we want to meaningfully govern the data that we use for AI development
media — pass: No media candidates were supplied, and no media is selected, which is consistent.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
Ai2 develops AI research tools and models, including the openly documented Olmo family.
Explore Ai2A model that works with more than one kind of input or output, such as text, images and audio.
Definition referenceAi2 develops AI research tools and models, including the openly documented Olmo family.
Explore Ai2Ai2 develops AI research tools and models, including the openly documented Olmo family.
Explore Ai2Ai2 develops AI research tools and models, including the openly documented Olmo family.
Explore Ai2A model that works with more than one kind of input or output, such as text, images and audio.
Definition referenceAi2 develops AI research tools and models, including the openly documented Olmo family.
Explore Ai2Ai2 develops AI research tools and models, including the openly documented Olmo family.
Explore Ai2