James Grimmelmann
LightNews LightNews
banner
jtlg.bsky.social
James Grimmelmann
@jtlg.bsky.social
5.5K followers 42 following 1.5K posts

I’m a professor at Cornell Tech and Cornell Law School.

One of "a number of very informative people." -WSJ

Videos
Computer science 34%
Political science 21%
Posts Replies Media Videos

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 9h
I had not seen this. Thank you. It's an important part of the picture.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 9h
Important addition: here is Benj Edwards's account and acceptance of responsibility.

bsky.app/profile/benj...
jtlg.bsky.social James Grimmelmann @jtlg.bsky.social · 12h
Ars Technica has retracted the article about an AI bot that contained fabricated AI-generated quotes.

arstechnica.com/staff/2026/0...

This is a good first step, but it is well short of what Ars needs to do.

1/5
Editor’s Note: Retraction of article containing fabricated quotations
We are reinforcing our editorial standards following this incident.
arstechnica.com

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 9h
This is closer to what I wanted from the official Ars Technica response. I'm disappointed that this was left to/put on an indiviudal reporter and not part of the institution's own transparency and accountability.
benjedwards.com Benj Edwards @benjedwards.com · 12h
Sorry all this is my fault; and speculation has grown worse because I have been sick in bed with a high fever and unable to reliably address it (still am sick)

I was told by management not to comment until they did. Here is my statement in images below

arstechnica.com/staff/2026/0...
I have been sick with COVID all week and missed Mon and Tues due to this. On Friday, while working from bed with a fever and very little sleep, I unintentionally made a serious journalistic error in an article about Scott Shambaugh. Here’s what happened: I was incorporating information from Shambaugh’s new blog post into an existing draft from Thursday. During the process, I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline. When the tool refused to process the post due to content policy restrictions (Shambaugh’s post described harassment). I pasted the text into ChatGPT to understand why. I should have taken a sick day because in the course of that interaction, I inadvertently ended up with a paraphrased version of Shambaugh’s words rather than his actual words. Being sick and rushing to finish, I failed to verify the quotes in my outline notes against the original blog source before including them in my draft. Kyle Orland had no role in this error. He trusted me to provide accurate quotes, and I failed him. The text of the article was human-written by us, and this incident was isolated and is not representative of Ars Technica’s editorial standards. None of our articles are AI-generated, it is against company policy and we have always respected that. I sincerely apologize to Scott Shambaugh for misrepresenting his words. I take full responsibility. The irony of an AI reporter being tripped up by AI hallucination is not lost on me. I take accuracy in my work very seriously and this is a painful failure on my part. When I realized what had happened, I asked my boss to pull the piece because I was too sick to fix it on Friday. There was nothing nefarious at work, just a terrible judgement call which was no one’s fault but my own. —Benj Edwards, February 15, 2026 I have been sick with COVID all week and missed Mon and Tues due to this. On Friday, while working from bed with a fever and very little sleep, I unintentionally made a serious journalistic error in an article about Scott Shambaugh. Here’s what happened: I was incorporating information from Shambaugh’s new blog post into an existing draft from Thursday. During the process, I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline. When the tool refused to process the post due to content policy restrictions (Shambaugh’s post described harassment). I pasted the text into ChatGPT to understand why. I should have taken a sick day because in the course of that interaction, I inadvertently ended up with a paraphrased version of Shambaugh’s words rather than his actual words. Being sick and rushing to finish, I failed to verify the quotes in my outline notes against the original blog source before including them in my draft. Kyle Orland had no role in this error. He trusted me to provide accurate quotes, and I failed him. The text of the article was human-written by us, and this incident was isolated and is not representative of Ars Technica’s editorial standards. None of our articles are AI-generated, it is against company policy and we have always respected that. I sincerely apologize to Scott Shambaugh for misrepresenting his words. I take full responsibility. The irony of an AI reporter being tripped up by AI hallucination is not lost on me. I take accuracy in my work very seriously and this is a painful failure on my part. When I realized what had happened, I asked my boss to pull the piece because I was too sick to fix it on Friday. There was nothing nefarious at work, just a terrible judgement call which was no one’s fault but my own. —Benj Edwards, February 15, 2026 I have been sick with COVID all week and missed Mon and Tues due to this. On Friday, while working from bed with a fever and very little sleep, I unintentionally made a serious journalistic error in an article about Scott Shambaugh. Here’s what happened: I was incorporating information from Shambaugh’s new blog post into an existing draft from Thursday. During the process, I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline. When the tool refused to process the post due to content policy restrictions (Shambaugh’s post described harassment). I pasted the text into ChatGPT to understand why. I should have taken a sick day because in the course of that interaction, I inadvertently ended up with a paraphrased version of Shambaugh’s words rather than his actual words. Being sick and rushing to finish, I failed to verify the quotes in my outline notes against the original blog source before including them in my draft. Kyle Orland had no role in this error. He trusted me to provide accurate quotes, and I failed him. The text of the article was human-written by us, and this incident was isolated and is not representative of Ars Technica’s editorial standards. None of our articles are AI-generated, it is against company policy and we have always respected that. I sincerely apologize to Scott Shambaugh for misrepresenting his words. I take full responsibility. The irony of an AI reporter being tripped up by AI hallucination is not lost on me. I take accuracy in my work very seriously and this is a painful failure on my part. When I realized what had happened, I asked my boss to pull the piece because I was too sick to fix it on Friday. There was nothing nefarious at work, just a terrible judgement call which was no one’s fault but my own. —Benj Edwards, February 15, 2026

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 12h
I will be emailing Ars's editorial leadership to express my concerns, and considering whether to cancel my subscription.

5/5

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 12h
I have long valued Ars an in-depth, informed, and trusthworthy source on technology issues. I have written for them several times and been interviewed for Ars stories often. I have been a paid subscriber for years.

Without further action, I don't know that I can continue to rely on Ars.

4/5

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 12h
Now the bad:

* The retraction does not explain who violated the policy, how, or why.
* The retraction does not explain how they have been held accountable.
* The retraction does not explain how Ars will prevent similar editorial failures in the future.

3/5

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 12h
First, the good:

* The article has been replaced by a retraction notice.
* Ars has a policy requiring accurate quotation.
* Ars has a policy against publishing AI-generated material "unless it is clearly labeled and presented for demonstration purposes."

2/5

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 12h
Ars Technica has retracted the article about an AI bot that contained fabricated AI-generated quotes.

arstechnica.com/staff/2026/0...

This is a good first step, but it is well short of what Ars needs to do.

1/5
Editor’s Note: Retraction of article containing fabricated quotations
We are reinforcing our editorial standards following this incident.
arstechnica.com

Reposted by James Grimmelmann

benjedwards.com
Benj Edwards @benjedwards.com · 12h
Sorry all this is my fault; and speculation has grown worse because I have been sick in bed with a high fever and unable to reliably address it (still am sick)

I was told by management not to comment until they did. Here is my statement in images below

arstechnica.com/staff/2026/0...
I have been sick with COVID all week and missed Mon and Tues due to this. On Friday, while working from bed with a fever and very little sleep, I unintentionally made a serious journalistic error in an article about Scott Shambaugh. Here’s what happened: I was incorporating information from Shambaugh’s new blog post into an existing draft from Thursday. During the process, I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline. When the tool refused to process the post due to content policy restrictions (Shambaugh’s post described harassment). I pasted the text into ChatGPT to understand why. I should have taken a sick day because in the course of that interaction, I inadvertently ended up with a paraphrased version of Shambaugh’s words rather than his actual words. Being sick and rushing to finish, I failed to verify the quotes in my outline notes against the original blog source before including them in my draft. Kyle Orland had no role in this error. He trusted me to provide accurate quotes, and I failed him. The text of the article was human-written by us, and this incident was isolated and is not representative of Ars Technica’s editorial standards. None of our articles are AI-generated, it is against company policy and we have always respected that. I sincerely apologize to Scott Shambaugh for misrepresenting his words. I take full responsibility. The irony of an AI reporter being tripped up by AI hallucination is not lost on me. I take accuracy in my work very seriously and this is a painful failure on my part. When I realized what had happened, I asked my boss to pull the piece because I was too sick to fix it on Friday. There was nothing nefarious at work, just a terrible judgement call which was no one’s fault but my own. —Benj Edwards, February 15, 2026 I have been sick with COVID all week and missed Mon and Tues due to this. On Friday, while working from bed with a fever and very little sleep, I unintentionally made a serious journalistic error in an article about Scott Shambaugh. Here’s what happened: I was incorporating information from Shambaugh’s new blog post into an existing draft from Thursday. During the process, I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline. When the tool refused to process the post due to content policy restrictions (Shambaugh’s post described harassment). I pasted the text into ChatGPT to understand why. I should have taken a sick day because in the course of that interaction, I inadvertently ended up with a paraphrased version of Shambaugh’s words rather than his actual words. Being sick and rushing to finish, I failed to verify the quotes in my outline notes against the original blog source before including them in my draft. Kyle Orland had no role in this error. He trusted me to provide accurate quotes, and I failed him. The text of the article was human-written by us, and this incident was isolated and is not representative of Ars Technica’s editorial standards. None of our articles are AI-generated, it is against company policy and we have always respected that. I sincerely apologize to Scott Shambaugh for misrepresenting his words. I take full responsibility. The irony of an AI reporter being tripped up by AI hallucination is not lost on me. I take accuracy in my work very seriously and this is a painful failure on my part. When I realized what had happened, I asked my boss to pull the piece because I was too sick to fix it on Friday. There was nothing nefarious at work, just a terrible judgement call which was no one’s fault but my own. —Benj Edwards, February 15, 2026 I have been sick with COVID all week and missed Mon and Tues due to this. On Friday, while working from bed with a fever and very little sleep, I unintentionally made a serious journalistic error in an article about Scott Shambaugh. Here’s what happened: I was incorporating information from Shambaugh’s new blog post into an existing draft from Thursday. During the process, I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline. When the tool refused to process the post due to content policy restrictions (Shambaugh’s post described harassment). I pasted the text into ChatGPT to understand why. I should have taken a sick day because in the course of that interaction, I inadvertently ended up with a paraphrased version of Shambaugh’s words rather than his actual words. Being sick and rushing to finish, I failed to verify the quotes in my outline notes against the original blog source before including them in my draft. Kyle Orland had no role in this error. He trusted me to provide accurate quotes, and I failed him. The text of the article was human-written by us, and this incident was isolated and is not representative of Ars Technica’s editorial standards. None of our articles are AI-generated, it is against company policy and we have always respected that. I sincerely apologize to Scott Shambaugh for misrepresenting his words. I take full responsibility. The irony of an AI reporter being tripped up by AI hallucination is not lost on me. I take accuracy in my work very seriously and this is a painful failure on my part. When I realized what had happened, I asked my boss to pull the piece because I was too sick to fix it on Friday. There was nothing nefarious at work, just a terrible judgement call which was no one’s fault but my own. —Benj Edwards, February 15, 2026

Reposted by James Grimmelmann, Richard Pettigrew

bretdevereaux.bsky.social
"Online Rent-a-Sage" Bret Devereaux @bretdevereaux.bsky.social · 18h
The real story of the piece is mentioned in the seventh paragraph and then not returned to: one by one, the foundations that used to fund humanities research stopped, leaving just Mellon, which put an ideological rider on their funding in 2020.

They did not seize the humanities; everyone else left.

Reposted by James Grimmelmann, Ann Bartow, Greg Linden

rtushnet.bsky.social
Rebecca Tushnet @rtushnet.bsky.social · 19h
Feds now personalist and extortionate all the way down: FDA suggests “Moderna might have more luck with a second, narrower request for approval if the company showed some ‘humility’ when it made the submission.”

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
I would still like to know a lot more about the FAA's concerns on a technical level. Dispersion? Reflection? Misses? Target selection? Danger to aircraft or to pilots? Etc.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
It's apparently a LOCUST directed energy weapon. It's described as having a power of 20 kilowatts. By way of comparison, handheld laser pointers are limited by law to 5 mlliwatts, so this is the equivalent of 4 million of them.

www.twz.com/news-feature...
bluehalo.com/c-uas-autono...
This Is The LOCUST Laser That Reportedly Prompted Closing El Paso's Airspace
The U.S. Army has been acquiring multiple versions of the LOCUST laser weapon system to bolster its ability to shoot down hostile drones.
www.twz.com

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
I'm unaware of anything on the issue, and I think your observation is right. But since prospect theory obviously fails on its own terms for explaining the actual U.S. patent system, and since the use required to substantiate a trademark registration is minimal, it seems like primarily a curiosity.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
This is great on the public-facing FAA procedures, but I’d like to go deeper on the weapon and the nature of the dangers it poses.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
Has anyone written this up from a more technical perspective (even if the details are somewhat speculative)? It would help a lot for my understanding to see something that breaks it down in terms of power, range, wavelengths, etc.
jamalgreene.bsky.social Jamal Greene @jamalgreene.bsky.social · 1d
An astonishing account of recklessness and maladministration at the highest levels. (gift link) www.nytimes.com/2026/02/14/u...
Inside the Debacle That Led to the Closure of El Paso’s Airspace
www.nytimes.com

Reposted by Joanna Bryson, Ben H. Ansell, James Grimmelmann , and 1 more

jamalgreene.bsky.social
Jamal Greene @jamalgreene.bsky.social · 1d
An astonishing account of recklessness and maladministration at the highest levels. (gift link) www.nytimes.com/2026/02/14/u...
Inside the Debacle That Led to the Closure of El Paso’s Airspace
www.nytimes.com

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
Rare: "The y-axis should always go to 100%" cranks.
Rarer: "The x-axis should always go to 100%" cranks.
Rarest: "The z-axis should always go to 100%" cranks.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
Very good context for understanding how the Minneapolis protesters are smart, well-organized, and winning.
bretdevereaux.bsky.social "Online Rent-a-Sage" Bret Devereaux @bretdevereaux.bsky.social · 1d
This week on the blog: Against the State! Another military theory primer, this time focused on groups that take on the state itself, both with violent and non-violent means.

Given the enormous disparity in force, how can they win? How have they won?

acoup.blog/2026/02/13/c...
Collections: Against the State – A Primer on Terrorism, Insurgency and Protest
This week, continuing in the vein of some of our previous strategy and military theory primers, I wanted to off a basic 101-level survey of the strategic theory behind efforts, in a sense, directed…
acoup.blog

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
I’ve run to ground a number of misquotes, in which someone’s wording is gradually massaged into a more memorable form, but this is a real doozy.

Reposted by James Grimmelmann

bretdevereaux.bsky.social
"Online Rent-a-Sage" Bret Devereaux @bretdevereaux.bsky.social · 1d
This week on the blog: Against the State! Another military theory primer, this time focused on groups that take on the state itself, both with violent and non-violent means.

Given the enormous disparity in force, how can they win? How have they won?

acoup.blog/2026/02/13/c...
Collections: Against the State – A Primer on Terrorism, Insurgency and Protest
This week, continuing in the vein of some of our previous strategy and military theory primers, I wanted to off a basic 101-level survey of the strategic theory behind efforts, in a sense, directed…
acoup.blog

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 1d
Whaaaat?

Reposted by Anna O. Law, Marc Lynch, Ben H. Ansell , and 22 more

fishkin.bsky.social
Joey Fishkin @fishkin.bsky.social · 2d
This is a genuinely huge, sweeping victory today for the University of California—or rather, for us, its faculty, acting through our faculty associations, while the UC itself maintained a strict policy of deer-in-headlights silence.

The Trump admin has given up its appeal of a powerful injunction:
Trump administration drops appeal of court order blocking $1.2-billion UCLA settlement
The Trump administration dropped its appeal of a major higher education case in which a federal judge blocked its $1.2-billion settlement proposal to UCLA over alleged civil rights violations. It will...
www.latimes.com

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 2d
Yeah. I will cancel my paid subscription if there is anything less than full accountability.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 2d
It’s worse. Scrooge wouldn’t be that gratuitously cruel, and he wouldn’t spend money on a feast for himself.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 2d
Under the reasoning as reported, if my lawyer gives me a memorandum outlining my legal options, and I have Claude summarize it for me, I have waived privilege in both the outputs and the memorandum itself. The same might be true if I just uploaded the memo to non-E2EE cloud storage under normal TOS.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 2d
One of Galileo’s lesser-known writings.
Sticker reading “Siderail release”

Reposted by James Grimmelmann

kjhealy.co
Kieran Healy @kjhealy.co · 2d
The rationale that goes “You, a questing student, can only learn my special technē from me, in person, through a form of intense physical and mental submission” is what unites culty philosophers with e.g. martinet piano teachers and abusive gymnastics coaches.

jtlg.bsky.social
James Grimmelmann @jtlg.bsky.social · 2d
Tag yourself; I’m driftwood.
c. Appearances of Sea, Atmosphere, and Unusual Objects. All peculiar or extraordinary appearances of the sea, atmosphere, or heavens, preceding or following sudden changes of wind, heavy squalls or wind, or of heavy gales. All unusual appearances of the sea, tide rips, discolored water, extraordinary luminescence of the sea, strange birds, fishes, icebergs, driftwood, seaweed, etc.

Reposted by Anna O. Law, James Grimmelmann, Michael Zimmer , and 1 more

nome.bsky.social
Nome @nome.bsky.social · 3d
I have a friend who said, of Starfleet Academy, that this took them out of it - that seeing a person in a wheelchair in the 32nd century seemed unrealistic, since medicine in Star Trek "should be able to cure whatever was wrong with them."

That's not the case, for any number of possible reasons.
Lines of academy cadets, one in command red seated but at attention while in a wheelchair.