Daniel Lowd
@lowd.cc
CS Prof at the University of Oregon, studying adversarial machine learning, data poisoning, interpretable AI, probabilistic and relational models, and more. Avid unicyclist and occasional singer-songwriter. He/him
I think I could live on egg nog. It would be like that whole Soylent craze, except so much better. Not to mention more festive.
November 14, 2025 at 9:04 PM
I think I could live on egg nog. It would be like that whole Soylent craze, except so much better. Not to mention more festive.
Reposted by Daniel Lowd
She’s right. The government started putting gender on passports in 1972 out of concern that men and women were getting too hard to tell apart, what with all the hippies with long hair and feminists in pants and whatnot.
Strike gender from IDs. Everywhere. It serves no purpose. @murray.senate.gov
Strike gender from IDs. Everywhere. It serves no purpose. @murray.senate.gov
@murray.senate.gov When the government is back in session, will you introduce a bill to take “sex/gender” off passports? There is no reason for it to be there, and, indeed, it’s only been there because of the misogyny of our government in 1972?
I’d appreciate a response.
I’d appreciate a response.
November 6, 2025 at 9:18 PM
She’s right. The government started putting gender on passports in 1972 out of concern that men and women were getting too hard to tell apart, what with all the hippies with long hair and feminists in pants and whatnot.
Strike gender from IDs. Everywhere. It serves no purpose. @murray.senate.gov
Strike gender from IDs. Everywhere. It serves no purpose. @murray.senate.gov
I’ve reached the point of actively encouraging my college-aged kid to ask ChatGPT for help, because 20 minutes of searching through Petrucci’s chem textbook to find the definition and formula for partial charge of a molecular bond is NOT educational or a good use of time.
November 6, 2025 at 7:21 AM
I’ve reached the point of actively encouraging my college-aged kid to ask ChatGPT for help, because 20 minutes of searching through Petrucci’s chem textbook to find the definition and formula for partial charge of a molecular bond is NOT educational or a good use of time.
Reposted by Daniel Lowd
Wrote a short piece arguing that higher ed must help steer AI. TLDR: If we outsource this to tech, we outsource our whole business. But rejectionism is basically stalling. If we want to survive, schools themselves must proactively shape AI for education & research. [1/6, unpaywalled at 5/6] +
Opinion | AI Is the Future. Higher Ed Should Shape It.
If we want to stay at the forefront of knowledge production, we must fit technology to our needs.
www.chronicle.com
November 4, 2025 at 7:55 PM
Wrote a short piece arguing that higher ed must help steer AI. TLDR: If we outsource this to tech, we outsource our whole business. But rejectionism is basically stalling. If we want to survive, schools themselves must proactively shape AI for education & research. [1/6, unpaywalled at 5/6] +
I’m on sabbatical this year, but I haven’t been doing a startup, new projects with distant colleagues, or writing a book.
No, I’ve just been catching my breath after 22 years of academia (grad school+prof life) and 18 years as a parent.
And I’m finding that… maybe I have space to think and learn!
No, I’ve just been catching my breath after 22 years of academia (grad school+prof life) and 18 years as a parent.
And I’m finding that… maybe I have space to think and learn!
November 4, 2025 at 6:57 AM
I’m on sabbatical this year, but I haven’t been doing a startup, new projects with distant colleagues, or writing a book.
No, I’ve just been catching my breath after 22 years of academia (grad school+prof life) and 18 years as a parent.
And I’m finding that… maybe I have space to think and learn!
No, I’ve just been catching my breath after 22 years of academia (grad school+prof life) and 18 years as a parent.
And I’m finding that… maybe I have space to think and learn!
Ok, I finally got my own domain (lowd.cc), set up a semi-modern site using Hugo and claimed my BlueSky handle.
I used GPT-5 Codex for this too. It did ok at translating my content from static html to Hugo .md and .yaml, but debugging it was still a pain.
I used GPT-5 Codex for this too. It did ok at translating my content from static html to Hugo .md and .yaml, but debugging it was still a pain.
Welcome
I am a Professor of Computer Science at the University of Oregon. I work on adversarial machine learning, tractable probabilistic modeling, and statistical relational learning.
CV · Publications · Res...
lowd.cc
November 4, 2025 at 6:43 AM
Ok, I finally got my own domain (lowd.cc), set up a semi-modern site using Hugo and claimed my BlueSky handle.
I used GPT-5 Codex for this too. It did ok at translating my content from static html to Hugo .md and .yaml, but debugging it was still a pain.
I used GPT-5 Codex for this too. It did ok at translating my content from static html to Hugo .md and .yaml, but debugging it was still a pain.
I made a thing!
Since the NaNoWriMo website is gone (RIP), I vibe-coded a NaNo-style word tracker, including many stats, UI options, and a cat picture reward for each daily milestone.
Since the NaNoWriMo website is gone (RIP), I vibe-coded a NaNo-style word tracker, including many stats, UI options, and a cat picture reward for each daily milestone.
November 1, 2025 at 11:35 PM
I made a thing!
Since the NaNoWriMo website is gone (RIP), I vibe-coded a NaNo-style word tracker, including many stats, UI options, and a cat picture reward for each daily milestone.
Since the NaNoWriMo website is gone (RIP), I vibe-coded a NaNo-style word tracker, including many stats, UI options, and a cat picture reward for each daily milestone.
Reposted by Daniel Lowd
Sometimes when I put a song into Suno, what I get back is pure magic. My Flowers for Algernon-esque story, “Techno Babel” (co-written with @lowd.bsky.social) is one of those cases…
Also, new wave as a genre of music can be kind of amazing.
youtu.be/GCxLMWIMFKY?...
Also, new wave as a genre of music can be kind of amazing.
youtu.be/GCxLMWIMFKY?...
Techno Babel
YouTube video by Mary E. Lowd - Topic
youtu.be
October 27, 2025 at 4:37 AM
Sometimes when I put a song into Suno, what I get back is pure magic. My Flowers for Algernon-esque story, “Techno Babel” (co-written with @lowd.bsky.social) is one of those cases…
Also, new wave as a genre of music can be kind of amazing.
youtu.be/GCxLMWIMFKY?...
Also, new wave as a genre of music can be kind of amazing.
youtu.be/GCxLMWIMFKY?...
In case you haven’t been following closely, there’s been amazing progress in AI models for music!
My partner has been adapting her original spec fic story collections into whole albums, 1 song per story. (Kind of like how Elton John’s “Rocketman” is based on the Bradbury short.)
My partner has been adapting her original spec fic story collections into whole albums, 1 song per story. (Kind of like how Elton John’s “Rocketman” is based on the Bradbury short.)
This is a song inspired by the first story I ever sold — “Forget Me Not” — which is about a man addicted to a memory drug. I love how jazzy and catchy it turned out.
youtu.be/w3ui-XMrFcQ?...
youtu.be/w3ui-XMrFcQ?...
Forget Me Not
YouTube video by Mary E. Lowd - Topic
youtu.be
October 21, 2025 at 9:50 AM
In case you haven’t been following closely, there’s been amazing progress in AI models for music!
My partner has been adapting her original spec fic story collections into whole albums, 1 song per story. (Kind of like how Elton John’s “Rocketman” is based on the Bradbury short.)
My partner has been adapting her original spec fic story collections into whole albums, 1 song per story. (Kind of like how Elton John’s “Rocketman” is based on the Bradbury short.)
Reposted by Daniel Lowd
I can remember analogous panics about Dungeons and Dragons.
Everything that teenagers start doing in large numbers is dangerous. It always causes them to lose touch with reality and estranges them from their families. I doubt that civilization can survive if we keep letting teenagers do new stuff.
Everything that teenagers start doing in large numbers is dangerous. It always causes them to lose touch with reality and estranges them from their families. I doubt that civilization can survive if we keep letting teenagers do new stuff.
October 16, 2025 at 3:43 AM
I can remember analogous panics about Dungeons and Dragons.
Everything that teenagers start doing in large numbers is dangerous. It always causes them to lose touch with reality and estranges them from their families. I doubt that civilization can survive if we keep letting teenagers do new stuff.
Everything that teenagers start doing in large numbers is dangerous. It always causes them to lose touch with reality and estranges them from their families. I doubt that civilization can survive if we keep letting teenagers do new stuff.
Reposted by Daniel Lowd
Yeah to be clear I expect slop to end up like kitsch in that it becomes its own countercultural art movement that people defiantly brand their work as to get attention and cater to the audience that's looking for it.
October 2, 2025 at 1:53 PM
Yeah to be clear I expect slop to end up like kitsch in that it becomes its own countercultural art movement that people defiantly brand their work as to get attention and cater to the audience that's looking for it.
I worry that, in an effort to prevent chatbots from supporting self-harm, we will end up with LLMs that refuse to comment on a draft of a murder mystery novel. Worse, all LLMs will be controlled by large companies, because they’re the only entities that can afford the liability.
September 25, 2025 at 2:32 PM
I worry that, in an effort to prevent chatbots from supporting self-harm, we will end up with LLMs that refuse to comment on a draft of a murder mystery novel. Worse, all LLMs will be controlled by large companies, because they’re the only entities that can afford the liability.
Reposted by Daniel Lowd
I thought I wouldn‘t be one of those academics super into outreach talks, but I just put together something about understanding LLMs for laypeople and I get to talk about results that I don’t really focus on in any of my technical talks! It’s actually really cool. I made this lil takeaway slide
September 10, 2025 at 1:21 PM
I thought I wouldn‘t be one of those academics super into outreach talks, but I just put together something about understanding LLMs for laypeople and I get to talk about results that I don’t really focus on in any of my technical talks! It’s actually really cool. I made this lil takeaway slide
Reposted by Daniel Lowd
melt down the guns
forge the steel into beams
raise the beams to build schools
but first
melt down all the guns
forge the steel into beams
raise the beams to build schools
but first
melt down all the guns
September 10, 2025 at 8:56 PM
melt down the guns
forge the steel into beams
raise the beams to build schools
but first
melt down all the guns
forge the steel into beams
raise the beams to build schools
but first
melt down all the guns
Using a verification email for 2FA turns a 10-second login experience into a 10-minute login experience.
April 23, 2025 at 9:56 PM
Using a verification email for 2FA turns a 10-second login experience into a 10-minute login experience.
I'm thousands of miles from home, but my laptop connects to the wifi with no hiccups. I love eduroam!
April 9, 2025 at 8:17 AM
I'm thousands of miles from home, but my laptop connects to the wifi with no hiccups. I love eduroam!
I'm at @satml.org this week! Great conference, great community, and hosted in a great country this year (Denmark).
April 9, 2025 at 8:17 AM
I'm at @satml.org this week! Great conference, great community, and hosted in a great country this year (Denmark).
Reposted by Daniel Lowd
A very exciting day for open-source AI! We're releasing our biggest open source model yet -- OLMo 2 32B -- and it beats the latest GPT 3.5, GPT 4o mini, and leading open weight models like Qwen and Mistral. As usual, all data, weights, code, etc. are available.
March 13, 2025 at 6:16 PM
A very exciting day for open-source AI! We're releasing our biggest open source model yet -- OLMo 2 32B -- and it beats the latest GPT 3.5, GPT 4o mini, and leading open weight models like Qwen and Mistral. As usual, all data, weights, code, etc. are available.
UO faculty have voted to authorize a strike.
89% turnout, 92% in favor.
Why? During the pandemic, inflation soared while salaries were stagnant. UO admin refuses to acknowledge the problem and wants us to just accept the pay cut.
Thanks to my union, @uauoregon.bsky.social, for representing us!
89% turnout, 92% in favor.
Why? During the pandemic, inflation soared while salaries were stagnant. UO admin refuses to acknowledge the problem and wants us to just accept the pay cut.
Thanks to my union, @uauoregon.bsky.social, for representing us!
March 15, 2025 at 1:55 AM
UO faculty have voted to authorize a strike.
89% turnout, 92% in favor.
Why? During the pandemic, inflation soared while salaries were stagnant. UO admin refuses to acknowledge the problem and wants us to just accept the pay cut.
Thanks to my union, @uauoregon.bsky.social, for representing us!
89% turnout, 92% in favor.
Why? During the pandemic, inflation soared while salaries were stagnant. UO admin refuses to acknowledge the problem and wants us to just accept the pay cut.
Thanks to my union, @uauoregon.bsky.social, for representing us!
Reposted by Daniel Lowd
New research from Costello, Pennycook, and Rand. Can conversations with AI help reduce belief in conspiracy theories? Quite possibly. What’s the mechanism? Evidence production. Let me explain my little-t theory about how this might work.
Last year, we published a paper showing that AI models can "debunk" conspiracy theories via personalized conversations. That paper raised a major question: WHY are the human<>AI convos so effective? In a new working paper, we have some answers.
TLDR: facts
osf.io/preprints/ps...
TLDR: facts
osf.io/preprints/ps...
February 20, 2025 at 5:00 PM
New research from Costello, Pennycook, and Rand. Can conversations with AI help reduce belief in conspiracy theories? Quite possibly. What’s the mechanism? Evidence production. Let me explain my little-t theory about how this might work.
Reposted by Daniel Lowd
New blog post: Stop talking about AGI, it's lazy and misleading
I argue that we all could do better than using the slippery, undefinable term Artificial General Intelligence. We would have a much better discussion by focusing on concrete tasks and capabilities
togelius.blogspot.com/2025/01/stop...
I argue that we all could do better than using the slippery, undefinable term Artificial General Intelligence. We would have a much better discussion by focusing on concrete tasks and capabilities
togelius.blogspot.com/2025/01/stop...
January 23, 2025 at 6:56 PM
New blog post: Stop talking about AGI, it's lazy and misleading
I argue that we all could do better than using the slippery, undefinable term Artificial General Intelligence. We would have a much better discussion by focusing on concrete tasks and capabilities
togelius.blogspot.com/2025/01/stop...
I argue that we all could do better than using the slippery, undefinable term Artificial General Intelligence. We would have a much better discussion by focusing on concrete tasks and capabilities
togelius.blogspot.com/2025/01/stop...
Reposted by Daniel Lowd
Here are my notes on OpenAI's new ChatGPT Operator browser "agent", including initial thoughts on their approach to mitigating prompt injection risks simonwillison.net/2025/Jan/23/...
Introducing Operator
OpenAI released their "research preview" today of Operator, a cloud-based browser automation platform rolling out today to $200/month ChatGPT Pro subscribers. They're calling this their first "agent"....
simonwillison.net
January 23, 2025 at 7:16 PM
Here are my notes on OpenAI's new ChatGPT Operator browser "agent", including initial thoughts on their approach to mitigating prompt injection risks simonwillison.net/2025/Jan/23/...
Reposted by Daniel Lowd
Here's the talk. I hope you enjoy it
www.youtube.com/watch?v=rwfY...
www.youtube.com/watch?v=rwfY...
January 23, 2025 at 3:55 PM
Here's the talk. I hope you enjoy it
www.youtube.com/watch?v=rwfY...
www.youtube.com/watch?v=rwfY...
If the current AI gold rush is anything like the dot com boom, some of these ventures will end up like pets dot com and some will end up like Amazon dot com.
January 21, 2025 at 6:29 AM
If the current AI gold rush is anything like the dot com boom, some of these ventures will end up like pets dot com and some will end up like Amazon dot com.
Reposted by Daniel Lowd
I think that's a key question here: do you find AI tools useful or not?
People who don't find them useful think their energy use is a complete waste
People (like myself) who use them every day are much more accepting of their energy costs
People who don't find them useful think their energy use is a complete waste
People (like myself) who use them every day are much more accepting of their energy costs
January 19, 2025 at 4:52 PM
I think that's a key question here: do you find AI tools useful or not?
People who don't find them useful think their energy use is a complete waste
People (like myself) who use them every day are much more accepting of their energy costs
People who don't find them useful think their energy use is a complete waste
People (like myself) who use them every day are much more accepting of their energy costs