Matt Gallivan
@theunderstanders.com
User researcher at Anthropic • opinions mine etc • SF/NH • ☕️🎸🌁
theunderstanders.com
theunderstanders.com
Reposted by Matt Gallivan
Just dropped! The first of 9 podcast episodes all about scaling research & my book, 'Research That Scales' (@rosenfeldmedia.com, 2024). In this episode, Matt Gallivan shares his best & worst examples of scaling research. Listen here: katetowsey.substack.com/p/episode-1-... / @theunderstanders.com
Research That Scales: Episode 1–Matt Gallivan
Matt Has Built & Led Research Teams at Meta, Airbnb, and Slack
katetowsey.substack.com
April 28, 2025 at 9:16 PM
Just dropped! The first of 9 podcast episodes all about scaling research & my book, 'Research That Scales' (@rosenfeldmedia.com, 2024). In this episode, Matt Gallivan shares his best & worst examples of scaling research. Listen here: katetowsey.substack.com/p/episode-1-... / @theunderstanders.com
Any user researchers who follow me here: what's your remote research stack these days? Curious about every and any tool you love to plan, recruit, conduct, analyze and write/report!
April 25, 2025 at 5:45 PM
Any user researchers who follow me here: what's your remote research stack these days? Curious about every and any tool you love to plan, recruit, conduct, analyze and write/report!
"The common thread in all these viral conspiracy theories on TikTok is that they are fueled by distrust of institutions — from schools to the National Weather Service to the medical establishment. And that... carries over to the media: Only 16 percent of Gen Zers have strong confidence in the news."
Growing up with the internet did not produce an "internet savvy" generation. It produced a generation of credulous rubes.
How Gen Z Became the Most Gullible Generation
The almighty algorithm is fueling conspiracy theories among young people and ruining their ability to tell fact from fiction on the internet.
www.politico.com
April 25, 2025 at 5:27 PM
"The common thread in all these viral conspiracy theories on TikTok is that they are fueled by distrust of institutions — from schools to the National Weather Service to the medical establishment. And that... carries over to the media: Only 16 percent of Gen Zers have strong confidence in the news."
People are growing more comfortable and intimate with AI chatbots-- sharing secrets, growing feelings of connection, and even being romantically intimate-- and there's research suggesting the social mechanics are in place for this to get stronger and more common over time. Today's post:
📝 Research summary: Our willingness to disclose private information to AI companions
A new paper shows that many people think sharing intimate things with AI companions is worth the risk
www.theunderstanders.com
April 25, 2025 at 4:23 PM
People are growing more comfortable and intimate with AI chatbots-- sharing secrets, growing feelings of connection, and even being romantically intimate-- and there's research suggesting the social mechanics are in place for this to get stronger and more common over time. Today's post:
I’ve been on a real Paul Simon kick lately, for reasons I can’t really explain. I have tender affection for him as a representative and personification of my parents’ generation. Just listened to Malcolm Gladwell’s lovely audiobook Miracle and Wonder and really enjoyed this piece on his new tour:
In 2018, Paul Simon did a farewell tour. Seven years later, he's back on the road. I wrote about seeing his show. Also wrote about "late" career work by icons, the 1980 film ONE-TRICK PONY, the 1983 album HEARTS AND BONES, and more.
uproxx.com/indie/paul-s...
uproxx.com/indie/paul-s...
Paul Simon Is Back On The Road And Searching For God
The iconic singer-songwriter did a ’farewell’ tour in 2018, but he’s back again at 83 with new songs to play.
uproxx.com
April 23, 2025 at 1:00 AM
I’ve been on a real Paul Simon kick lately, for reasons I can’t really explain. I have tender affection for him as a representative and personification of my parents’ generation. Just listened to Malcolm Gladwell’s lovely audiobook Miracle and Wonder and really enjoyed this piece on his new tour:
🐂🐂🐂
April 21, 2025 at 10:45 PM
🐂🐂🐂
Reposted by Matt Gallivan
Have information you think the public should know about the Census Bureau, Postal Service or other parts of the federal government?
I'm a @npr.org correspondent who can be reached on Signal at hansi.01 (Please use a nonwork device)
I'm a @npr.org correspondent who can be reached on Signal at hansi.01 (Please use a nonwork device)
Brain drain at Census Bureau has employees warning about the country's statistics
Staff departures and survey cuts are roiling the federal agency in charge of producing census results, job numbers and other key statistics as Trump officials continue to slash the U.S. government.
www.npr.org
April 18, 2025 at 3:38 PM
Have information you think the public should know about the Census Bureau, Postal Service or other parts of the federal government?
I'm a @npr.org correspondent who can be reached on Signal at hansi.01 (Please use a nonwork device)
I'm a @npr.org correspondent who can be reached on Signal at hansi.01 (Please use a nonwork device)
I conducted my second piece of original primary research for my newsletter The Understanders, exploring the concerns underlying AI skepticism
open.substack.com/pub/mattgall...
open.substack.com/pub/mattgall...
🧠 The Understanders research: The concerns driving AI skepticism
We interviewed 105 people to test our earlier hypotheses about AI skepticism and, well... let's just say we proved the value of testing one's hypotheses
open.substack.com
April 6, 2025 at 5:31 PM
I conducted my second piece of original primary research for my newsletter The Understanders, exploring the concerns underlying AI skepticism
open.substack.com/pub/mattgall...
open.substack.com/pub/mattgall...
Some really good stuff in this research on workers' views of AI from Pew last month:
www.pewresearch.org/social-trend...
www.pewresearch.org/social-trend...
April 1, 2025 at 5:51 PM
Some really good stuff in this research on workers' views of AI from Pew last month:
www.pewresearch.org/social-trend...
www.pewresearch.org/social-trend...
Wrote about a new paper on the effects of perceived “creepiness” of digital assistants 🧟
📝 Research summary: What yesterday's digital assistants can teach us about tomorrow
A new paper shows that even Alexa and Siri give some of us the ick
open.substack.com
February 13, 2025 at 9:13 PM
Wrote about a new paper on the effects of perceived “creepiness” of digital assistants 🧟
Inspired by a common observation from @carnage4life.bsky.social @thebenedictevans.bsky.social and @caseynewton.bsky.social, and because I don't know how to simply sit back and enjoy a career break, I ran a study of how people think about AI model selection. www.theunderstanders.com/p/the-unders...
🧠 The Understanders research: User perceptions of model selection in ChatGPT
People started talking about the model selection UI in ChatGPT last weekend, so we interviewed 51 paying users to understand how they feel about model switching
www.theunderstanders.com
February 6, 2025 at 7:13 PM
Inspired by a common observation from @carnage4life.bsky.social @thebenedictevans.bsky.social and @caseynewton.bsky.social, and because I don't know how to simply sit back and enjoy a career break, I ran a study of how people think about AI model selection. www.theunderstanders.com/p/the-unders...
Wrote about how some new YouGov data on attitudes about AI showcase some classic signs of optimism bias:
📝 Research summary: How do Americans feel about AI? It's complicated.
New data from YouGov paints a classic picture of optimism bias in the wild
open.substack.com
February 3, 2025 at 6:08 PM
Wrote about how some new YouGov data on attitudes about AI showcase some classic signs of optimism bias:
Just stumbled upon this early discussion paper that's a nice Friday evening mindf**k: arxiv.org/pdf/2501.16513
February 1, 2025 at 12:21 AM
Just stumbled upon this early discussion paper that's a nice Friday evening mindf**k: arxiv.org/pdf/2501.16513
Wrote about a new paper on how product designers respond to AI-powered personas, and a bit about the idea of doing research with synthetic users generally:
📝 Research summary: How product designers respond to AI-powered personas
AI representations of real people seem to be showing small glimmers of promise as a research tool... with some important caveats
open.substack.com
January 30, 2025 at 2:27 AM
Wrote about a new paper on how product designers respond to AI-powered personas, and a bit about the idea of doing research with synthetic users generally:
I took a look at some research about AI voice assistants and let myself think about where we're headed for a little too long today.
📝 Research summary: User preferences and trust in AI voice assistants
Homophily, influence, and anthropomorphism, oh my!
www.theunderstanders.com
January 24, 2025 at 10:36 PM
I took a look at some research about AI voice assistants and let myself think about where we're headed for a little too long today.
Arguably the most important dimension of the AI story overall, and certainly one of the most fascinating.
I don't have anything snide or snarky to say here. I just think it's worth reading this article.
Gift link.
Gift link.
She Is in Love With ChatGPT (Gift Article)
A 28-year-old woman with a busy social life spends hours on end talking to her A.I. boyfriend for advice and consolation. And yes, they do have sex.
www.nytimes.com
January 16, 2025 at 7:09 PM
Arguably the most important dimension of the AI story overall, and certainly one of the most fascinating.
First they mock you for buying your own coffee gear and storing it in your in-laws’ basement but then… well, actually, they’ll still mock you and drink their own terrible coffee, but you won’t mind as much because you’ll have a proper cup for yourself
November 27, 2024 at 2:30 PM
First they mock you for buying your own coffee gear and storing it in your in-laws’ basement but then… well, actually, they’ll still mock you and drink their own terrible coffee, but you won’t mind as much because you’ll have a proper cup for yourself
I don’t think we ever got a Bonfire of the Vanities for the 2010s era of Big Tech, and that is a shame. We got some good first hand nonfiction and some mediocre pointed satire but nothing as sweeping and damning as the era deserved.
Tom Wolfe, man. The dude was good at his job.
Tom Wolfe, man. The dude was good at his job.
November 26, 2024 at 1:45 PM
I don’t think we ever got a Bonfire of the Vanities for the 2010s era of Big Tech, and that is a shame. We got some good first hand nonfiction and some mediocre pointed satire but nothing as sweeping and damning as the era deserved.
Tom Wolfe, man. The dude was good at his job.
Tom Wolfe, man. The dude was good at his job.
I have yet to successfully make a single functional thing with Replit's AI coding agent, yet I still find it so fun to play with. Like cosplaying an engineer.
Also, once tools like this get better, it's going to be an amazing and profound shift. Bespoke apps and software for everyone. Wild.
Also, once tools like this get better, it's going to be an amazing and profound shift. Bespoke apps and software for everyone. Wild.
November 25, 2024 at 1:06 AM
I have yet to successfully make a single functional thing with Replit's AI coding agent, yet I still find it so fun to play with. Like cosplaying an engineer.
Also, once tools like this get better, it's going to be an amazing and profound shift. Bespoke apps and software for everyone. Wild.
Also, once tools like this get better, it's going to be an amazing and profound shift. Bespoke apps and software for everyone. Wild.
Our relationship to this technology— how and when we fight it, when we cede ground to it, if some of us will ever trust it— is going to be the most interesting technology story of the next decade
This is a fascinating study, and speaks to the vast perception gap between what LLMs are capable of and what people think they are capable of.
Doctors using ChatGPT for diagnosis performed *worse* than ChatGPT alone, because they kept second-guessing the model.
www.nytimes.com/2024/11/17/h...
Doctors using ChatGPT for diagnosis performed *worse* than ChatGPT alone, because they kept second-guessing the model.
www.nytimes.com/2024/11/17/h...
A.I. Chatbots Defeated Doctors at Diagnosing Illness
A small study found ChatGPT outdid human physicians when assessing medical case histories, even when those doctors were using a chatbot.
www.nytimes.com
November 19, 2024 at 5:14 PM
Our relationship to this technology— how and when we fight it, when we cede ground to it, if some of us will ever trust it— is going to be the most interesting technology story of the next decade
This is wild. User researchers (myself definitely included) scoffed at the idea of doing research with “synthetic users” a year or so ago. Might be time to start thinking differently about the idea.
Crazy interesting paper in many ways:
1) Voice-enabled GPT-4o conducted 2 hour
interviews of 1,052 people
2) GPT-4o agents were given the transcripts & prompted to simulate the people
3) The agents were given surveys & tasks. They achieved 85% accuracy in simulating interviewees real answers!
1) Voice-enabled GPT-4o conducted 2 hour
interviews of 1,052 people
2) GPT-4o agents were given the transcripts & prompted to simulate the people
3) The agents were given surveys & tasks. They achieved 85% accuracy in simulating interviewees real answers!
November 19, 2024 at 3:22 PM
This is wild. User researchers (myself definitely included) scoffed at the idea of doing research with “synthetic users” a year or so ago. Might be time to start thinking differently about the idea.
Two San Francisco hidden gems that one should *absolutely never* spill the beans about on social media: the Cole Valley sandwich from Cheese Plus and the beach at Kirby Cove
November 15, 2024 at 10:07 PM
Two San Francisco hidden gems that one should *absolutely never* spill the beans about on social media: the Cole Valley sandwich from Cheese Plus and the beach at Kirby Cove
I smashed the Follow All button on this bad boy and if you follow me because of user research, I encourage you to do the same!
📣 I created a starter pack for the UX research community! I made so many friends and learned so much over the years on Twitter. Let's recreate that magic here. ✨ go.bsky.app/K7LqGXU
November 14, 2024 at 6:57 PM
I smashed the Follow All button on this bad boy and if you follow me because of user research, I encourage you to do the same!
Surely, more social media is the answer! So here I am.
November 11, 2024 at 7:17 PM
Surely, more social media is the answer! So here I am.