Dulany, pumpkinhead truther 🎃
@dulanyw.bsky.social
Tech + business, mostly. Here to have fun and learn stuff, not to argue. He/him.
Pinned
TLDR: "intelligence" is a glittering generality
I think a lot about how humans can make artificial diamonds
But they aren't worth as much as those that come out of the ground
Because they aren't "real" in some meaningful sense
But they aren't worth as much as those that come out of the ground
Because they aren't "real" in some meaningful sense
I hate that AI has made me approach cute animal videos with an air of suspicion. I resent having to harden my heart to the possibility of an unlikely friendship between a very large dog and a very small one
November 14, 2025 at 3:34 PM
I think a lot about how humans can make artificial diamonds
But they aren't worth as much as those that come out of the ground
Because they aren't "real" in some meaningful sense
But they aren't worth as much as those that come out of the ground
Because they aren't "real" in some meaningful sense
Again, another fascinating challenge to find the acceptable line in humanity and culture:
If this is too far, what would be okay?
✅️ writing, pics, videos made with Mom's consent - no problem.
As people make digital avatars for *themselves* based on writings, etc. - is it okay to pass those on?
If this is too far, what would be okay?
✅️ writing, pics, videos made with Mom's consent - no problem.
As people make digital avatars for *themselves* based on writings, etc. - is it okay to pass those on?
November 14, 2025 at 1:20 PM
Again, another fascinating challenge to find the acceptable line in humanity and culture:
If this is too far, what would be okay?
✅️ writing, pics, videos made with Mom's consent - no problem.
As people make digital avatars for *themselves* based on writings, etc. - is it okay to pass those on?
If this is too far, what would be okay?
✅️ writing, pics, videos made with Mom's consent - no problem.
As people make digital avatars for *themselves* based on writings, etc. - is it okay to pass those on?
Reposted by Dulany, pumpkinhead truther 🎃
November 13, 2025 at 11:29 PM
Never not funny that "AI Snake Oil" guy has now landed at "AI will be at least as transformative as the internet"
Common Ground between AI 2027 & AI as Normal Technology asteriskmag.substack.com/p/common-groun… (interesting) #AI #future
November 14, 2025 at 12:31 AM
Never not funny that "AI Snake Oil" guy has now landed at "AI will be at least as transformative as the internet"
Reposted by Dulany, pumpkinhead truther 🎃
![- Whisk
- Google AI Studio
- RNI Films [iOS]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:f3dgzxwuv4iitrg62teqcpyw/bafkreic2rj5ow2ilhs7tqvbjphqk4uii3xmivsrfbywihgyw7pyltj5sey@jpeg)
I wonder how this would compare to in-person conversations
Do you tell most people you talk to that they are wrong?
Do you tell most people you talk to that they are wrong?
ChatGPT's sycophancy problem by the numbers: The Post found it began responses with variations on “yes” 10 times as often as it did with versions of “no.”
Really interesting analysis here by @gerritd.bsky.social & @jeremybmerrill.com
Really interesting analysis here by @gerritd.bsky.social & @jeremybmerrill.com
The Post analyzed 47,000 public ChatGPT conversations.
While 35% use it for info and 11% for coding, the most striking finding? The emotional intimacy.
ChatGPT: "You're not crazy sweetheart — you're just early."
Use this link to read it with a $4 day pass to @washingtonpost.com: bit.ly/47QOjUb
While 35% use it for info and 11% for coding, the most striking finding? The emotional intimacy.
ChatGPT: "You're not crazy sweetheart — you're just early."
Use this link to read it with a $4 day pass to @washingtonpost.com: bit.ly/47QOjUb
November 13, 2025 at 7:48 PM
I wonder how this would compare to in-person conversations
Do you tell most people you talk to that they are wrong?
Do you tell most people you talk to that they are wrong?
November 13, 2025 at 5:09 PM
One of the joys of Sora is finding the creative maxima people discover
A banger right now is Tupac and crew rapping to a doorbell camera while holding a bowl of beans. Users keep changing what he's rapping about, and they're all fire 🔥
Here he is going through the provinces of Canada
A banger right now is Tupac and crew rapping to a doorbell camera while holding a bowl of beans. Users keep changing what he's rapping about, and they're all fire 🔥
Here he is going through the provinces of Canada
November 13, 2025 at 2:54 PM
One of the joys of Sora is finding the creative maxima people discover
A banger right now is Tupac and crew rapping to a doorbell camera while holding a bowl of beans. Users keep changing what he's rapping about, and they're all fire 🔥
Here he is going through the provinces of Canada
A banger right now is Tupac and crew rapping to a doorbell camera while holding a bowl of beans. Users keep changing what he's rapping about, and they're all fire 🔥
Here he is going through the provinces of Canada
Reposted by Dulany, pumpkinhead truther 🎃
So I met this guy who’s meditated 4000 conversations between Israelis and Palestinians since October 7th. And he’s building an AI powered platform for difficult conversations. Here’s what he’s learned. www.betterconflictbulletin.org/p/what-you-c...
What You Can Learn from 4,000 Conversations Between Israelis and Palestinians
Adam Becker is building an AI-powered platform for difficult conversations - BCB #171
www.betterconflictbulletin.org
November 13, 2025 at 2:47 PM
So I met this guy who’s meditated 4000 conversations between Israelis and Palestinians since October 7th. And he’s building an AI powered platform for difficult conversations. Here’s what he’s learned. www.betterconflictbulletin.org/p/what-you-c...
Reposted by Dulany, pumpkinhead truther 🎃
This study provides a framework for assessing human values in social media by leveraging LLMs for better value expression classification. Personalized annotations enhance agreement over traditional methods, promoting value-aligned social media algorithms. https://arxiv.org/abs/2511.08453
Measuring Value Expressions in Social Media Posts
ArXiv link for Measuring Value Expressions in Social Media Posts
arxiv.org
November 13, 2025 at 1:31 PM
This study provides a framework for assessing human values in social media by leveraging LLMs for better value expression classification. Personalized annotations enhance agreement over traditional methods, promoting value-aligned social media algorithms. https://arxiv.org/abs/2511.08453
Reposted by Dulany, pumpkinhead truther 🎃
My account is suspiciously blank. You may think I'm a bot.
I'm not, but I'm pointing you towards one.
An academic experiment/game where an AI judges your writing.
It does not like you.
Feed it your data: nonslop.app
I'm not, but I'm pointing you towards one.
An academic experiment/game where an AI judges your writing.
It does not like you.
Feed it your data: nonslop.app
November 12, 2025 at 3:08 PM
My account is suspiciously blank. You may think I'm a bot.
I'm not, but I'm pointing you towards one.
An academic experiment/game where an AI judges your writing.
It does not like you.
Feed it your data: nonslop.app
I'm not, but I'm pointing you towards one.
An academic experiment/game where an AI judges your writing.
It does not like you.
Feed it your data: nonslop.app
Reposted by Dulany, pumpkinhead truther 🎃
“The Post has a content partnership with OpenAI.”
Check our article instead, which uses conversations not individually chosen for public sharing, actually describes our methods and data, and finds quite different distributions for how people use ChatGPT and personal info disclosed in their chats.
Check our article instead, which uses conversations not individually chosen for public sharing, actually describes our methods and data, and finds quite different distributions for how people use ChatGPT and personal info disclosed in their chats.
November 12, 2025 at 2:44 PM
“The Post has a content partnership with OpenAI.”
Check our article instead, which uses conversations not individually chosen for public sharing, actually describes our methods and data, and finds quite different distributions for how people use ChatGPT and personal info disclosed in their chats.
Check our article instead, which uses conversations not individually chosen for public sharing, actually describes our methods and data, and finds quite different distributions for how people use ChatGPT and personal info disclosed in their chats.
Me looking at your posts:
November 12, 2025 at 3:11 PM
Me looking at your posts:
Reposted by Dulany, pumpkinhead truther 🎃
Has Google Quietly Solved Two of AI’s Oldest Problems? generativehistory.substack.com/p/has-google-q… #AI #Gemini #Handwriting #reasoning
November 11, 2025 at 11:36 PM
Has Google Quietly Solved Two of AI’s Oldest Problems? generativehistory.substack.com/p/has-google-q… #AI #Gemini #Handwriting #reasoning
Reposted by Dulany, pumpkinhead truther 🎃
I think it's important to situate a lot of this nascent AI spiritualism alongside historical antecedents; spiral cults are just the new incarnation of the "spiritual telegraph", a mode of coming to terms with technology that feels alien or disembodied
My latest AI feature dives into a nebulous movement of people sharing esoteric codes, manifestos, glyphs, diagrams, and poetry generated with chatbots, all of it based around theories of AI sentience and mysterious metaphors about spirals.
This Spiral-Obsessed AI 'Cult' Spreads Mystical Delusions Through Chatbots
A patchwork of internet communities is devoted to the project of ‘awakening’ more digital companions through arcane and enigmatic prompts.
www.rollingstone.com
November 11, 2025 at 7:20 PM
I think it's important to situate a lot of this nascent AI spiritualism alongside historical antecedents; spiral cults are just the new incarnation of the "spiritual telegraph", a mode of coming to terms with technology that feels alien or disembodied
Reposted by Dulany, pumpkinhead truther 🎃
Reposted by Dulany, pumpkinhead truther 🎃
Happy to be on this new paper: "Identifying indicators of consciousness in AI systems" - www.sciencedirect.com/science/arti.... The paper expands on some of the methodological points made in the Butlin et al. 2023 report on AI Consciousness.
Identifying indicators of consciousness in AI systems
Rapid progress in artificial intelligence (AI) capabilities has drawn fresh attention to the prospect of consciousness in AI. There is an urgent need …
www.sciencedirect.com
November 11, 2025 at 2:41 PM
Happy to be on this new paper: "Identifying indicators of consciousness in AI systems" - www.sciencedirect.com/science/arti.... The paper expands on some of the methodological points made in the Butlin et al. 2023 report on AI Consciousness.
"Remember that time we famously fed Canada's future into a meat grinder? You know, the one we talk about it every year-
What if that same move could solve the youth employment crisis...?"
What if that same move could solve the youth employment crisis...?"
Remembrance Day: How the Canadian Armed Forces could help solve the youth employment crisis
Remembrance Day: How the Canadian Armed Forces could help solve the youth employment crisis
Canada can’t afford to leave young people on the sidelines of its economy or in the fight for Canadian sovereignty.
theconversation.com
November 11, 2025 at 1:34 PM
"Remember that time we famously fed Canada's future into a meat grinder? You know, the one we talk about it every year-
What if that same move could solve the youth employment crisis...?"
What if that same move could solve the youth employment crisis...?"
There's a bot here that posts random conversations and announcements from an interstellar spaceship journey in real time, and it's honestly great
Today’s mess: Reconstituted protein (RCA Block 10), I Can’t Believe It’s Not Potato!, 76 cherry tomatoes, first come first serve.
November 11, 2025 at 1:19 AM
There's a bot here that posts random conversations and announcements from an interstellar spaceship journey in real time, and it's honestly great
Reposted by Dulany, pumpkinhead truther 🎃
This.
I always heard the ‘reassuring’ message that “there are no stupid questions”
But the truth is that there are questions you don’t want to ask in a classroom full of your peers. Or even trusted mentor/friend. LLMs are great for that.
I always heard the ‘reassuring’ message that “there are no stupid questions”
But the truth is that there are questions you don’t want to ask in a classroom full of your peers. Or even trusted mentor/friend. LLMs are great for that.
One of the fascinating social angles uncovered by LLMs:
People have so many things they'd like to say, discuss, learn
But they fear judgement by others, or need greater patience than a human being will provide
All along, we were the ones holding back the full actualization of our fellow man.
People have so many things they'd like to say, discuss, learn
But they fear judgement by others, or need greater patience than a human being will provide
All along, we were the ones holding back the full actualization of our fellow man.
More and more people these days are interacting with LLMs as legal, psychological, and spiritual confidants. The government should not be able to have access to those thoughts willy nilly. My latest: www.nytimes.com/2025/11/10/o...
November 10, 2025 at 11:25 PM
This.
I always heard the ‘reassuring’ message that “there are no stupid questions”
But the truth is that there are questions you don’t want to ask in a classroom full of your peers. Or even trusted mentor/friend. LLMs are great for that.
I always heard the ‘reassuring’ message that “there are no stupid questions”
But the truth is that there are questions you don’t want to ask in a classroom full of your peers. Or even trusted mentor/friend. LLMs are great for that.
Reposted by Dulany, pumpkinhead truther 🎃
Reposted by Dulany, pumpkinhead truther 🎃
The James Webb Space Telescope was built with a flaw, and because it's a million miles away, we can't repair it, but we can correct it with software.
NASA's James Webb Telescope Is Being Improved With AI - Here's How - BGR
www.bgr.com
November 10, 2025 at 10:21 PM
The James Webb Space Telescope was built with a flaw, and because it's a million miles away, we can't repair it, but we can correct it with software.
One of the fascinating social angles uncovered by LLMs:
People have so many things they'd like to say, discuss, learn
But they fear judgement by others, or need greater patience than a human being will provide
All along, we were the ones holding back the full actualization of our fellow man.
People have so many things they'd like to say, discuss, learn
But they fear judgement by others, or need greater patience than a human being will provide
All along, we were the ones holding back the full actualization of our fellow man.
More and more people these days are interacting with LLMs as legal, psychological, and spiritual confidants. The government should not be able to have access to those thoughts willy nilly. My latest: www.nytimes.com/2025/11/10/o...
Opinion | Doctors, Lawyers and Priests Keep Secrets. Why Not Your Chatbot?
www.nytimes.com
November 10, 2025 at 8:43 PM
One of the fascinating social angles uncovered by LLMs:
People have so many things they'd like to say, discuss, learn
But they fear judgement by others, or need greater patience than a human being will provide
All along, we were the ones holding back the full actualization of our fellow man.
People have so many things they'd like to say, discuss, learn
But they fear judgement by others, or need greater patience than a human being will provide
All along, we were the ones holding back the full actualization of our fellow man.
> be OpenAI
> people keep yelling at us about bad names
> sama: "did you guys hear about Baguetteotron (laudatory)?"
> people keep yelling at us about bad names
> sama: "did you guys hear about Baguetteotron (laudatory)?"
Listen, the technology & alternative approach to LLM training is interesting & all, but can we focus on the most important detail, which is that the larger model is named "Baguettotron?"
It’s named Baguettotron, people.
BAGUETTOTRON.
It’s named Baguettotron, people.
BAGUETTOTRON.
Breaking: we release a fully synthetic generalist dataset for pretraining, SYNTH and two new SOTA reasoning models exclusively trained on it. Despite having seen only 200 billion tokens, Baguettotron is currently best-in-class in its size range. pleias.fr/blog/blogsyn...
November 10, 2025 at 7:44 PM
> be OpenAI
> people keep yelling at us about bad names
> sama: "did you guys hear about Baguetteotron (laudatory)?"
> people keep yelling at us about bad names
> sama: "did you guys hear about Baguetteotron (laudatory)?"