Melanie Mitchell
@melaniemitchell.bsky.social
Professor, Santa Fe Institute. Research on AI, cognitive science, and complex systems.
Website: https://melaniemitchell.me
Substack: https://aiguide.substack.com/
Website: https://melaniemitchell.me
Substack: https://aiguide.substack.com/
Straight from the horse's mouth:
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
November 25, 2025 at 8:58 PM
Straight from the horse's mouth:
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
“These models somehow just generalize dramatically worse than people. It's a very fundamental thing.”
I'm looking forward to the upcoming "Journal of Adversarial Poetry"
November 20, 2025 at 9:52 PM
I'm looking forward to the upcoming "Journal of Adversarial Poetry"
Reposted by Melanie Mitchell
If you use GMail, AI (Gemini) was turned on yesterday by default and now scans all of your content for machine learning. To turn off, go to Settings>General and scroll down. Uncheck the box for "Smart features."
There's other "Smart" add-ons as well, but that's the one that reads your content.
There's other "Smart" add-ons as well, but that's the one that reads your content.
November 20, 2025 at 5:32 PM
If you use GMail, AI (Gemini) was turned on yesterday by default and now scans all of your content for machine learning. To turn off, go to Settings>General and scroll down. Uncheck the box for "Smart features."
There's other "Smart" add-ons as well, but that's the one that reads your content.
There's other "Smart" add-ons as well, but that's the one that reads your content.
It's sad that AI conference reviewers use "incremental" as reason to reject a paper -- e.g., "the contribution of this paper is incremental; reject". Where do they think most progress in science comes from, and what eventually fuels big discoveries?
November 19, 2025 at 9:54 PM
It's sad that AI conference reviewers use "incremental" as reason to reject a paper -- e.g., "the contribution of this paper is incremental; reject". Where do they think most progress in science comes from, and what eventually fuels big discoveries?
Good article bringing a reality-check to the claim of an automonous AI cyber-attack.
arstechnica.com/security/202...
arstechnica.com/security/202...
Researchers question Anthropic claim that AI-assisted attack was 90% autonomous
The results of AI-assisted hacking aren’t as impressive as many might have us believe.
arstechnica.com
November 15, 2025 at 8:40 PM
Good article bringing a reality-check to the claim of an automonous AI cyber-attack.
arstechnica.com/security/202...
arstechnica.com/security/202...
November 11, 2025 at 10:51 PM
I just called my senators, leaving messages asking them not to vote for any deal to end the shutdown without a deal to extend ACA subsidies. It is shocking to me that they are about to cave without this.
Following up on this. This remains a ‘cave is minutes or hours away’ situation. If you want to impact how this plays out you need to contact senators literally now. I want to add additional points of context. What I’m relaying is what I’ve picked up from highly reliable sources.
Sounds like the squishes in the senate caucus are ready to pull the plug with no ACA changes. This is real. If you want to register your opinion you shld call yr senator in the next hour. They not only want to reopen w/nothing. They want cover from their colleagues who still want to hold out.
November 9, 2025 at 11:13 PM
I just called my senators, leaving messages asking them not to vote for any deal to end the shutdown without a deal to extend ACA subsidies. It is shocking to me that they are about to cave without this.
Beautiful remembrance of Brian Cantwell Smith, from Faculty of Information, University of Toronto: ischool.utoronto.ca/news/obituar...
Obituary: Brian Cantwell Smith (1950 to 2025) - Faculty of Information
Former dean wove philosophy and computation into a lifelong inquiry into the human dimensions of intelligence, judgment, and meaning Born into a prominent Canadian family known for both its intellectu...
ischool.utoronto.ca
November 4, 2025 at 12:52 AM
Beautiful remembrance of Brian Cantwell Smith, from Faculty of Information, University of Toronto: ischool.utoronto.ca/news/obituar...
Or, much more likely, they are role-playing, based on their training data.
www.theguardian.com/technology/2...
www.theguardian.com/technology/2...
AI models may be developing their own ‘survival drive’, researchers say
Like 2001: A Space Odyssey’s HAL 9000, some AIs seem to resist being turned off and will even sabotage shutdown
www.theguardian.com
October 25, 2025 at 5:14 PM
Or, much more likely, they are role-playing, based on their training data.
www.theguardian.com/technology/2...
www.theguardian.com/technology/2...
I'm so thrilled and honored to be selected for this award, and really grateful to @scicommexcellence.nationalacademies.org for their support of science communication.
www.nationalacademies.org/awards/excel...
#scicomm
www.nationalacademies.org/awards/excel...
#scicomm
www.nationalacademies.org
October 23, 2025 at 3:36 PM
I'm so thrilled and honored to be selected for this award, and really grateful to @scicommexcellence.nationalacademies.org for their support of science communication.
www.nationalacademies.org/awards/excel...
#scicomm
www.nationalacademies.org/awards/excel...
#scicomm
Excellent and insightful critique of the recent "Definition of AGI" paper, from @philipcball.bsky.social
philipball86.substack.com/p/can-we-fin...
philipball86.substack.com/p/can-we-fin...
Can we find a good definition of AGI?
I don't think it is this one.
philipball86.substack.com
October 21, 2025 at 3:03 PM
Excellent and insightful critique of the recent "Definition of AGI" paper, from @philipcball.bsky.social
philipball86.substack.com/p/can-we-fin...
philipball86.substack.com/p/can-we-fin...
Reposted by Melanie Mitchell
Tomorrow: the final event in SFI’s 2025 Community Lecture Series!
Join us for Alison Gopnik's lecture, “Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?” — 7:30 pm at The Lensic.
Free tickets: lensic.org/events
Livestream: youtube.com/@SFIScience
Join us for Alison Gopnik's lecture, “Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?” — 7:30 pm at The Lensic.
Free tickets: lensic.org/events
Livestream: youtube.com/@SFIScience
Transmission Versus Truth: What Will It Take to Make an Al as Smart as a 4-Year-Old?
There is no such thing as general intelligence — artificial or natural — argues Alison Gopnik. Instead, there are multiple intelligences, each with its own trade-offs. Three different types of cogniti...
www.santafe.edu
October 20, 2025 at 5:50 PM
Tomorrow: the final event in SFI’s 2025 Community Lecture Series!
Join us for Alison Gopnik's lecture, “Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?” — 7:30 pm at The Lensic.
Free tickets: lensic.org/events
Livestream: youtube.com/@SFIScience
Join us for Alison Gopnik's lecture, “Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?” — 7:30 pm at The Lensic.
Free tickets: lensic.org/events
Livestream: youtube.com/@SFIScience
👀
The viral "Definition of AGI" paper tells you to read fake references which do not exist!
Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
Take this as a warning to not use LMs to generate your references!
Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.
Take this as a warning to not use LMs to generate your references!
October 18, 2025 at 5:13 PM
👀
Reposted by Melanie Mitchell
This is the one, folks. If you possibly, possibly can, we need you to be there. Find your peaceful gathering at nokings.org
October 17, 2025 at 2:04 PM
This is the one, folks. If you possibly, possibly can, we need you to be there. Find your peaceful gathering at nokings.org
Well-worth watching. ⬇️
Earlier this year, after receiving emails from people in the throes of AI psychosis, I stayed in touch with one man as he journeyed to recovery.
I'm grateful to James that he was willing to go on record. He wanted people to know: humanlike chatbots are dangerous.
www.youtube.com/watch?v=zkGk...
I'm grateful to James that he was willing to go on record. He wanted people to know: humanlike chatbots are dangerous.
www.youtube.com/watch?v=zkGk...
He Lost His Mind Using ChatGPT. Then It Told Him to Contact Me.
YouTube video by More Perfect Union
www.youtube.com
October 15, 2025 at 6:16 PM
Well-worth watching. ⬇️
The lifecycle of software objects
Sam Altman says ChatGPT will soon sext with verified adults
Sam Altman says ChatGPT will soon sext with verified adults
ChatGPT is about to get flirty.
buff.ly
October 14, 2025 at 9:01 PM
The lifecycle of software objects
Reposted by Melanie Mitchell
Reserve your free tickets to SFI’s next Community Lecture with renowned developmental psychologist Alison Gopnik:
“Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?”
October 21, 7:30 pm at The Lensic Performing Arts Center.
Tickets: lensic.org/events/aliso...
“Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?”
October 21, 7:30 pm at The Lensic Performing Arts Center.
Tickets: lensic.org/events/aliso...
October 7, 2025 at 7:02 PM
Reserve your free tickets to SFI’s next Community Lecture with renowned developmental psychologist Alison Gopnik:
“Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?”
October 21, 7:30 pm at The Lensic Performing Arts Center.
Tickets: lensic.org/events/aliso...
“Transmission Versus Truth: What Will It Take to Make an AI as Smart as a 4-Year-Old?”
October 21, 7:30 pm at The Lensic Performing Arts Center.
Tickets: lensic.org/events/aliso...
Do AI reasoning models abstract and reason like humans?
New paper on this from my group:
arxiv.org/abs/2510.02125
🧵 1/10
New paper on this from my group:
arxiv.org/abs/2510.02125
🧵 1/10
Do AI Models Perform Human-like Abstract Reasoning Across Modalities?
OpenAI's o3-preview reasoning model exceeded human accuracy on the ARC-AGI benchmark, but does that mean state-of-the-art models recognize and reason with the abstractions that the task creators inten...
arxiv.org
October 6, 2025 at 9:27 PM
Do AI reasoning models abstract and reason like humans?
New paper on this from my group:
arxiv.org/abs/2510.02125
🧵 1/10
New paper on this from my group:
arxiv.org/abs/2510.02125
🧵 1/10
Reposted by Melanie Mitchell
Since it resonated with the audience, I’ll recap my main argument against AGI here. ‘General intelligence’ is like phlogiston, or the aether. It’s an outmoded scientific concept that does not refer to anything real. Any explanatory work it did can be done better by a richer scientific frame. 1/3
This was a truly heartening day, with deeply thoughtful challenges to the dominant narrative framed around AGI, coming from across disciplines and perspectives. Felt like the tide might finally be turning a bit, at least among the scientific community. Thanks @royalsociety.org!
1/2 I'm looking forward to taking part in a panel on AGI and the Turing Test, tomorrow afternoon (Thurs 2nd Oct) at the @royalsociety.org, w/ Dame Wendy Hall, Shannon Vallor, William Isaac, & Sir Nigel Shadbolt. royalsociety.org/science-even...
October 2, 2025 at 10:09 PM
Since it resonated with the audience, I’ll recap my main argument against AGI here. ‘General intelligence’ is like phlogiston, or the aether. It’s an outmoded scientific concept that does not refer to anything real. Any explanatory work it did can be done better by a richer scientific frame. 1/3
In honor of the 75th anniversary of the Turing test, I'm re-upping my short essay, "The Turing Test and our shifting conceptions of intelligence".
www.science.org/doi/10.1126/...
www.science.org/doi/10.1126/...
The Turing Test and our shifting conceptions of intelligence
“Can machines think?” So asked Alan Turing in his 1950 paper, “Computing Machinery and Intelligence.” Turing quickly noted that, given the difficulty of defining thinking, the question is “too meaning...
www.science.org
October 2, 2025 at 2:50 PM
In honor of the 75th anniversary of the Turing test, I'm re-upping my short essay, "The Turing Test and our shifting conceptions of intelligence".
www.science.org/doi/10.1126/...
www.science.org/doi/10.1126/...
In 2016 Hinton predicted that AI would replace all radiologists in five years. Ten years later, why hasn't it happened? This post is a great explainer.
www.understandingai.org/p/ai-isnt-re...
www.understandingai.org/p/ai-isnt-re...
AI isn't replacing radiologists
Radiology combines digital images, clear benchmarks, and repeatable tasks. But demand for human radiologists is at an all-time high.
www.understandingai.org
October 1, 2025 at 2:35 PM
In 2016 Hinton predicted that AI would replace all radiologists in five years. Ten years later, why hasn't it happened? This post is a great explainer.
www.understandingai.org/p/ai-isnt-re...
www.understandingai.org/p/ai-isnt-re...
Andrew Ng: "AI is the new electricity!"
Cory Doctorow: "AI is the asbestos we are shoveling into the walls of our society and our descendants will be digging it out for generations."
¯\_(ツ)_/¯
(pluralistic.net/2025/09/27/e...)
Cory Doctorow: "AI is the asbestos we are shoveling into the walls of our society and our descendants will be digging it out for generations."
¯\_(ツ)_/¯
(pluralistic.net/2025/09/27/e...)
September 30, 2025 at 11:19 PM
Andrew Ng: "AI is the new electricity!"
Cory Doctorow: "AI is the asbestos we are shoveling into the walls of our society and our descendants will be digging it out for generations."
¯\_(ツ)_/¯
(pluralistic.net/2025/09/27/e...)
Cory Doctorow: "AI is the asbestos we are shoveling into the walls of our society and our descendants will be digging it out for generations."
¯\_(ツ)_/¯
(pluralistic.net/2025/09/27/e...)
⬇️⬇️
Please everyone post comments on this rule!
This rule change will be devastating to our valued international students and scholars who contribute tremendously to US universities and other institutions with STEM research and development.
This rule change will be devastating to our valued international students and scholars who contribute tremendously to US universities and other institutions with STEM research and development.
DHS is proposing a rule that would end "duration of status" for J, F and other visa holders. This would decimate the international graduate student and postdoc population, which are a crucial part of the United States biomedical workforce. There's still time for you to comment on the proposed rule.
September 29, 2025 at 5:34 PM
⬇️⬇️