Olivia Guest · Ολίβια Γκεστ
@olivia.science
https://olivia.science
assistant professor of computational cognitive science · she/they · cypriot/kıbrıslı/κυπραία · σὺν Ἀθηνᾷ καὶ χεῖρα κίνει
assistant professor of computational cognitive science · she/they · cypriot/kıbrıslı/κυπραία · σὺν Ἀθηνᾷ καὶ χεῖρα κίνει
Pinned
CAIL = prerequisite knowledge for a critical perspective, such as to tell apart nonsense hype from true theoretical computer scientific claims. For example, the idea that human-like systems are a sensible or possible goal is the result of circular reasoning and anthropomorphism. olivia.science/ai
🧵
🧵
Reposted by Olivia Guest · Ολίβια Γκεστ
So Musk didn't accidentally spend billions making what became a child porn machine, he purposely spent billions to make a child porn machine.
Grok producing vile sexualized deepfakes of women and children isn’t an accident
www.cnn.com/2026/01/08/t...
www.cnn.com/2026/01/08/t...
January 9, 2026 at 2:32 AM
So Musk didn't accidentally spend billions making what became a child porn machine, he purposely spent billions to make a child porn machine.
Reposted by Olivia Guest · Ολίβια Γκεστ
Holy fuck they actually did it
January 8, 2026 at 11:44 PM
Holy fuck they actually did it
Reposted by Olivia Guest · Ολίβια Γκεστ
to avoid making certain kinds of fuck ups (in highly idealised scenarios). But it doesn't help me much if I want to do the actual hard thing, which is to find out if my model can safely be used in the ways I would like it to be.
January 4, 2026 at 8:11 PM
to avoid making certain kinds of fuck ups (in highly idealised scenarios). But it doesn't help me much if I want to do the actual hard thing, which is to find out if my model can safely be used in the ways I would like it to be.
Reposted by Olivia Guest · Ολίβια Γκεστ
have a DAG and they know about the do-operator. But it does not much for me RE the things that are most important about doing good statistical science. At least, that's my gut feeling but I am a cynic at heart...
I guess it just reads to me like causal inference is a framework
I guess it just reads to me like causal inference is a framework
January 4, 2026 at 8:11 PM
have a DAG and they know about the do-operator. But it does not much for me RE the things that are most important about doing good statistical science. At least, that's my gut feeling but I am a cynic at heart...
I guess it just reads to me like causal inference is a framework
I guess it just reads to me like causal inference is a framework
Reposted by Olivia Guest · Ολίβια Γκεστ
psychology will be saved. Well no because it it no way solves the things that are actually hard about doing good science. Bad science can, and will, replicate.
I just get a similar feeling. Bad science can (and will) be dressed up to look like evidence for causal theories because, hey they
I just get a similar feeling. Bad science can (and will) be dressed up to look like evidence for causal theories because, hey they
January 4, 2026 at 8:11 PM
psychology will be saved. Well no because it it no way solves the things that are actually hard about doing good science. Bad science can, and will, replicate.
I just get a similar feeling. Bad science can (and will) be dressed up to look like evidence for causal theories because, hey they
I just get a similar feeling. Bad science can (and will) be dressed up to look like evidence for causal theories because, hey they
Reposted by Olivia Guest · Ολίβια Γκεστ
Some half-formed thoughts on causal inference bc I've been reading about it very casually. (fucking yes, managed to spell them both right first go)
I get similar feelings to when I was reading about how the replication crisis would be solved by pre-registration and so
I get similar feelings to when I was reading about how the replication crisis would be solved by pre-registration and so
January 4, 2026 at 8:11 PM
Some half-formed thoughts on causal inference bc I've been reading about it very casually. (fucking yes, managed to spell them both right first go)
I get similar feelings to when I was reading about how the replication crisis would be solved by pre-registration and so
I get similar feelings to when I was reading about how the replication crisis would be solved by pre-registration and so
Reposted by Olivia Guest · Ολίβια Γκεστ
Remembering today that having your heart broken is a necessary step on the path to becoming fully human. Whichever heartbreak is your first, it’s probably critical that a state break your heart so that you can develop a political imagination. If this is your first, I’m sorry and also welcome.
January 8, 2026 at 10:33 PM
Remembering today that having your heart broken is a necessary step on the path to becoming fully human. Whichever heartbreak is your first, it’s probably critical that a state break your heart so that you can develop a political imagination. If this is your first, I’m sorry and also welcome.
Reposted by Olivia Guest · Ολίβια Γκεστ
In one of the most appalling developments of past day, Elon Musk's AI chatbot desecrated Renee Nicole Good's body. That's in addition to harassing women and sexually exploiting minors. By @katieherchenroeder.bsky.social
www.motherjones.com/politics/202...
www.motherjones.com/politics/202...
Grok deepfaked Renee Nicole Good's body into a bikini
Hours after an ICE agent killed the mother of three, Elon Musk’s chatbot was undressing her.
www.motherjones.com
January 8, 2026 at 9:41 PM
In one of the most appalling developments of past day, Elon Musk's AI chatbot desecrated Renee Nicole Good's body. That's in addition to harassing women and sexually exploiting minors. By @katieherchenroeder.bsky.social
www.motherjones.com/politics/202...
www.motherjones.com/politics/202...
Reposted by Olivia Guest · Ολίβια Γκεστ
Wait wait wait, while Americans have to pay hundreds of billions of dollars in new tariffs, Trump is going send over $10,000 to each person in *Greenland*?????????
This is a joke right ?
www.reuters.com/world/europe...
This is a joke right ?
www.reuters.com/world/europe...
Exclusive: Trump administration mulls payments to sway Greenlanders to join US
U.S. officials have discussed sending lump sum payments to Greenlanders as part of a bid to convince them to secede from Denmark and potentially join the United States, according to four sources famil...
www.reuters.com
January 8, 2026 at 8:12 PM
Wait wait wait, while Americans have to pay hundreds of billions of dollars in new tariffs, Trump is going send over $10,000 to each person in *Greenland*?????????
This is a joke right ?
www.reuters.com/world/europe...
This is a joke right ?
www.reuters.com/world/europe...
Reposted by Olivia Guest · Ολίβια Γκεστ
I feel like this is extremely relevant again
Not “this was not self-defense” but “this was murder”
Not “she didn’t run the ICE agent over” but “she was murdered while trying to drive away”
Not “this was not self-defense” but “this was murder”
Not “she didn’t run the ICE agent over” but “she was murdered while trying to drive away”
i made a helpful reference
January 8, 2026 at 9:24 PM
I feel like this is extremely relevant again
Not “this was not self-defense” but “this was murder”
Not “she didn’t run the ICE agent over” but “she was murdered while trying to drive away”
Not “this was not self-defense” but “this was murder”
Not “she didn’t run the ICE agent over” but “she was murdered while trying to drive away”
Reposted by Olivia Guest · Ολίβια Γκεστ
start with what it is (not what it isn't)
focus on why they're lying (not their lies)
make it about them (not her)
eat their frame
focus on why they're lying (not their lies)
make it about them (not her)
eat their frame
I feel like this is extremely relevant again
Not “this was not self-defense” but “this was murder”
Not “she didn’t run the ICE agent over” but “she was murdered while trying to drive away”
Not “this was not self-defense” but “this was murder”
Not “she didn’t run the ICE agent over” but “she was murdered while trying to drive away”
i made a helpful reference
January 8, 2026 at 9:32 PM
start with what it is (not what it isn't)
focus on why they're lying (not their lies)
make it about them (not her)
eat their frame
focus on why they're lying (not their lies)
make it about them (not her)
eat their frame
Reposted by Olivia Guest · Ολίβια Γκεστ
"We have built a MACHINE with real HUMAN INTELLIGENCE!" ...
"Sir, the product you use is literally called 'Mechanical Turk'"
"Sir, the product you use is literally called 'Mechanical Turk'"
January 8, 2026 at 9:18 PM
"We have built a MACHINE with real HUMAN INTELLIGENCE!" ...
"Sir, the product you use is literally called 'Mechanical Turk'"
"Sir, the product you use is literally called 'Mechanical Turk'"
Reposted by Olivia Guest · Ολίβια Γκεστ
The implication is severe: careless responding on MTurk isn’t rare noise; it’s systematic enough to flip the sign of relationships and generate results that are the opposite of what they really are.
Wow; this is pretty damning.
link.springer.com/article/10.3...
Wow; this is pretty damning.
link.springer.com/article/10.3...
January 8, 2026 at 8:49 PM
The implication is severe: careless responding on MTurk isn’t rare noise; it’s systematic enough to flip the sign of relationships and generate results that are the opposite of what they really are.
Wow; this is pretty damning.
link.springer.com/article/10.3...
Wow; this is pretty damning.
link.springer.com/article/10.3...
Reposted by Olivia Guest · Ολίβια Γκεστ
Researchers have long argued about whether Amazon Mechanical Turk (MTurk) survey data can be trusted.
This paper takes a simple approach to evaluating the quality of data currently produced by MTurk.
The author gives respondents pairs of questions that are obviously contradictory.
This paper takes a simple approach to evaluating the quality of data currently produced by MTurk.
The author gives respondents pairs of questions that are obviously contradictory.
January 8, 2026 at 8:47 PM
Researchers have long argued about whether Amazon Mechanical Turk (MTurk) survey data can be trusted.
This paper takes a simple approach to evaluating the quality of data currently produced by MTurk.
The author gives respondents pairs of questions that are obviously contradictory.
This paper takes a simple approach to evaluating the quality of data currently produced by MTurk.
The author gives respondents pairs of questions that are obviously contradictory.
Reposted by Olivia Guest · Ολίβια Γκεστ
“These findings provide clear evidence that data collected on MTurk simply cannot be trusted.”
January 8, 2026 at 8:47 PM
“These findings provide clear evidence that data collected on MTurk simply cannot be trusted.”
Reposted by Olivia Guest · Ολίβια Γκεστ
I reached out to more than three dozen X advertisers, xAI investors, politicians, and government agencies about the growing library of sexualized deepfakes generated with Grok—now including Renee Good. I only heard back from four, with three declining to comment.
spitfirenews.com/p/grok-deepf...
spitfirenews.com/p/grok-deepf...
Why isn't there a bigger Grok boycott?
Advertisers, politicians, and investors are still all-in on X, despite a sexual abuse crisis.
spitfirenews.com
January 8, 2026 at 9:04 PM
I reached out to more than three dozen X advertisers, xAI investors, politicians, and government agencies about the growing library of sexualized deepfakes generated with Grok—now including Renee Good. I only heard back from four, with three declining to comment.
spitfirenews.com/p/grok-deepf...
spitfirenews.com/p/grok-deepf...
Reposted by Olivia Guest · Ολίβια Γκεστ
Anthropomorphizing language conceals the limitations of AI, promoting misplaced trust. @emilymbender.bsky.social & @nannainie.bsky.social suggest focusing on a system’s functionalities: instead of saying a model is “good at” something, say what it is “good for." www.techpolicy.press/we-need-to-t...
We Need to Talk About How We Talk About 'AI' | TechPolicy.Press
We share a responsibility to create and use empowering metaphors rather than misleading language, write Emily M. Bender and Nanna Inie.
www.techpolicy.press
January 8, 2026 at 8:14 PM
Anthropomorphizing language conceals the limitations of AI, promoting misplaced trust. @emilymbender.bsky.social & @nannainie.bsky.social suggest focusing on a system’s functionalities: instead of saying a model is “good at” something, say what it is “good for." www.techpolicy.press/we-need-to-t...
Reposted by Olivia Guest · Ολίβια Γκεστ
Excellent thread validating what many of us find we need to do to make our environment less disruptive & overwhelming.
Boundaries are so important. Even (especially?) with ourselves.
Boundaries are so important. Even (especially?) with ourselves.
Going to start a thread of things I’m doing that are helping me get out of autistic (auDHD possibly, we’ll see) burnout. TW mental health
January 8, 2026 at 8:16 PM
Excellent thread validating what many of us find we need to do to make our environment less disruptive & overwhelming.
Boundaries are so important. Even (especially?) with ourselves.
Boundaries are so important. Even (especially?) with ourselves.
Reposted by Olivia Guest · Ολίβια Γκεστ
*most students
Students get it:
> She also argues that the use of AI can mean that “people aren’t critically thinking” any more. She feels that students who use GenAI consistently “can’t see another viewpoint”.
www.irishtimes.com/life-style/p...
> She also argues that the use of AI can mean that “people aren’t critically thinking” any more. She feels that students who use GenAI consistently “can’t see another viewpoint”.
www.irishtimes.com/life-style/p...
Students’ complicated relationship with AI: ‘Avoiding using it probably put me at a bit of a disadvantage’
As students get stuck into the academic year, there are strong feelings about whether artificial intelligence should play a role in their education
www.irishtimes.com
January 8, 2026 at 9:32 PM
*most students
They crave this for free since they cannot get it any other way. They are scamming the golden goose because no farmer would let those eggs go even for money. Please please please don't do it.
I cannot stress enough that you should not do this.
OpenAI launches ChatGPT Health, encouraging users to connect their medical records — The Verge
OpenAI has been dropping hints this week about AI’s role as a “healthcare ally” — and today, the company is announcing a product to go along with that idea: ChatGPT Health.
apple.news
January 8, 2026 at 9:28 PM
They crave this for free since they cannot get it any other way. They are scamming the golden goose because no farmer would let those eggs go even for money. Please please please don't do it.
Reposted by Olivia Guest · Ολίβια Γκεστ
good god, please do not do this
Do not trust this company with ANYTHING sensitive or personal.
Or any of them.
Do not trust this company with ANYTHING sensitive or personal.
Or any of them.
I cannot stress enough that you should not do this.
OpenAI launches ChatGPT Health, encouraging users to connect their medical records — The Verge
OpenAI has been dropping hints this week about AI’s role as a “healthcare ally” — and today, the company is announcing a product to go along with that idea: ChatGPT Health.
apple.news
January 8, 2026 at 9:04 PM
good god, please do not do this
Do not trust this company with ANYTHING sensitive or personal.
Or any of them.
Do not trust this company with ANYTHING sensitive or personal.
Or any of them.
Reposted by Olivia Guest · Ολίβια Γκεστ
Going to start a thread of things I’m doing that are helping me get out of autistic (auDHD possibly, we’ll see) burnout. TW mental health
January 7, 2026 at 1:20 PM
Going to start a thread of things I’m doing that are helping me get out of autistic (auDHD possibly, we’ll see) burnout. TW mental health
Reposted by Olivia Guest · Ολίβια Γκεστ
this has been brewing for years & we've been saying as much for years only to be dismissed over & over as paranoid, hysterical, alarmist, or as somehow unfaithful witnesses to our own lived experiences.
they want us all dead.
they will try to kill as many of us as they can get away with.
they want us all dead.
they will try to kill as many of us as they can get away with.
“I think what they’re doing here is they’re trying to destroy a gender group,” he added. “And so I do think it’s genocidal. I think that the objective here is literally, physically to destroy this group.”
www.lemkininstitute.com/single-post/...
www.lemkininstitute.com/single-post/...
Experts Warn U.S. in Early Stages of Genocide Against Trans Americans
Genocide scholars say policies targeting transgender Americans match early warning signs of mass atrocity.Genocide scholars are sounding the alarm over what they describe as escalating attacks targeti...
www.lemkininstitute.com
January 8, 2026 at 7:19 PM
this has been brewing for years & we've been saying as much for years only to be dismissed over & over as paranoid, hysterical, alarmist, or as somehow unfaithful witnesses to our own lived experiences.
they want us all dead.
they will try to kill as many of us as they can get away with.
they want us all dead.
they will try to kill as many of us as they can get away with.
Reposted by Olivia Guest · Ολίβια Γκεστ
Worth waiting for! @alexhanna.bsky.social and I did a LOT of podcast guesting last spring & summer and this one, one of my favorites, is finally available:
www.youtube.com/watch?v=LJmk...
www.youtube.com/watch?v=LJmk...
Thinking Through...The AI Con & Deconstructing the Hype
YouTube video by Thinking Through Podcast
www.youtube.com
January 8, 2026 at 3:29 PM
Worth waiting for! @alexhanna.bsky.social and I did a LOT of podcast guesting last spring & summer and this one, one of my favorites, is finally available:
www.youtube.com/watch?v=LJmk...
www.youtube.com/watch?v=LJmk...
Reposted by Olivia Guest · Ολίβια Γκεστ
The number of people who believe CSI was real and the state can just “biometric” up the identity of the Minnesota killer, and therefore the fact they haven’t done so in less than 24 hours is a failure of political will that means nothing will ever be done is, uh, certainly something.
January 8, 2026 at 6:29 PM
The number of people who believe CSI was real and the state can just “biometric” up the identity of the Minnesota killer, and therefore the fact they haven’t done so in less than 24 hours is a failure of political will that means nothing will ever be done is, uh, certainly something.