![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
kradeelav* neat gamedev writeup on LoZ: Ocarina of Time talking about some meta-gameplay difficulties the devs had to work through in regards to designing the first 3D action game. Fascinating to see how most modern games like Elden Ring today still operate off of similar spatial principles.
* i was talking to a friend on bsky about LLM's and reducing friction of delusions; bellingcat's founder Elliot H. posted a really good thread a while back that I keep thinking about. (I have my reservations about bellingcat in general and where they get their funding but this feels insightful.)
copy/pasted thread text below:
"Bellingcat’s contact email has always been a magnet for people with fairly unusual views; paranoid delusions, sprawling conspiracies, the works. But recently, the pattern has shifted, we’re seeing more and more emails clearly written with ChatGPT. Some explicitly say the AI recommended they contact us. That makes sense, the models scrape our work and cite us as a reputable source, so the AI has inadvertently become part of the referral pathway for people spiralling down a rabbit hole. The bigger change is presentation, historically, these emails were visually chaotic: sudden colour changes, ALL CAPS, screenshots pasted at random. Classic signs of disordered thinking. Now they’re neat, structured, and grammatically consistent. The delusion hasn't changed, but the formatting has. LLMs give people with delusional belief systems a way to organise those beliefs. The model fills in gaps, supplies connective tissue, and wraps the madness in formal language. It doesn't make the claims more credible to us, it just tidies them up.
Previously, these belief systems tended to collapse under their own weight. Contradictions and memory gaps made them hard to sustain. Now the AI removes that friction, it produces timelines, summaries, and "forensic" explanations on demand. I've written about "disordered counterpublics" groups that reinforce shared delusions Here, the LLM becomes a 1 person counterpublic. It mirrors assumptions, absorbs contradictions, & provides the sense of a responsive, confirming audience. It is the architecture of disordered discourse, automated.
The content is still obviously delusional to anyone reading it. It’s just organised crazy. But the real danger is the internal effect on the user. The AI provides coherence and validation, stabilising the narrative they’re caught in. We’re watching a shift where AI simulates the cohesion and reinforcement that used to require an entire online community. A single person can now build a complete parallel reality with nothing more than a prompt window. Which seems bad."
* Bart Ehrman's last lecture - "The Most Significant Discovery in the History of Biblical Studies”. haven't been in these circles for a while, this was a refreshingly thoughtful lecture/sermon of sorts.
* one of the latest pieces of news this week is firefox has a new CEO who apparently is commited to driving a perfectly decent product off a cliff; they're forcing AI upgrades and there's rumors of disabling extensions in the future, but that's less concrete. While there is a follow-up clarification that the devs are trying hard to keep AI explicitly opt in / have a kill switch for it, i'm less certian that they'll be able to keep that from CEO pressure.
if you feel strongly about this, i've given feedback at https://connect.mozilla.org/ and suggest you do as well. For immediate AI-free alternatives, waterfox and vivaldi are the two strongest ones I consistently see in rec lists.
* thin desires vs thick desires - thoughtful articulation on a concept i've been seeing a lot lately. (though i have my reservations that quote unquote pornography in its entirety is a thin desire since to me sexual desire in media is pretty nuanced; i see where the thought line comes from, but that's a side tangent.)
* escape clause - a blogger decides to move away from the apple ecosystem after seeing another dev elsewhere get locked out of mission critical email accounts. (i really liked the empathy to apple devs at the same time of concretely outlining why he's moving away)
* i was talking to a friend on bsky about LLM's and reducing friction of delusions; bellingcat's founder Elliot H. posted a really good thread a while back that I keep thinking about. (I have my reservations about bellingcat in general and where they get their funding but this feels insightful.)
copy/pasted thread text below:
"Bellingcat’s contact email has always been a magnet for people with fairly unusual views; paranoid delusions, sprawling conspiracies, the works. But recently, the pattern has shifted, we’re seeing more and more emails clearly written with ChatGPT. Some explicitly say the AI recommended they contact us. That makes sense, the models scrape our work and cite us as a reputable source, so the AI has inadvertently become part of the referral pathway for people spiralling down a rabbit hole. The bigger change is presentation, historically, these emails were visually chaotic: sudden colour changes, ALL CAPS, screenshots pasted at random. Classic signs of disordered thinking. Now they’re neat, structured, and grammatically consistent. The delusion hasn't changed, but the formatting has. LLMs give people with delusional belief systems a way to organise those beliefs. The model fills in gaps, supplies connective tissue, and wraps the madness in formal language. It doesn't make the claims more credible to us, it just tidies them up.
Previously, these belief systems tended to collapse under their own weight. Contradictions and memory gaps made them hard to sustain. Now the AI removes that friction, it produces timelines, summaries, and "forensic" explanations on demand. I've written about "disordered counterpublics" groups that reinforce shared delusions Here, the LLM becomes a 1 person counterpublic. It mirrors assumptions, absorbs contradictions, & provides the sense of a responsive, confirming audience. It is the architecture of disordered discourse, automated.
The content is still obviously delusional to anyone reading it. It’s just organised crazy. But the real danger is the internal effect on the user. The AI provides coherence and validation, stabilising the narrative they’re caught in. We’re watching a shift where AI simulates the cohesion and reinforcement that used to require an entire online community. A single person can now build a complete parallel reality with nothing more than a prompt window. Which seems bad."
* Bart Ehrman's last lecture - "The Most Significant Discovery in the History of Biblical Studies”. haven't been in these circles for a while, this was a refreshingly thoughtful lecture/sermon of sorts.
* one of the latest pieces of news this week is firefox has a new CEO who apparently is commited to driving a perfectly decent product off a cliff; they're forcing AI upgrades and there's rumors of disabling extensions in the future, but that's less concrete. While there is a follow-up clarification that the devs are trying hard to keep AI explicitly opt in / have a kill switch for it, i'm less certian that they'll be able to keep that from CEO pressure.
if you feel strongly about this, i've given feedback at https://connect.mozilla.org/ and suggest you do as well. For immediate AI-free alternatives, waterfox and vivaldi are the two strongest ones I consistently see in rec lists.
* thin desires vs thick desires - thoughtful articulation on a concept i've been seeing a lot lately. (though i have my reservations that quote unquote pornography in its entirety is a thin desire since to me sexual desire in media is pretty nuanced; i see where the thought line comes from, but that's a side tangent.)
* escape clause - a blogger decides to move away from the apple ecosystem after seeing another dev elsewhere get locked out of mission critical email accounts. (i really liked the empathy to apple devs at the same time of concretely outlining why he's moving away)
