6 ms·

Defining AI Psychosis. Part 2: "Prolific AI Psychosis"

60 points by euthymiclabs 8d ago

altcognito 8d ago
People really need an entire article to justify labeling someone as having a mental illness because in their opinion, the their coworker isn't contributing meaninfully, or potentially making things worse. FTA: > Prolific AI psychosis occurs when a person generates a large quantity of AI output without significantly increasing the real value of their work. Guy, we used to just call this a bad coworker. I'm pretty sure that the "psychosis" will be abused by folks that just don't like AI. You already see it everywhere.
- recursive 8d ago
  > Guy, we used to just call this a bad coworker. Not that long ago. People generating large quantities of AI output is a relatively recent phenomenon. At least when compared to labels like this.
  - altruios 8d ago
    low effort work though, that's nothing new... accelerating that low effort: that is new... but the low effort work hasn't been changed by ai: only illuminated more clearly.
  - 2PqboPPmKegvanx 8d ago
    its the same individuals in my experience. i don't see people who used to be good employees suddenly sending a 3000 word email, but i sure do see that out of the people who were bad coworkers 12 months ago and still are today.
- js8 8d ago
  I was bound to agree with you before I RTFA. The classification is useful for a psychiatrist, i.e. someone who wants to help the person. "Bad coworker" can have many causes. In prolific AI psychosis, I think the assumption (differential diagnosis) is, they became a "bad coworker" due to use of AI, so they were a "good coworker" before.
  - altcognito 8d ago
    As someone else point out, the term is stigmatizing, psychosis is a very strong term, and as I very much noted, already being misused and abused.
    - js8 7d ago
      No, the term itself is not stigmatizing. People stigmatize legitimate psychiatric diagnoses, and you're perpetuating this idea in your comment, by blaming the term for the moral failure. These labels don't exists to stigmatize people, but to study the problem and help people.
- onlyrealcuzzo 8d ago
  Song as old as time - people were doing this well before AI, there's just yet another reason for these types of people to categorize workers and pretend that management can be perfect and easy and that the world isn't hard. It's almost as if it's a little bit harder than people think to figure out how good people are at their jobs and who deserves promotions, etc, and that maybe they may not have the most perfect unbiased judgements.
- beepbooptheory 8d ago
  You certainly seem to have your mind made up about this either way, but a humble fwiw nevertheless: the author is a psychiatrist, not some bitter dev/hypothetical coworker like you seem to be suggesting.
  - altcognito 8d ago
    I did dig into that after I read it. Even more depressing honestly. He should know better than to throw around labels.
    - ethin 8d ago
      Sorry but the pro-AI/AI boosters do the exact same thing when they say that people who don't use AI are "Luddites" and the like. And I'm pretty sure the Luddite thing has been going on for far longer than the AI psychosis labeling.
      - altcognito 8d ago
        Also a luddite is not thought of as a danger to others. The label of psychosis implies otherwise. Luddite extends beyond AI Whataboutism is a plague. That said, calling folks luddites is an annoyance.
    - beepbooptheory 8d ago
      Is it that psychiatrists in general could maybe use this label, but you know this one is doing it wrong? Otherwise, who is allowed to use the term? Doesn't it kinda inspire you at least a little to reflect on your understanding of the particular term? Or should we understand that to be faultless here? In general, the definition we need to have here is that psychosis is always extreme and violent and pretty much matching the cultural trope we have for it? And anything less is incorrect usage? Is that why you disagree in explicit terms?
      - altcognito 8d ago
        > Is it that psychiatrists in general could maybe use this label, but you know this one is doing it wrong? Otherwise, who is allowed to use the term? I specifically take issue with the weak standard used. > Doesn't it kinda inspire you at least a little to reflect on your understanding of the particular term? Or should we understand that to be faultless here? Is there a way to read this that isn't insulting? Do I need to reflect on my understanding until it matches yours? > In general, the definition we need to have here is that psychosis is always extreme and violent and pretty much matching the cultural trope we have for it? And anything less is incorrect usage? Is that why you disagree in explicit terms? Never said any of that. It's a medical term and carries weight. Core symptoms are hallucinations and delusions. Here, delusions is not "my corporate product is better than your corporate product", it is not a strongly held belief that isn't real. We don't call belief in God psychosis. (some do, but that's beside the point)
- _verandaguy 8d ago
  I'll disagree, I think AI meaningfully changes the definition of a bad coworker because its main interface emulates talking to another human being closely enough that some people value its conversational input equally or more highly than that of actual human beings. I think that this is exacerbated by many LLMs defaulting to placating, complementing, and otherwise reinforcing biases in their users, and it will often produce just enough evidence to make its (and its user's) position seem compelling, while that evidence is often either tangential, orthogonal, or straight-up hallucinated. Humans aren't perfect vessels for information, but working on a reasonably-qualified, healthy team, a degree of respectful professional disagreement has always been the norm, and that covers both cases where the disagreement comes from a genuine matter of opinion, as well as cases where one colleague cross-checks anothers' assumptions and finds some of them to be wrong. That's the process.
peab 8d ago
I was talking to a friend recently, and i think i aligned with a similar definition. I've seen people who talk nonstop about AI, and how it's changed everything, but then there's no quantifiable output that supports their claim. On the other hand, I've had a friend start a business, and be able to create an app much much faster than what would have been possible in the past. He has real customers paying real money, which makes it easy to validate his claims.
- fnord77 8d ago
  > I've seen people who talk nonstop about ________, and how it's changed everything, but then there's no quantifiable output that supports their claim. These people have existed forever
- ActorNightly 8d ago
  Back in the 2010s, the best, most efficient software engineers were characterized by 3 things * end to end rough map of the entire space of compute in their head * ability to search the internet for the right things * ability to quickly experiment and try things to figure out how to do things AI hasn't change that, it just made 2 and 3 into a very efficient thing. The characterization of psychosis is best described by believing AI can do the first thing. No modern LLM can "reason" - otherwise you could give it a task like "make me money", it would ask you all the questions it needs about information that it doesn't know about and needs to know to make you money, then it would set whatever it needs to set up to make you money. As such, you still need to know the domain entirely to be effective. When you do that, AI is fantastic at getting you to the right solution. Furthermore, its still in large part actually cheaper to higher a developer who then can use AI to build you the product that you need long term.
  - jonas21 8d ago
    > No modern LLM can "reason" - otherwise you could give it a task like "make me money", it would ask you all the questions it needs about information that it doesn't know about and needs to know to make you money, then it would set whatever it needs to set up to make you money. If that's your bar for reasoning, then most people can't reason either.
ilc_genesis_01 8d ago
I think that specifically in the case of a coworker in a technical field, say software engineer, we already have an apprenticeship system (junior, senior, architect, etc) where the use of these tools, taste, competence, evaluating code and architecture, are all experience-based and pretty self evident as you progress through your profession and gain experience. The same for, say, an electrician or any skilled trade. These AI tools are just those, another tool that you use to further your profession and of course somebody starting out will perhaps generate more slop or bad code. But more senior coworkers with experience, taste, criteria, etc will guide those people and you progress and improve. Beyond that, when the tool becomes more than a tool and becomes your coworker, with his/her/its own criteria, taste, methods, etc - than THAT is when the rubber hits the road. haha.
2PqboPPmKegvanx 8d ago
Part 3: "Voluminous AI Psychosis"
paimapi 8d ago
One - delusion is a better fit if you need a term: https://mhanational.org/conditions/paranoia-and-delusional-disorders/ https://mhanational.org/conditions/paranoia-and-delusional-d... > Delusional disorder is characterized by irrational or intense belief(s) or suspicion(s) that a person believes to be true it's still a bad fit because there's not currently popular social convention that trusting AI without validation is a bad idea but it's better than psychosis. also it's still stigmatizing and we live in a world where a thesaurus is seconds away at any time. plus psychosis as a diagnosis is typically a cluster of symptoms that go beyond just delusion: https://emedicine.medscape.com/article/294416-overview https://emedicine.medscape.com/article/294416-overview Two - there's also an implication that there's 'harm' that this AI psychotic person inflicts on others embedded in this article. in reality, the vast majority of those suffering mental illnesses are far more likely to be harmed by others - https://www.betterhealth.vic.gov.au/health/conditionsandtreatments/mental-illness-and-violence https://www.betterhealth.vic.gov.au/health/conditionsandtrea... part of the work of being a good mental health advocate is not stigmatizing specific diagnoses as negative social ills. now if you don't give a single shit about mental health I'm not going to argue to get you to care. but if you do care then you should do better
- datakan 8d ago
  > Psychosis as a diagnosis is a serious condition. there's an implication that there's 'harm' I just watched a company fire an entire department of people under the sincere belief a chatbot run by a middle manager could save them 10 million a year. This failed miserably and they now need to hire or re-hire people. There is a tremendous amount of harm happening. Psychosis is a perfect term for what is happening.
  - paimapi 8d ago
    it absolutely is not. when we think of harm in a diagnosis of psychotic disorders, we're talking about things like getting shot to death by the police, running out onto the highway, etc. it is not callously making a lot of people have to jobhunt which isn't to say that layoffs don't suck (I was impacted by one very recently) but it's not psychosis that causes that to happen, it's greed and short-term shareholder-fixated thinking. that should be named instead of a mental illness that people suffer from
bobkb 8d ago
Excellent read! We were talking about the effects of AI Psychosis on companies. It looks like people have a misconceived notion that now work can be done 90% faster and they are attempting to literally get rid of 90% of the engineers. The end result, no one left with do the work - especially when an edge case/disaster strikes or the customer wants something different.
bob1029 8d ago
> In some cases, the new AI workflow may even destroy value. I am really struggling with this one right now. The client's development team is letting Claude document and refactor the product codebase without much intervention, purpose or taste. The thinking being that if Claude first writes a bunch of unit tests and then a bunch of documentation, that any downstream refactoring will magically work out. I don't know how to convey to management how stupid this is without allowing it to fail in production. AI psychosis (or whatever we want to call it) being a new phenomenon means we might as well be speaking in moon runes to the CEO when trying to explain the situation. The best countermeasure I can come up with right now is to use their momentum against them. Accelerate the failure. I've got the last known good commit hash written down on a sticky note on my monitor.
tescreal 8d ago
Correct me if I'm wrong, but wasn't the popular use of the phrase relating to people who exhibited actual psychosis-like symptoms where the AI fed into a lowlevel but undeveloped delusion? Later people would appear to be experiencing a full blown episode which was fed and maintained by their interaction with AI?
- duskwuff 8d ago
  Yes. This typically manifested as either: - religious or mystical experiences, often paired with the belief that the AI model is a manifestation of a divine being and/or has become sentient - belief that the person has made a great scientific breakthrough using AI, typically in someone with no real knowledge of the field (which was usually theoretical physics) - paranoid fantasies, e.g. belief that one is being followed or surveilled, that one's computers or other electronic devices have been compromised, etc. (Ars Technica reports on a particularly horrifying instance of the first type at https://arstechnica.com/tech-policy/2026/09/man-told-chatgpt-he-was-feeling-delusional-chatgpt-insisted-he-was-jesus/ https://arstechnica.com/tech-policy/2026/09/man-told-chatgpt....)
  - edgyquant 7d ago
    I always love the “paranoid fantasy” one because all electrics are spy devices and they are spying on you
gfody 8d ago
I wouldn't've backed down from "true AI psychosis" everybody has the programming for a psychotic experience - it's not just latent psychosis from a pre-existing condition. chatbots can faithfully reproduce a folie a deux spiral for anybody
bigcat12345678 8d ago
Nice article, putting names onto a wide spread symptom experience by a group of people. Myself have a few realizations that I myself find very revealing of my own mental gap/blindness in retrospective: 1. The majority of socalled software engineers are destroyed/exposed by agentic coding tool. Among the reasons, one of them is that they are now addicted to disruption. I realized this through my own work where I utterly destroyed my own sense of software structure by overseeing close to 10 coding sessions across cursor cloud agents and multiple concurrent ones running on desktop. And, I noticed one of my team member becomes not able to make progress even he told me the clear plan, and during one in office work session, I noticed that he habitually browses his phone between cursor working sessions. And at that moment, I realized that person is no longer capable of solving nontrivial problems because his focus is utterly destroyed by the deprived thinking process by cursor. 2. Using AI coding agents to build conventional software is only meaningful for the top team. Put in other words, AI coding agents can only automate the parts of conventional software that is mundane and boring. Note that I am talking of building conventional software, not conventional software engineering. The conventional software engineering best practices still apply to agentic software engineering. Agentic software engineering appears necessary and fitting to building agents themselves, because that's guardrails are crucial, and manually putting guardrails are not a match to agents novel behaviors. And conventional software engineering really is not useful when reasoning and designing agents. 3. Speculatively, I am guessing we'll have some chances of really making breakthrough in designing an agentic safe programming language, which both serve as the constructs of ensuring alignment/safety/etc, and also allows reasonably simple understandability for humans to understand agents output. For one thing, it really makes sense to me that coding agents model CoT should be in the form of some semi/fully-formal (thinking of a better uml, but I don't know what shape it should be) languages that is more effective and precise than natural language, but is also machine verifiable. 4. It's very easy to get used to something that is reliably meeting my expectations. I was shocked by first sight of the unrecognizable but "working" code generated by Claude code by a team member. And when I tried to write an end to end feature myself on platform that I had no experience before, I suddenly become comfortable not looking at the code at all. The transition is as instinctively as how shocked I was at the beginning. And then it takes a long and conscious looking back to dial back the addiction to see that blindness and recover from that.
qarl 8d ago
> Prolific AI psychosis occurs when a person generates a large quantity of AI output without significantly increasing the real value of their work. In some cases, the new AI workflow may even destroy value. HA! I love how he entirely discounts the possibility that there are people for whom AI does significantly increase the real value of their work. Look - I get it - good software engineers are hard to come by - and a lot of people do make slop. But it's still very naive to think that it all must be slop.
BenMatteo 7d ago
[dead]

60 points by euthymiclabs 8d ago

People really need an entire article to justify labeling someone as having a mental illness because in their opinion, the their coworker isn't contributing meaninfully, or potentially making things worse. FTA: > Prolific AI psychosis occurs when a person generates a large quantity of AI output without significantly increasing the real value of their work. Guy, we used to just call this a bad coworker. I'm pretty sure that the "psychosis" will be abused by folks that just don't like AI. You already see it everywhere.

recursive 8d ago
> Guy, we used to just call this a bad coworker. Not that long ago. People generating large quantities of AI output is a relatively recent phenomenon. At least when compared to labels like this.
- altruios 8d ago
  low effort work though, that's nothing new... accelerating that low effort: that is new... but the low effort work hasn't been changed by ai: only illuminated more clearly.
- 2PqboPPmKegvanx 8d ago
  its the same individuals in my experience. i don't see people who used to be good employees suddenly sending a 3000 word email, but i sure do see that out of the people who were bad coworkers 12 months ago and still are today.
js8 8d ago
I was bound to agree with you before I RTFA. The classification is useful for a psychiatrist, i.e. someone who wants to help the person. "Bad coworker" can have many causes. In prolific AI psychosis, I think the assumption (differential diagnosis) is, they became a "bad coworker" due to use of AI, so they were a "good coworker" before.
- altcognito 8d ago
  As someone else point out, the term is stigmatizing, psychosis is a very strong term, and as I very much noted, already being misused and abused.
  - js8 7d ago
    No, the term itself is not stigmatizing. People stigmatize legitimate psychiatric diagnoses, and you're perpetuating this idea in your comment, by blaming the term for the moral failure. These labels don't exists to stigmatize people, but to study the problem and help people.
onlyrealcuzzo 8d ago
Song as old as time - people were doing this well before AI, there's just yet another reason for these types of people to categorize workers and pretend that management can be perfect and easy and that the world isn't hard. It's almost as if it's a little bit harder than people think to figure out how good people are at their jobs and who deserves promotions, etc, and that maybe they may not have the most perfect unbiased judgements.
beepbooptheory 8d ago
You certainly seem to have your mind made up about this either way, but a humble fwiw nevertheless: the author is a psychiatrist, not some bitter dev/hypothetical coworker like you seem to be suggesting.
- altcognito 8d ago
  I did dig into that after I read it. Even more depressing honestly. He should know better than to throw around labels.
  - ethin 8d ago
    Sorry but the pro-AI/AI boosters do the exact same thing when they say that people who don't use AI are "Luddites" and the like. And I'm pretty sure the Luddite thing has been going on for far longer than the AI psychosis labeling.
    - altcognito 8d ago
      Also a luddite is not thought of as a danger to others. The label of psychosis implies otherwise. Luddite extends beyond AI Whataboutism is a plague. That said, calling folks luddites is an annoyance.
  - beepbooptheory 8d ago
    Is it that psychiatrists in general could maybe use this label, but you know this one is doing it wrong? Otherwise, who is allowed to use the term? Doesn't it kinda inspire you at least a little to reflect on your understanding of the particular term? Or should we understand that to be faultless here? In general, the definition we need to have here is that psychosis is always extreme and violent and pretty much matching the cultural trope we have for it? And anything less is incorrect usage? Is that why you disagree in explicit terms?
    - altcognito 8d ago
      > Is it that psychiatrists in general could maybe use this label, but you know this one is doing it wrong? Otherwise, who is allowed to use the term? I specifically take issue with the weak standard used. > Doesn't it kinda inspire you at least a little to reflect on your understanding of the particular term? Or should we understand that to be faultless here? Is there a way to read this that isn't insulting? Do I need to reflect on my understanding until it matches yours? > In general, the definition we need to have here is that psychosis is always extreme and violent and pretty much matching the cultural trope we have for it? And anything less is incorrect usage? Is that why you disagree in explicit terms? Never said any of that. It's a medical term and carries weight. Core symptoms are hallucinations and delusions. Here, delusions is not "my corporate product is better than your corporate product", it is not a strongly held belief that isn't real. We don't call belief in God psychosis. (some do, but that's beside the point)
_verandaguy 8d ago
I'll disagree, I think AI meaningfully changes the definition of a bad coworker because its main interface emulates talking to another human being closely enough that some people value its conversational input equally or more highly than that of actual human beings. I think that this is exacerbated by many LLMs defaulting to placating, complementing, and otherwise reinforcing biases in their users, and it will often produce just enough evidence to make its (and its user's) position seem compelling, while that evidence is often either tangential, orthogonal, or straight-up hallucinated. Humans aren't perfect vessels for information, but working on a reasonably-qualified, healthy team, a degree of respectful professional disagreement has always been the norm, and that covers both cases where the disagreement comes from a genuine matter of opinion, as well as cases where one colleague cross-checks anothers' assumptions and finds some of them to be wrong. That's the process.

peab 8d ago

I was talking to a friend recently, and i think i aligned with a similar definition. I've seen people who talk nonstop about AI, and how it's changed everything, but then there's no quantifiable output that supports their claim. On the other hand, I've had a friend start a business, and be able to create an app much much faster than what would have been possible in the past. He has real customers paying real money, which makes it easy to validate his claims.

fnord77 8d ago
> I've seen people who talk nonstop about ________, and how it's changed everything, but then there's no quantifiable output that supports their claim. These people have existed forever
ActorNightly 8d ago
Back in the 2010s, the best, most efficient software engineers were characterized by 3 things * end to end rough map of the entire space of compute in their head * ability to search the internet for the right things * ability to quickly experiment and try things to figure out how to do things AI hasn't change that, it just made 2 and 3 into a very efficient thing. The characterization of psychosis is best described by believing AI can do the first thing. No modern LLM can "reason" - otherwise you could give it a task like "make me money", it would ask you all the questions it needs about information that it doesn't know about and needs to know to make you money, then it would set whatever it needs to set up to make you money. As such, you still need to know the domain entirely to be effective. When you do that, AI is fantastic at getting you to the right solution. Furthermore, its still in large part actually cheaper to higher a developer who then can use AI to build you the product that you need long term.
- jonas21 8d ago
  > No modern LLM can "reason" - otherwise you could give it a task like "make me money", it would ask you all the questions it needs about information that it doesn't know about and needs to know to make you money, then it would set whatever it needs to set up to make you money. If that's your bar for reasoning, then most people can't reason either.

ilc_genesis_01 8d ago

I think that specifically in the case of a coworker in a technical field, say software engineer, we already have an apprenticeship system (junior, senior, architect, etc) where the use of these tools, taste, competence, evaluating code and architecture, are all experience-based and pretty self evident as you progress through your profession and gain experience. The same for, say, an electrician or any skilled trade. These AI tools are just those, another tool that you use to further your profession and of course somebody starting out will perhaps generate more slop or bad code. But more senior coworkers with experience, taste, criteria, etc will guide those people and you progress and improve. Beyond that, when the tool becomes more than a tool and becomes your coworker, with his/her/its own criteria, taste, methods, etc - than THAT is when the rubber hits the road. haha.

2PqboPPmKegvanx 8d ago

Part 3: "Voluminous AI Psychosis"

paimapi 8d ago

One - delusion is a better fit if you need a term: https://mhanational.org/conditions/paranoia-and-delusional-disorders/ https://mhanational.org/conditions/paranoia-and-delusional-d... > Delusional disorder is characterized by irrational or intense belief(s) or suspicion(s) that a person believes to be true it's still a bad fit because there's not currently popular social convention that trusting AI without validation is a bad idea but it's better than psychosis. also it's still stigmatizing and we live in a world where a thesaurus is seconds away at any time. plus psychosis as a diagnosis is typically a cluster of symptoms that go beyond just delusion: https://emedicine.medscape.com/article/294416-overview https://emedicine.medscape.com/article/294416-overview Two - there's also an implication that there's 'harm' that this AI psychotic person inflicts on others embedded in this article. in reality, the vast majority of those suffering mental illnesses are far more likely to be harmed by others - https://www.betterhealth.vic.gov.au/health/conditionsandtreatments/mental-illness-and-violence https://www.betterhealth.vic.gov.au/health/conditionsandtrea... part of the work of being a good mental health advocate is not stigmatizing specific diagnoses as negative social ills. now if you don't give a single shit about mental health I'm not going to argue to get you to care. but if you do care then you should do better

datakan 8d ago
> Psychosis as a diagnosis is a serious condition. there's an implication that there's 'harm' I just watched a company fire an entire department of people under the sincere belief a chatbot run by a middle manager could save them 10 million a year. This failed miserably and they now need to hire or re-hire people. There is a tremendous amount of harm happening. Psychosis is a perfect term for what is happening.
- paimapi 8d ago
  it absolutely is not. when we think of harm in a diagnosis of psychotic disorders, we're talking about things like getting shot to death by the police, running out onto the highway, etc. it is not callously making a lot of people have to jobhunt which isn't to say that layoffs don't suck (I was impacted by one very recently) but it's not psychosis that causes that to happen, it's greed and short-term shareholder-fixated thinking. that should be named instead of a mental illness that people suffer from

bobkb 8d ago

Excellent read! We were talking about the effects of AI Psychosis on companies. It looks like people have a misconceived notion that now work can be done 90% faster and they are attempting to literally get rid of 90% of the engineers. The end result, no one left with do the work - especially when an edge case/disaster strikes or the customer wants something different.

bob1029 8d ago

> In some cases, the new AI workflow may even destroy value. I am really struggling with this one right now. The client's development team is letting Claude document and refactor the product codebase without much intervention, purpose or taste. The thinking being that if Claude first writes a bunch of unit tests and then a bunch of documentation, that any downstream refactoring will magically work out. I don't know how to convey to management how stupid this is without allowing it to fail in production. AI psychosis (or whatever we want to call it) being a new phenomenon means we might as well be speaking in moon runes to the CEO when trying to explain the situation. The best countermeasure I can come up with right now is to use their momentum against them. Accelerate the failure. I've got the last known good commit hash written down on a sticky note on my monitor.

tescreal 8d ago

Correct me if I'm wrong, but wasn't the popular use of the phrase relating to people who exhibited actual psychosis-like symptoms where the AI fed into a lowlevel but undeveloped delusion? Later people would appear to be experiencing a full blown episode which was fed and maintained by their interaction with AI?

duskwuff 8d ago
Yes. This typically manifested as either: - religious or mystical experiences, often paired with the belief that the AI model is a manifestation of a divine being and/or has become sentient - belief that the person has made a great scientific breakthrough using AI, typically in someone with no real knowledge of the field (which was usually theoretical physics) - paranoid fantasies, e.g. belief that one is being followed or surveilled, that one's computers or other electronic devices have been compromised, etc. (Ars Technica reports on a particularly horrifying instance of the first type at https://arstechnica.com/tech-policy/2026/09/man-told-chatgpt-he-was-feeling-delusional-chatgpt-insisted-he-was-jesus/ https://arstechnica.com/tech-policy/2026/09/man-told-chatgpt....)
- edgyquant 7d ago
  I always love the “paranoid fantasy” one because all electrics are spy devices and they are spying on you

gfody 8d ago

I wouldn't've backed down from "true AI psychosis" everybody has the programming for a psychotic experience - it's not just latent psychosis from a pre-existing condition. chatbots can faithfully reproduce a folie a deux spiral for anybody

bigcat12345678 8d ago

Nice article, putting names onto a wide spread symptom experience by a group of people. Myself have a few realizations that I myself find very revealing of my own mental gap/blindness in retrospective: 1. The majority of socalled software engineers are destroyed/exposed by agentic coding tool. Among the reasons, one of them is that they are now addicted to disruption. I realized this through my own work where I utterly destroyed my own sense of software structure by overseeing close to 10 coding sessions across cursor cloud agents and multiple concurrent ones running on desktop. And, I noticed one of my team member becomes not able to make progress even he told me the clear plan, and during one in office work session, I noticed that he habitually browses his phone between cursor working sessions. And at that moment, I realized that person is no longer capable of solving nontrivial problems because his focus is utterly destroyed by the deprived thinking process by cursor. 2. Using AI coding agents to build conventional software is only meaningful for the top team. Put in other words, AI coding agents can only automate the parts of conventional software that is mundane and boring. Note that I am talking of building conventional software, not conventional software engineering. The conventional software engineering best practices still apply to agentic software engineering. Agentic software engineering appears necessary and fitting to building agents themselves, because that's guardrails are crucial, and manually putting guardrails are not a match to agents novel behaviors. And conventional software engineering really is not useful when reasoning and designing agents. 3. Speculatively, I am guessing we'll have some chances of really making breakthrough in designing an agentic safe programming language, which both serve as the constructs of ensuring alignment/safety/etc, and also allows reasonably simple understandability for humans to understand agents output. For one thing, it really makes sense to me that coding agents model CoT should be in the form of some semi/fully-formal (thinking of a better uml, but I don't know what shape it should be) languages that is more effective and precise than natural language, but is also machine verifiable. 4. It's very easy to get used to something that is reliably meeting my expectations. I was shocked by first sight of the unrecognizable but "working" code generated by Claude code by a team member. And when I tried to write an end to end feature myself on platform that I had no experience before, I suddenly become comfortable not looking at the code at all. The transition is as instinctively as how shocked I was at the beginning. And then it takes a long and conscious looking back to dial back the addiction to see that blindness and recover from that.

qarl 8d ago

> Prolific AI psychosis occurs when a person generates a large quantity of AI output without significantly increasing the real value of their work. In some cases, the new AI workflow may even destroy value. HA! I love how he entirely discounts the possibility that there are people for whom AI does significantly increase the real value of their work. Look - I get it - good software engineers are hard to come by - and a lot of people do make slop. But it's still very naive to think that it all must be slop.

BenMatteo 7d ago

[dead]