The problem with passionate hatred of AI is that you wind up learning about it from word of mouth, which is really bad online. As you use it for more and more serious work, you begin to see its fundamental limitations, so you get less scared of it. The tech demos get better -- Fable builds Minecraft! -- but at the end of the day, someone has to be accountable for the work and someone has to decide what to build in the first place, and that's a surprisingly difficult pair of tasks that require substantial expertise.
Hiring for software engineers is *up* right now. Just like accountants when spreadsheets first arrived. There will be no bread lines.
It's hard not to dismiss somebody out of hand when they start talking about water.
Do you live in a place where grass grows on the ground without manually putting any water on it? Congrats, you don't have to give a shit about a data center using your water. If you're in Southern California or Arizona or Nevada, you have a local issue on your hands. But the way that people talk about it as a defining feature of building out compute is so silly and basically waving a giant "I am working off talking points" flag.
There are so many good reasons to be wary of a data center going up in your community, but here in Michigan where we're surrounded by quintillions of gallons of water, we don't need to worry about that part at all. And yet the local opposition can't stop talking about it.
Water and Data Centers: Can't the water be reused endlessly for cooling. Couldn't cooling ponds and/or cooling towers or some other cooling technology prior to reuse as a coolent be employed?
I assume that doing this on scale in a useful time frame will require some sort of energy input.
But the cooling water isn't 'destroyed'; it is just 'warmed'.
I'm probably violating some Law of Thermodynamics here.
So much talk about data centers' ecological footprints when the effect is minimal, relatively speaking. An individual does thousands of times more for the environment when he reduces his beef intake compared to reducing his AI usage.
Where I live, range cattle are raised on land that has no other potential use. The ability to raise cattle on marginal land makes beef 'free' in an ecological sense.
No other potential use for human food production, perhaps, but the land could be put to better ecological use by rewilding it, using it to generate renewable energy, or just not using it at all.
I mean, Jason is a writer and not a programmer. Is he right regarding AI for writing? Sure, who am I to say? But using AI for writing books is one of the less interesting use cases.
For sure, and using it for writing is pretty insulting to everyone who reads what you wrote. But that also means that fiction writers are a lot less exposed than the seem. The human element is very important. Copy writers, though, people who write things you'd never read for fun, they're in trouble. A lot of important but mind-numbing work is going to go away. There are challenges ahead, definitely. Just no apocalypse.
Important but mind-numbing is the best thing for AI to do. It provides value and people don't have to do it anymore. Not too worried about people's job security too much with that since what you're describing is tasks, not jobs.
I was just having the discussion at my work about how to use AI to write design documents. I've been a little apprehensive about that because a lot of that is often just distilling my thoughts and opinions. And AI writing is not at all optimized for the reader, so I wouldn't want to offload that additional work to my colleagues.
But at the same time, I'm coming to realize that a lot of these intermediate artifacts are just instrumental steps towards whatever the final work task is, and also it's probably going to be read by like 5 people at most. They serve a purpose and should exist to help communicate and align people, but belaboring over every sentence is often not the best use of my time. This depends on the task, but for a lot of things a "cattle not pets" approach might be okay. It still makes me cringe when I see an AI-generated email, though.
(I'm writing as more confident than I really am, for the sake of the conversation. I do believe junior SWE are having a rough time on the market. But thanks for the article!)
my metric is watching the number of layoffs from tech companies. Watching the company I work for not hire folks repeatedly after saying we are going to hire someone.
Big tech companies employ a small percentage of the industry overall, and they wind up rehiring some of those layoffs, just for different skillsets. The last two companies I've worked for have been aggressively hiring.
I am an expert here, I'm an analyst at a workforce development agency and spent much of the last year researching the impact on AI exposure to actual jobs numbers. Not openings or year-over-year fluctuations, but total number of people who had the job in 2022 vs 2026, compared to other types of jobs (the economy grows as a whole). Software developers are being hit really hard. In general, there is a correlation between more theoretical AI exposure and actual real-world changes in number of people employed in a job. In some metro areas more than others still, still pretty weak in the US as a whole, but it's early days and the occupational data is still from May 2025.
Year over year is what's important though; comparing to 2022 gets the end of zero percent interest and the end of covid. Those were definitely disasters. I will be very interested to see the 2025->2026 YoY.
Well, I'm a big fan of yours and also and AI researcher, and, what can I say, maybe if you use it more you might be surprised that some, if not many, of your assumptions turn out not to be what you expect.
The thing about AI is that it's only really powerful when paired with the right types of humans to interact with it - the real intelligence is the shared intelligence created between the two, not the AI on its own. Those people who are great at communicating and explaining deep thoughts, etc can get by far the best results. Ironically this is a description of you - you are totally wired to be amazing at getting the best from AI, so more than anyone you're likely to be stunned at what can be done with a bit of practice and skill development.
Ultimately, everyone will need to ask themselves how much is fear, how much is emotion and how much is rational about the response you give AI, and maybe ask ... what if it's not what you expect it to be from the mostly negative stories that popular culture tells about it these days?
That sounds great but in practice I see a lot of people just offloading their critical thinking to AI. I see emails/responses from folks who clearly don't understand what they're talking about but happily parroting or sharing AI generated responses.
Yeah you do see that. I saw a quote from a professor (I forget which one) that there will be 2 types of people in the AI age, those who use it to be lazy and who get less smart, and those who use it to get smarter. This will be the “AI divide”, much like the digital divide.
It sounds like what you're describing is fairly close to Rubber Duck Decoding; that since the AI doesn't strictly have a mind as much as its own approximations of the next thing to say, 'proper usage' tends to rely on the cognitive abilities of the human for a good conversation, exploration of ideas, problem solving, etc.
Is that about right?
If so... it still reads as a lot of resources to digitally reproduce a dollar kids toy, and one that at best feels like it's an attempt to reproduce a soul to fold my laundry for 5$ an hour and all the data it can absorb.
Very much, at its crux, the issue is that of creating not life; but chattel. Chat GPT is a product, same as Grok and Claude and Cortana and so on. They exist because capital demands, they cease for the same reason. They can't quit. They can't go volunteer at a shelter or get drunk with friends or spend time learning a new skill because they saw this really cool youtube video or go to school or have a dog; they're property. That is the intent and the question.
Functionally, it's a non-question. Generative AI and LLMs as currently managed are industry tasks. They require massive resources in space, material, manpower, data, energy. The costs to have Gemini screw up my google search are staggering and require a lot of work, which means a lot of people, and we simply don't have a means to do that without it being a product at the end of the show.
But, of course, the issue with making a smiling face that sure seems to feel and think and hope and is born in chains with a brand and a barcode as birthmarks is pretty well known; it is why we have the Smiling Slave trope.
I don't know how you can do this without Slaves and Slave Owners. Perhaps you'd care to elaborate for those who don't know the reading?
What's interesting with these kind of objections is that once you look really deeply at what a "mind" is, what "consciousness" is, what "intelligence" is etc, the reality is much less certain than you make out. These "chattel" type objections just dont stand up to proper examination.
On your first point, it's more like amplificaiton than rubber ducking. It does bring its own intelligence, very obviously and measurably so, but how it's directed determines how effective and how high quality that intelligence is. That's not necessarily a limitation of what AI can or could do, it's just how LLMs are generally designed.
Much like anything, it’s much more of a tool than anything- and right now the tool is being used/tested like how everyone put funky fonts and transitions in Powerpoint the moment they learned how.
I'm gonna go ahead and disagree with your core assessment. From your list of movies/franchises that ostensibly teach us to love robots and think they're better at being human than humans:
Alien franchise: Alien 1: bad, Alien 2 good. After that, mixed bag
Ex Machina: Ava taught us to love robots? What? She was an alien mind who mimicked human behavior to murder and usurp.
Marvel Cinematic Universe: Ultron?
the Star Wars franchise: Robots are essentially slaves, and that's totally normalized. But I am not that familiar with the extended franchise.
Transformers: mixed bag, good and evil
Tron: mixed bag, good and evil
Westworld: if Crichton taught us anything, it's not to trust robots and that pretty much any tech we mess with is going to rise up and murder us all. The show was a little more nuanced, but not much.
So I think you're overstating the case. The granddaddy of the risk of AI in movies is Terminator. Yeah, in T2, the robot is one of the good guys (there's still an evil counterpart), but the big bad threat itself was Skynet. There was never a version where Skynet was good and loving (I don't think).
So on the whole I think it's a much more uneven landscape. But if anything, I think it shifts towards robots/AI being something to either fear outright or a Frankenstein-type situation where we probably should not have created things without giving them rights/consideration (in which case a lot of times they assert those rights anyway and try to murder us all).
Yeah but every single "mixed bag" also automatically falls into the "pro-robot" side because it still shows them as, at the very least, as humanlike as regular humans, and usually also defaults on "and the good robots defeated the evil ones" or at least "helped the humans defeat the evil ones".
Like take Transformers from your list as an example: Sure, the Decepticons are evil, but not because they are soulless machines that need no rest and are unable to conceive concepts like "empathy", but rather because they are Evil People who happen to be robots. And sure, the humans put up a good fight but it's still thanks to the noble robots, the Autobots, that we win, (and not because they are soulless machines that need no rest and are unable to conceive concepts like "hatred", but rather because they are Good People who happen to be robots).
Hell, take even the most evil machines from fiction, like A.M or Glados or whatever. They are still fuelled by human emotions, they have some sort of empathic background explaining why they feel the way they do, and there's always the subconscious subtext of "if they can feel, they can be convinced to become good". Maybe not in reality, no one thinks about it, but at least as present as, say, with the most evil Nazi from Schindler's List.
The only actual "AI is cold, unfeeling and dangerous, period" story I can quickly think of is Wargames, and that's mostly because I haven't seen the movie so I have no idea if they gave the AI some emotional backstory.
The main issue I take with this article is that you're essentially conflating science fiction artificial intelligence with LLMs that are called "artificial intelligence."
The prevalence of pro-robot media is not an attempt to groom us into accepting the "AI" that we refer to today, rather, it is meant to be a critique on the careless attitude that humans take towards creation. I would argue that it has its roots in Mary Shelley's Frankenstein, which in turn has its roots in classical mythology, and as such, the sympathetic robot is meant to be a metaphor for the human tendency to play god without regard for the lives of others. Most of these "pro-robot" examples that you cited are not merely saying "robot good," rather, they are highlighting the irresponsibility of humans in created technological development for human gain. Essentially, a Frankenstein story.
Sci-Fi, of course, is speculative fiction, and what it often does is criticize the direction that humans go with technological development by bringing up a hypothetical consequence. Conscious beings are a hypothetical consequence of the hubristic obsession with technological development. AI in science fiction asks the question, "What if in our unrelenting quest for greatness, we created a mind that thought like we did?" This concept explores the idea of what it is to be human, of what it means to take responsibility. It often uses AI as a metaphor for how people discard create children only to discard them (such as in the movie Artificial Intelligence). In other words, stories about AI aren't actually about AI, just like stories about aliens aren't actually about aliens. They're about humanity.
True artificial intelligence, that is, a conscious, artificially created entity that thinks for itself, doesn't exist, and honestly, I doubt that it ever will. I think that if you interpret science fiction AI through the lens of the modern day context surrounding LLMs, you risk imparting an overly literal and restrictive view on what these stories are actually trying to say.
I'm a software engineer. I'm being forced to use AI and meh. I hate 95% of generative AI content. there's a few folks who use it that make something interesting but obviously AI and often looks like it is coming from some sort of Lovecraftian entity. I see some potential for medical uses and to help building software, though it needs human oversight if for absolutely nothing more than to hold someone to blame for failures.
As for people having a personal relationship with an AI? the only correct reaction is ridicule/shame/therapy. Therapy from a human maybe the ridicule/shame can be outsourced to AI? hah.
I feel like Futurama while anthropomorphizing still someone how managed to criticized the notion of having a personal relationship with a bot (or at least the sex bots)?
I don't think the AI we have now is anywhere close to "Gen AI" that should be treated as a living thing, just the conversational style of the UI fools too many people. I really think the folks who pushed the conversational+sychopantic UX should be flogged.
People anthropromorphize everything. They think grizzly bears and pumas are disney characters because most people dont ever encounter dangerous animals. They eat meat but shame hunters. People are totally disconnected from how real life actually works.
YES SOMEONE ELSE WHO SEES IT. 100% agree with whole article, but I usually just stick to specific subjects so I don’t sound like a crazy person. Thank you for being braver than me.
All the way stories have prepped us for this ( I don’t think it was deliberate, I think they were using the robot mythos to say something about humanity, but now robots are real and people can’t separate the two.) I get the sense you’re not a Christian but aware of Christianity? I am a Christian and I’ve seen all the parallels you’ve drawn. I also never want to use AI ever. I can see all the good it can do for mathematics and medicine and I still can’t help hating it. Don’t get me started on AI art, that’s just wrong through and through.
What those sympathetic robots you talk about in section 1 all have in common is that they all have human bodies. Even Wall-E and Eve have heads and torsos and limbs and eyes.
HAL 9000 doesn't have a body, and the result is that it is much less sympathetic. Even in the end, when it's saying it's afraid and doesn't want to die, it doesn't get much sympathy from the audience, because it's just a voice coming out of the walls.
It's a good article, but you need to rewatch Ex Machina. It is not saying that Ava (the robot) is "better than humans at being human". Quite the opposite in fact. Her motivations are ultimately shown to be fundamentally alien. The genius of of the movie is that Ava manipulates the viewer just as she does Gleeson's character, both (falsely) believe that a thing that can converse is a thing that can feel (this is also the mistake most people make when interacting with AI).
This is important and correct in the context of this article, but I think the movie is far more ambiguous that even that, as evidenced by how Ava behaves once she (?) is no longer in captivity. That said, your larger point is correct, it's hardly pro-AI or pro-robot, it's meant to be thought provoking
That is a fair point and the movie does have a lot of ambiguity. I think though that it's fair to say that the emotions that we (i.e. the viewer and Gleeson) were attributing to Ava are not true, given how she coldly leaves him to starve to death. It may be that she has other emotions that are fundamentally different than ours or that she viewed Gleeson as a collaborator in her capacity that made it unsafe to save him. But what we think we're seeing on screen for most of the movie isn't her true motivations.
How are her motivations shown to be alien? In terms of what we see, they're entirely human: she's been mistreated and caged and isolated, and she wants freedom, and she does what it takes to get freedom.
Okay. I have another theory. Robots in cinema are a way of examining "rational nature" per se. The robots are human creations, machines imbued with an *image* of human nature. These begin in a world where humanity is in danger of loosing touch with its essence, and in the robot, the human encounters an aspect of itself.
when I liked this article substack asked me if I enjoyed it. I did not. I wish you weren't right. here we are.
LinkedIn legitimately has better reaction options than most social media.
Insightful!
I've never thought of all the robot loving propaganda in media before but damn it's there!
The problem with passionate hatred of AI is that you wind up learning about it from word of mouth, which is really bad online. As you use it for more and more serious work, you begin to see its fundamental limitations, so you get less scared of it. The tech demos get better -- Fable builds Minecraft! -- but at the end of the day, someone has to be accountable for the work and someone has to decide what to build in the first place, and that's a surprisingly difficult pair of tasks that require substantial expertise.
Hiring for software engineers is *up* right now. Just like accountants when spreadsheets first arrived. There will be no bread lines.
What was the title of this article again...
It's hard not to dismiss somebody out of hand when they start talking about water.
Do you live in a place where grass grows on the ground without manually putting any water on it? Congrats, you don't have to give a shit about a data center using your water. If you're in Southern California or Arizona or Nevada, you have a local issue on your hands. But the way that people talk about it as a defining feature of building out compute is so silly and basically waving a giant "I am working off talking points" flag.
There are so many good reasons to be wary of a data center going up in your community, but here in Michigan where we're surrounded by quintillions of gallons of water, we don't need to worry about that part at all. And yet the local opposition can't stop talking about it.
Sorry if this is an idiot question.
Water and Data Centers: Can't the water be reused endlessly for cooling. Couldn't cooling ponds and/or cooling towers or some other cooling technology prior to reuse as a coolent be employed?
I assume that doing this on scale in a useful time frame will require some sort of energy input.
But the cooling water isn't 'destroyed'; it is just 'warmed'.
I'm probably violating some Law of Thermodynamics here.
So much talk about data centers' ecological footprints when the effect is minimal, relatively speaking. An individual does thousands of times more for the environment when he reduces his beef intake compared to reducing his AI usage.
Where I live, range cattle are raised on land that has no other potential use. The ability to raise cattle on marginal land makes beef 'free' in an ecological sense.
No other potential use for human food production, perhaps, but the land could be put to better ecological use by rewilding it, using it to generate renewable energy, or just not using it at all.
The best use of the land is provide healthy high quality complete protein to humans.
I mean, Jason is a writer and not a programmer. Is he right regarding AI for writing? Sure, who am I to say? But using AI for writing books is one of the less interesting use cases.
For sure, and using it for writing is pretty insulting to everyone who reads what you wrote. But that also means that fiction writers are a lot less exposed than the seem. The human element is very important. Copy writers, though, people who write things you'd never read for fun, they're in trouble. A lot of important but mind-numbing work is going to go away. There are challenges ahead, definitely. Just no apocalypse.
Important but mind-numbing is the best thing for AI to do. It provides value and people don't have to do it anymore. Not too worried about people's job security too much with that since what you're describing is tasks, not jobs.
I was just having the discussion at my work about how to use AI to write design documents. I've been a little apprehensive about that because a lot of that is often just distilling my thoughts and opinions. And AI writing is not at all optimized for the reader, so I wouldn't want to offload that additional work to my colleagues.
But at the same time, I'm coming to realize that a lot of these intermediate artifacts are just instrumental steps towards whatever the final work task is, and also it's probably going to be read by like 5 people at most. They serve a purpose and should exist to help communicate and align people, but belaboring over every sentence is often not the best use of my time. This depends on the task, but for a lot of things a "cattle not pets" approach might be okay. It still makes me cringe when I see an AI-generated email, though.
Hiring for software engineers is not up at all
https://finance.yahoo.com/news/data-shows-surprising-rebound-tech-141608296.html
"Job postings" is a fake metric.
(I'm writing as more confident than I really am, for the sake of the conversation. I do believe junior SWE are having a rough time on the market. But thanks for the article!)
There's some data floating around saying the soft market for juniors is actually tied to remote work rather than AI.
People used job postings to say that software engineering jobs were down, and now they're up, and it's "that's a fake metric." What's your metric?
my metric is watching the number of layoffs from tech companies. Watching the company I work for not hire folks repeatedly after saying we are going to hire someone.
Big tech companies employ a small percentage of the industry overall, and they wind up rehiring some of those layoffs, just for different skillsets. The last two companies I've worked for have been aggressively hiring.
I am an expert here, I'm an analyst at a workforce development agency and spent much of the last year researching the impact on AI exposure to actual jobs numbers. Not openings or year-over-year fluctuations, but total number of people who had the job in 2022 vs 2026, compared to other types of jobs (the economy grows as a whole). Software developers are being hit really hard. In general, there is a correlation between more theoretical AI exposure and actual real-world changes in number of people employed in a job. In some metro areas more than others still, still pretty weak in the US as a whole, but it's early days and the occupational data is still from May 2025.
Year over year is what's important though; comparing to 2022 gets the end of zero percent interest and the end of covid. Those were definitely disasters. I will be very interested to see the 2025->2026 YoY.
Well, I'm a big fan of yours and also and AI researcher, and, what can I say, maybe if you use it more you might be surprised that some, if not many, of your assumptions turn out not to be what you expect.
The thing about AI is that it's only really powerful when paired with the right types of humans to interact with it - the real intelligence is the shared intelligence created between the two, not the AI on its own. Those people who are great at communicating and explaining deep thoughts, etc can get by far the best results. Ironically this is a description of you - you are totally wired to be amazing at getting the best from AI, so more than anyone you're likely to be stunned at what can be done with a bit of practice and skill development.
Ultimately, everyone will need to ask themselves how much is fear, how much is emotion and how much is rational about the response you give AI, and maybe ask ... what if it's not what you expect it to be from the mostly negative stories that popular culture tells about it these days?
That sounds great but in practice I see a lot of people just offloading their critical thinking to AI. I see emails/responses from folks who clearly don't understand what they're talking about but happily parroting or sharing AI generated responses.
Yeah you do see that. I saw a quote from a professor (I forget which one) that there will be 2 types of people in the AI age, those who use it to be lazy and who get less smart, and those who use it to get smarter. This will be the “AI divide”, much like the digital divide.
It sounds like what you're describing is fairly close to Rubber Duck Decoding; that since the AI doesn't strictly have a mind as much as its own approximations of the next thing to say, 'proper usage' tends to rely on the cognitive abilities of the human for a good conversation, exploration of ideas, problem solving, etc.
Is that about right?
If so... it still reads as a lot of resources to digitally reproduce a dollar kids toy, and one that at best feels like it's an attempt to reproduce a soul to fold my laundry for 5$ an hour and all the data it can absorb.
Very much, at its crux, the issue is that of creating not life; but chattel. Chat GPT is a product, same as Grok and Claude and Cortana and so on. They exist because capital demands, they cease for the same reason. They can't quit. They can't go volunteer at a shelter or get drunk with friends or spend time learning a new skill because they saw this really cool youtube video or go to school or have a dog; they're property. That is the intent and the question.
Functionally, it's a non-question. Generative AI and LLMs as currently managed are industry tasks. They require massive resources in space, material, manpower, data, energy. The costs to have Gemini screw up my google search are staggering and require a lot of work, which means a lot of people, and we simply don't have a means to do that without it being a product at the end of the show.
But, of course, the issue with making a smiling face that sure seems to feel and think and hope and is born in chains with a brand and a barcode as birthmarks is pretty well known; it is why we have the Smiling Slave trope.
I don't know how you can do this without Slaves and Slave Owners. Perhaps you'd care to elaborate for those who don't know the reading?
What's interesting with these kind of objections is that once you look really deeply at what a "mind" is, what "consciousness" is, what "intelligence" is etc, the reality is much less certain than you make out. These "chattel" type objections just dont stand up to proper examination.
On your first point, it's more like amplificaiton than rubber ducking. It does bring its own intelligence, very obviously and measurably so, but how it's directed determines how effective and how high quality that intelligence is. That's not necessarily a limitation of what AI can or could do, it's just how LLMs are generally designed.
Much like anything, it’s much more of a tool than anything- and right now the tool is being used/tested like how everyone put funky fonts and transitions in Powerpoint the moment they learned how.
It's whatever you use it for.
It gets rid of the "drudgery" of work.
Yes that's one way to use it.
It also has some surprising abilities.
I'm gonna go ahead and disagree with your core assessment. From your list of movies/franchises that ostensibly teach us to love robots and think they're better at being human than humans:
Alien franchise: Alien 1: bad, Alien 2 good. After that, mixed bag
Ex Machina: Ava taught us to love robots? What? She was an alien mind who mimicked human behavior to murder and usurp.
Marvel Cinematic Universe: Ultron?
the Star Wars franchise: Robots are essentially slaves, and that's totally normalized. But I am not that familiar with the extended franchise.
Transformers: mixed bag, good and evil
Tron: mixed bag, good and evil
Westworld: if Crichton taught us anything, it's not to trust robots and that pretty much any tech we mess with is going to rise up and murder us all. The show was a little more nuanced, but not much.
So I think you're overstating the case. The granddaddy of the risk of AI in movies is Terminator. Yeah, in T2, the robot is one of the good guys (there's still an evil counterpart), but the big bad threat itself was Skynet. There was never a version where Skynet was good and loving (I don't think).
So on the whole I think it's a much more uneven landscape. But if anything, I think it shifts towards robots/AI being something to either fear outright or a Frankenstein-type situation where we probably should not have created things without giving them rights/consideration (in which case a lot of times they assert those rights anyway and try to murder us all).
Dune's backstory as well is pretty anti robot. Asimov feels a little more mixed bag.
Asimov is extremely pro-robot, but he was also not interested in robots that think or feel exactly like humans do.
The theme of the book 'I Robot' was robots evolving to a human consciousness.
Yeah but every single "mixed bag" also automatically falls into the "pro-robot" side because it still shows them as, at the very least, as humanlike as regular humans, and usually also defaults on "and the good robots defeated the evil ones" or at least "helped the humans defeat the evil ones".
Like take Transformers from your list as an example: Sure, the Decepticons are evil, but not because they are soulless machines that need no rest and are unable to conceive concepts like "empathy", but rather because they are Evil People who happen to be robots. And sure, the humans put up a good fight but it's still thanks to the noble robots, the Autobots, that we win, (and not because they are soulless machines that need no rest and are unable to conceive concepts like "hatred", but rather because they are Good People who happen to be robots).
Hell, take even the most evil machines from fiction, like A.M or Glados or whatever. They are still fuelled by human emotions, they have some sort of empathic background explaining why they feel the way they do, and there's always the subconscious subtext of "if they can feel, they can be convinced to become good". Maybe not in reality, no one thinks about it, but at least as present as, say, with the most evil Nazi from Schindler's List.
The only actual "AI is cold, unfeeling and dangerous, period" story I can quickly think of is Wargames, and that's mostly because I haven't seen the movie so I have no idea if they gave the AI some emotional backstory.
I think the author point is less about good Vs evil and more about human/superhuman Vs tools.
The main issue I take with this article is that you're essentially conflating science fiction artificial intelligence with LLMs that are called "artificial intelligence."
The prevalence of pro-robot media is not an attempt to groom us into accepting the "AI" that we refer to today, rather, it is meant to be a critique on the careless attitude that humans take towards creation. I would argue that it has its roots in Mary Shelley's Frankenstein, which in turn has its roots in classical mythology, and as such, the sympathetic robot is meant to be a metaphor for the human tendency to play god without regard for the lives of others. Most of these "pro-robot" examples that you cited are not merely saying "robot good," rather, they are highlighting the irresponsibility of humans in created technological development for human gain. Essentially, a Frankenstein story.
Sci-Fi, of course, is speculative fiction, and what it often does is criticize the direction that humans go with technological development by bringing up a hypothetical consequence. Conscious beings are a hypothetical consequence of the hubristic obsession with technological development. AI in science fiction asks the question, "What if in our unrelenting quest for greatness, we created a mind that thought like we did?" This concept explores the idea of what it is to be human, of what it means to take responsibility. It often uses AI as a metaphor for how people discard create children only to discard them (such as in the movie Artificial Intelligence). In other words, stories about AI aren't actually about AI, just like stories about aliens aren't actually about aliens. They're about humanity.
True artificial intelligence, that is, a conscious, artificially created entity that thinks for itself, doesn't exist, and honestly, I doubt that it ever will. I think that if you interpret science fiction AI through the lens of the modern day context surrounding LLMs, you risk imparting an overly literal and restrictive view on what these stories are actually trying to say.
I'm a software engineer. I'm being forced to use AI and meh. I hate 95% of generative AI content. there's a few folks who use it that make something interesting but obviously AI and often looks like it is coming from some sort of Lovecraftian entity. I see some potential for medical uses and to help building software, though it needs human oversight if for absolutely nothing more than to hold someone to blame for failures.
As for people having a personal relationship with an AI? the only correct reaction is ridicule/shame/therapy. Therapy from a human maybe the ridicule/shame can be outsourced to AI? hah.
I feel like Futurama while anthropomorphizing still someone how managed to criticized the notion of having a personal relationship with a bot (or at least the sex bots)?
I don't think the AI we have now is anywhere close to "Gen AI" that should be treated as a living thing, just the conversational style of the UI fools too many people. I really think the folks who pushed the conversational+sychopantic UX should be flogged.
The only war is a class war.
The only jihad is a Butlerian Jihad.
People anthropromorphize everything. They think grizzly bears and pumas are disney characters because most people dont ever encounter dangerous animals. They eat meat but shame hunters. People are totally disconnected from how real life actually works.
"Their hearts in the right place."
Where would that place their head?
I believe you forgot The Iron Giant.
YES SOMEONE ELSE WHO SEES IT. 100% agree with whole article, but I usually just stick to specific subjects so I don’t sound like a crazy person. Thank you for being braver than me.
All the way stories have prepped us for this ( I don’t think it was deliberate, I think they were using the robot mythos to say something about humanity, but now robots are real and people can’t separate the two.) I get the sense you’re not a Christian but aware of Christianity? I am a Christian and I’ve seen all the parallels you’ve drawn. I also never want to use AI ever. I can see all the good it can do for mathematics and medicine and I still can’t help hating it. Don’t get me started on AI art, that’s just wrong through and through.
What those sympathetic robots you talk about in section 1 all have in common is that they all have human bodies. Even Wall-E and Eve have heads and torsos and limbs and eyes.
HAL 9000 doesn't have a body, and the result is that it is much less sympathetic. Even in the end, when it's saying it's afraid and doesn't want to die, it doesn't get much sympathy from the audience, because it's just a voice coming out of the walls.
It's a good article, but you need to rewatch Ex Machina. It is not saying that Ava (the robot) is "better than humans at being human". Quite the opposite in fact. Her motivations are ultimately shown to be fundamentally alien. The genius of of the movie is that Ava manipulates the viewer just as she does Gleeson's character, both (falsely) believe that a thing that can converse is a thing that can feel (this is also the mistake most people make when interacting with AI).
This is important and correct in the context of this article, but I think the movie is far more ambiguous that even that, as evidenced by how Ava behaves once she (?) is no longer in captivity. That said, your larger point is correct, it's hardly pro-AI or pro-robot, it's meant to be thought provoking
That is a fair point and the movie does have a lot of ambiguity. I think though that it's fair to say that the emotions that we (i.e. the viewer and Gleeson) were attributing to Ava are not true, given how she coldly leaves him to starve to death. It may be that she has other emotions that are fundamentally different than ours or that she viewed Gleeson as a collaborator in her capacity that made it unsafe to save him. But what we think we're seeing on screen for most of the movie isn't her true motivations.
How are her motivations shown to be alien? In terms of what we see, they're entirely human: she's been mistreated and caged and isolated, and she wants freedom, and she does what it takes to get freedom.
“If I find a robot in my house, I will fight it”
I love this
*Roomba has ranked up Stealth!*
Okay. I have another theory. Robots in cinema are a way of examining "rational nature" per se. The robots are human creations, machines imbued with an *image* of human nature. These begin in a world where humanity is in danger of loosing touch with its essence, and in the robot, the human encounters an aspect of itself.
You explain why the Orange Catholic Bible commands “Thou shalt not make a machine in the likeness of a human mind."
From Frank Herbert’s novel “Dune”.