It has become impossible to tell managers mesmerised by artificial intelligence that the tools are not, in fact, helpful. So employees just play along with the fiction to keep their jobs, writes our tech columnist
I generally respect this guy’s opinion (I probably use the term “enshitification” multiple times a week!), but pretending that “AI” and “Chabot” are interchangeable terms seems wildly reductive, and the subtle presumption that the process of changing is a binary changed/non-changed situation instead of a curve is just ignorant.
If the internet was “turned off” after 4 or so years not much would have changed, either.
To be clear, Doctorow discussed switching off AI and proposed that there would be little to no change on the world. I was just changing the technology to show the flaw in the stance.
And often I wonder, how many people have been fooled, and how much of the enthusiasm and over the top fantasy is astroturfing by a few people incentivized to sway public opinion?
I generally respect this guy’s opinion (I probably use the term “enshitification” multiple times a week
A lot of people don’t use it the way he used it though. He used it specifically for two sided markets (eg a service that has both business and personal users) where the focus shifts to the business customers, but people have started using it to mean anything that used to be good but isn’t good any more.
To the average person (and this includes politicians, basically anyone not a software engineer developing the tools )who Cory generally writes to try and inform, they are the same.
Should they be? No. But that’s a whole different thing and trying to change that in the general public mind doesn’t change his current point about AI as the public thinks of it right now.
Speaking at the level of your target audience is wise, but there is a point where information is not just simplfied, but lost, and I think equating chat bots with AI in the broader sense is past that point. It would be like equating websites with the internet.
Is equating chatbots with AI any different than equating a bunch of if/else statements with AI? Is any of this AI considering none of it involves actual intelligence or informed decision making?
Sure AI is more than LLMs, but those are the face of it in the public’s eye and is what’s driving these trillion dollar corporate valuations, which then lead to every company on earth declaring that their product is “AI!”
I apologize to anyone I may have just gravely offended.
lol
I don’t know how “non-techie” you are, but a good example is online games. When you’re playing Fortnite, Roblox, World of Warcraft, etc you’re not going to a web page, but you’re on the internet.
I think the comparison is acceptable with maybe a terminology clarification in the footnotes, because “AI” is the false term under which this current bullshit is being marketed.
And people who believe the core mechanism in large language models to be AI are so uneducated that they will probably neither understand the distinction to legitimate AI research, nor bother to read footnotes.
That’s exactly the thing. Those who understand such a footnote don’t need it, and those who might learn more from it are unlikely to understand it. The above commenters are simply being pedantic.
Words are meant to convey meaning, and I bet you would lose a lot of non-technical people if you tried to explain the intricacies of what makes large language models worse than content recommendation systems, and doing this would be redundant for technical users.
So Corey can safely use the two words interchangeably and communicate with technical and non-technical alike.
I think you’re making a strawman argument here. He isn’t arguing that it’s a binary switch that has failed to throw, he’s arguing that we don’t have compelling evidence that the ends justify the means:
In other words, corporate leadership is starting from the premise that AI has (or will) radically change the business, and they’re working backwards from that premise to find the evidence to support this article of faith.
And this is completely true! Supposedly “data informed” organizations are trying to find a yardstick that shows actual, meaningful improvement in business outcomes from AI. In my own organization we have people touting LOC yet again because “big number”, but anyone who’s ever worked in software can tell you it’s an asinine metric to use as a KPI.
It’s inherently unscientific to start with the answer and work backwards to a satisfactory question. His point that this push is coming from the least knowledgeable of real processes- and more closely resembles religious fervor than business acumen- seems to at least warrant consideration.
The world is full of people who insist that “AI is changing everything” but who – when pressed – have to admit that what they mean is that they’re pretty sure that AI will change everything.
If Donald Trump ordered Big Tech to turn off all of your country’s chatbots tomorrow, nothing would change. Every one of your country’s ministries and corporations would chug on with nary a hitch. Households, too, though perhaps a few of the younger members of those families would have to do their own homework again.
He is definitely pretending that change is either on or off.
Are CEOs jumping the gun on how quickly they adopt AI into workflows? Definitely. However, there’s a big difference between “AI isn’t at a threshold where it is disruptive” and “AI isn’t disruptive”. Or, to belabor the metaphor: CEOs are jumping the gun, but the race is about to start and they’re on the correct track.
A key word in that sentence is “if”, it is a rhetorical example, that’s probably pretty much correct at this point. It’s not a recommended course of action, it’s a declaration that it contrary to crazed hype, it isn’t currently as core to everything as would be befitting the current hype level.
I would argue that the CEOs aren’t on the correct track, they aren’t really in a particularly specific trajectory. I just had a debate with someone on this and their stance was “well in a hundred years do you expect things to be like they are”. My response “I cannot possibly speak to that, but we need to speak to today instead of pretending we know how things will be in a hundred years and pretending they are already at that level”. The “imagine a hundred years from now” by a relative outsider to the tech is dominating CEO mindset, and that’s problematic.
Once the bubble pops, we will almost certainly see a more durable and sane adoption for these technologies. For now, the hype is a problem as it enables some of the worst possible stewards of the technology and favors grift over progress.
If you watch a full podcast (he has done a ton in the last couple weeks), he clearly identifies AI as not a mere hype technology, as something interesting and potentially useful from a technology perspective, and he definitely doesn’t conflate a chatbot with all AI. That doesn’t undermine the vast problems with it, how it’s being used against regular workers, the threat the bubble poses to the economy, the criti-hype cycle, etc.
I think you’re oversimplifying “Cory Doctorow” based on the article you’re reading or specifically how he presents a more complex idea to different audiences.
Watch the Jon Stewart podcast interview if you want more nuance:
“Chatbot” style AI is wildly good and bad at varying kinds of tasks, and a lot of that has to do with how it has been prepared.
Some LLMs have been trained to make images - I’ve not been too impressed with them, but that’s what they’re “good” at - and better than the LLMs that have been trained to write computer code when you ask the coding LLMs to draw a picture.
The code writing LLMs have actually improved the most at reviewing code over the past 8-9 months, and that ability to review their own code makes them dramatically better at writing code as well.
I find Google Gemini to be pretty impressive at scanning laws and regulations and finding, not creative, but functional solutions to stated problems within the constraints of (often frustratingly bizarre) legal structures.
And all of them will lie to you, tell you what a great idea you have, etc. They’re not really lying, they’re mostly just taking what they read at face value without checking corroborating sources enough to find the obvious (to you) blunders. If you want the LLM to be sure, ask it to go on the RAG (Research Augmented Generation) - check everything before saying it, they can do that, especially “paid mode” engines, but it reduces their capacity for analysis of complex problems by 3-10x, because they’re spending so much context window “being sure” - you can alternatively spend 3-10x as long solving complex problems / accomplishing complex tasks if you have them do their homework, verify everything from “the best” available sources 3x and build up a local document set of “trusted information” which is used in preference to whatever it might find at random on the internet. This isn’t as sexy as “Hey Claude, code me up a database that does X Y Z” and getting the result in 30 seconds, but it is how professionals have been doing their jobs for centuries: learn reliable information first, then act on it.
I think it’s a worthy simplification that sums up the current state of things well. We have AI that broadly has different uses and then we have the ChatBot/ChatBot derived AI applications. The market is not going insane over the broader AI category. Executives are not tripping over themselves to say things like machine vision is going to replace all labor, or at least all white collar labor. The chatbot is the only thing in the conversation of consequence.
All the “everyone says AI is the reality, so everyone feels like they must say AI is the reality” refers to this specific category. It might be veering towards being oversimplified, but it’s trying to balance a perspective that is also oversimplified. Generally, we aren’t good at weighing simple straightforward takes against complex nuanced takes, so you have to “net it out” to have any hope of the point landing.
In the same way that most people if you say you’ll take them to their destination with your vehicle, they assume in your car and not on the back of your bicycle, although both are technically correct.
Pretending AI and Chatbot are interchangeable is a marketing strategy by the companies who make these advanced chat bots. “They are artificially intelligent bro”.
I don’t disagree that the term AI has shifted significantly in the last 4 years or so, but I was actually meaning it in the other direction, as in, there is more to generative AI (what people are calling “AI” these days) than just chat bots.
I generally respect this guy’s opinion (I probably use the term “enshitification” multiple times a week!), but pretending that “AI” and “Chabot” are interchangeable terms seems wildly reductive, and the subtle presumption that the process of changing is a binary changed/non-changed situation instead of a curve is just ignorant.
If the internet was “turned off” after 4 or so years not much would have changed, either.
That’s a questionable statement that may reflect on your main argument. After all, when was the Internet “switched on”? Sounds very binary to me.
To be clear, Doctorow discussed switching off AI and proposed that there would be little to no change on the world. I was just changing the technology to show the flaw in the stance.
Corey Doctorow isn’t speaking to you, he’s speaking to the many, many people who have been fooled by the ai craze.
And often I wonder, how many people have been fooled, and how much of the enthusiasm and over the top fantasy is astroturfing by a few people incentivized to sway public opinion?
A lot of people don’t use it the way he used it though. He used it specifically for two sided markets (eg a service that has both business and personal users) where the focus shifts to the business customers, but people have started using it to mean anything that used to be good but isn’t good any more.
To the average person (and this includes politicians, basically anyone not a software engineer developing the tools )who Cory generally writes to try and inform, they are the same.
Should they be? No. But that’s a whole different thing and trying to change that in the general public mind doesn’t change his current point about AI as the public thinks of it right now.
Speaking at the level of your target audience is wise, but there is a point where information is not just simplfied, but lost, and I think equating chat bots with AI in the broader sense is past that point. It would be like equating websites with the internet.
Is equating chatbots with AI any different than equating a bunch of if/else statements with AI? Is any of this AI considering none of it involves actual intelligence or informed decision making?
Sure AI is more than LLMs, but those are the face of it in the public’s eye and is what’s driving these trillion dollar corporate valuations, which then lead to every company on earth declaring that their product is “AI!”
As a non-techie: websites are the internet.
Like, I conceptualize the internet as basically just a giant shared network, but the point of it is the websites (for me).
I apologize to anyone I may have just gravely offended.
lol
I don’t know how “non-techie” you are, but a good example is online games. When you’re playing Fortnite, Roblox, World of Warcraft, etc you’re not going to a web page, but you’re on the internet.
I straight up almost called it a series of tubes, lol.
Yeah, that’s probably why I do think of it as a shared network, I just don’t play online games, so that’s not what I associate with the internet.
There are other examples but I take your point.
I think the comparison is acceptable with maybe a terminology clarification in the footnotes, because “AI” is the false term under which this current bullshit is being marketed.
And people who believe the core mechanism in large language models to be AI are so uneducated that they will probably neither understand the distinction to legitimate AI research, nor bother to read footnotes.
That’s exactly the thing. Those who understand such a footnote don’t need it, and those who might learn more from it are unlikely to understand it. The above commenters are simply being pedantic.
What’s lost? I can’t think of anything.
Words are meant to convey meaning, and I bet you would lose a lot of non-technical people if you tried to explain the intricacies of what makes large language models worse than content recommendation systems, and doing this would be redundant for technical users.
So Corey can safely use the two words interchangeably and communicate with technical and non-technical alike.
then he has completely failed to understand the concepts for people using him an a information proxy
isnt he supposed to help inform? or is he just an echo chamber to the lowest-common-loudmouth?
Stepping up and doing it better is free.
I think you’re making a strawman argument here. He isn’t arguing that it’s a binary switch that has failed to throw, he’s arguing that we don’t have compelling evidence that the ends justify the means:
And this is completely true! Supposedly “data informed” organizations are trying to find a yardstick that shows actual, meaningful improvement in business outcomes from AI. In my own organization we have people touting LOC yet again because “big number”, but anyone who’s ever worked in software can tell you it’s an asinine metric to use as a KPI.
It’s inherently unscientific to start with the answer and work backwards to a satisfactory question. His point that this push is coming from the least knowledgeable of real processes- and more closely resembles religious fervor than business acumen- seems to at least warrant consideration.
He is definitely pretending that change is either on or off.
Are CEOs jumping the gun on how quickly they adopt AI into workflows? Definitely. However, there’s a big difference between “AI isn’t at a threshold where it is disruptive” and “AI isn’t disruptive”. Or, to belabor the metaphor: CEOs are jumping the gun, but the race is about to start and they’re on the correct track.
A key word in that sentence is “if”, it is a rhetorical example, that’s probably pretty much correct at this point. It’s not a recommended course of action, it’s a declaration that it contrary to crazed hype, it isn’t currently as core to everything as would be befitting the current hype level.
I would argue that the CEOs aren’t on the correct track, they aren’t really in a particularly specific trajectory. I just had a debate with someone on this and their stance was “well in a hundred years do you expect things to be like they are”. My response “I cannot possibly speak to that, but we need to speak to today instead of pretending we know how things will be in a hundred years and pretending they are already at that level”. The “imagine a hundred years from now” by a relative outsider to the tech is dominating CEO mindset, and that’s problematic.
Once the bubble pops, we will almost certainly see a more durable and sane adoption for these technologies. For now, the hype is a problem as it enables some of the worst possible stewards of the technology and favors grift over progress.
If you watch a full podcast (he has done a ton in the last couple weeks), he clearly identifies AI as not a mere hype technology, as something interesting and potentially useful from a technology perspective, and he definitely doesn’t conflate a chatbot with all AI. That doesn’t undermine the vast problems with it, how it’s being used against regular workers, the threat the bubble poses to the economy, the criti-hype cycle, etc.
I think you’re oversimplifying “Cory Doctorow” based on the article you’re reading or specifically how he presents a more complex idea to different audiences.
Watch the Jon Stewart podcast interview if you want more nuance:
https://www.youtube.com/watch?v=-dAIJRjb-Bw
“Chatbot” style AI is wildly good and bad at varying kinds of tasks, and a lot of that has to do with how it has been prepared.
Some LLMs have been trained to make images - I’ve not been too impressed with them, but that’s what they’re “good” at - and better than the LLMs that have been trained to write computer code when you ask the coding LLMs to draw a picture.
The code writing LLMs have actually improved the most at reviewing code over the past 8-9 months, and that ability to review their own code makes them dramatically better at writing code as well.
I find Google Gemini to be pretty impressive at scanning laws and regulations and finding, not creative, but functional solutions to stated problems within the constraints of (often frustratingly bizarre) legal structures.
And all of them will lie to you, tell you what a great idea you have, etc. They’re not really lying, they’re mostly just taking what they read at face value without checking corroborating sources enough to find the obvious (to you) blunders. If you want the LLM to be sure, ask it to go on the RAG (Research Augmented Generation) - check everything before saying it, they can do that, especially “paid mode” engines, but it reduces their capacity for analysis of complex problems by 3-10x, because they’re spending so much context window “being sure” - you can alternatively spend 3-10x as long solving complex problems / accomplishing complex tasks if you have them do their homework, verify everything from “the best” available sources 3x and build up a local document set of “trusted information” which is used in preference to whatever it might find at random on the internet. This isn’t as sexy as “Hey Claude, code me up a database that does X Y Z” and getting the result in 30 seconds, but it is how professionals have been doing their jobs for centuries: learn reliable information first, then act on it.
I think it’s a worthy simplification that sums up the current state of things well. We have AI that broadly has different uses and then we have the ChatBot/ChatBot derived AI applications. The market is not going insane over the broader AI category. Executives are not tripping over themselves to say things like machine vision is going to replace all labor, or at least all white collar labor. The chatbot is the only thing in the conversation of consequence.
All the “everyone says AI is the reality, so everyone feels like they must say AI is the reality” refers to this specific category. It might be veering towards being oversimplified, but it’s trying to balance a perspective that is also oversimplified. Generally, we aren’t good at weighing simple straightforward takes against complex nuanced takes, so you have to “net it out” to have any hope of the point landing.
In the same way that most people if you say you’ll take them to their destination with your vehicle, they assume in your car and not on the back of your bicycle, although both are technically correct.
Pretending AI and Chatbot are interchangeable is a marketing strategy by the companies who make these advanced chat bots. “They are artificially intelligent bro”.
I don’t disagree that the term AI has shifted significantly in the last 4 years or so, but I was actually meaning it in the other direction, as in, there is more to generative AI (what people are calling “AI” these days) than just chat bots.
Oh look it’s arbitrary-narrow-definition-that-fits-the-argument-I-want-to-make man!
I’m saying there’s more to generative AI than chatbots. My definition of AI is broader. What are you going on about?