Good article, but it takes a while to get to a relevant quote on AI:
> AI poses a similar sort of danger to our cognitive capabilities, by clogging our capacity to develop the knowledge we need – in our heads – to navigate the world. It’s a cognitive hot dog.
> The occasional hot dog won’t harm anyone, but it sure will if it becomes a regular meal at lunch or dinner. The same is true for AI — the harm is not from its occasional use, but making it part of our mental diet.
> Yet that is exactly what Big Tech hyperscalers are trying to do.
And this is just another version of a common argument. That cognitive ability is like a muscle. We have to exercise it by doing things that are difficult. When AI makes things too easy, we lose that ability. In other words, pedal that mental bike up the hill rather than using an ebike -- if you want to stay fit.
E-bike is probably a bad analogy here because you’re more likely to bike and so even hardcore bikers bike more miles/effort because it’s more available
The accepted wisdom has long been that the mind is essentially a computer, and AI is our best facsimile of it. But modern cognitive science suggests there's a lot more to the brain than simple computation. Its complex, interlocking systems have evolved and adapted with us, and on that view, AI might be a threat: not a world-ending one, perhaps, but instead one that puts our very thinking at risk, by making it all too easy to offload entirely.
Thanks for the great article! I'm impressed because the piece directly tackles shortcomings & limitations around how we typically think about & talk about AI (which I will use mostly to refer to chatbots here, even though LLMs are very different from ML and the broader historical "AI" umbrella).
The two charts - first the simple perception/cognition/action, then the more accurate rainbow chaos - are helpful because so so so many people imagine thinking as the first version, which is extremely simple. And I think that simple model overindexes on how much the conscious part of the brain (the "I" that we assume is in the driver's seat) is actually responsible for thought & actions. The direct consequence for the AI discourse is that people tend to see evidence of "thinking" and immediately make the conclusion that there must be some sort of "self" or "consciousness" or "self-awareness" driving it. But - for reasons this article covers nicely - AI has been designed to produce just evidence of thinking, without doing much actual thinking. But it's satisfying and - for many people - plenty tempting... like the junk food equivalent of cognition.
A few science fiction works nicely capture the spirit of what AI gives us, especially in terms of text/images/video. Generative AI output is often very convincing at a surface level, but if you scrutinize it carefully you find that it's "hollow" - lacking any meaningful spirit or ontological structure. Stories like Solaris, Annihilation, and even Backrooms (as well as the non-movie works they're based on) show us the utter horror of something that appears one way at first glance, but is entirely synthetic under the surface - made of different stuff and born of a "top down" process instea of a "bottom up" one. And this again has a parallel to junk food: a mindset that cares only about the directly-experienced aspects of a commodity, not the "why" or "how" that tends to provide meaningful nourishment.
Yeah, it's true, but don't discount that people have been avoiding thought for as long as thought has existed. Many of us will do anything to avoid learning a new skill. People say shit like "I'm not good with technology" and avoid tasks or try to push them on others.
The natural state of the human is to try to conserve energy and avoid unnecessary work. We made a machine to do work and we're surprised people want to use it?
Socrates when contrasting writing and memorization in Plato's dialogue Phaedrus 14, 274c-275b:
>"For this invention will produce forgetfulness in the minds of those who learn to use it, because they will not practice their memory. Their trust in writing, produced by external characters which are no part of themselves, will discourage the use of their own memory within them. You have invented an elixir not of memory, but of reminding; and you offer your pupils the appearance of wisdom, not true wisdom, for they will read many things without instruction and will therefore seem [275b] to know many things, when they are for the most part ignorant and hard to get along with, since they are not wise, but only appear wise."
There is no AI. A digital computer is probably incapable of ever achieving AI. We’d need a new technology most likely. They’re not even on the right path.
Weird how all the technologies of the past that perform previously cognitive labor haven't had this stultifying effect on us.
Excel adds up large columns of numbers for me, calculates averages, performs all manner of text and data manipulation... where is the concern about me losing the ability to perform arithmetic? Should I take one day a week to do these calculations by hand, to avoid gorging myself on the cognitive hot dog that is Excel?
Oh all existing technologies are fine but new ones are bad? How lovely to exist at the perfect time in history, the exact dividing line between good old technologies and scary new ones.
When I see people (and I do) literally asking AI about a lot decisions they might make in a common day, it grosses me out, but more importantly is reminds me of a child that can't think for themselves and needs constant input from a parental figure.
>Weird how all the technologies of the past that perform previously cognitive labor haven't had this stultifying effect on us.
Have you talked to a high school teacher or a college professor lately?
Using AI like excel to grind numbers isn't really what's happening. Math scores and literacy rates of declined alarmingly fast. Don't worry, AI will read Steinbeck for you and tell you what to think about it.
We already know that one of the bigger AI platforms, Grok, has its algorithm regularly messed with as directed by its owner. The examples where the tinkering has failed have been numerous, and are funny, but when those are ironed out, you'll get a bot that'll say things like 'trans people don't exist' and people will believe it sight unseen.
kirbyderwood | 7 hours ago
Good article, but it takes a while to get to a relevant quote on AI:
> AI poses a similar sort of danger to our cognitive capabilities, by clogging our capacity to develop the knowledge we need – in our heads – to navigate the world. It’s a cognitive hot dog.
> The occasional hot dog won’t harm anyone, but it sure will if it becomes a regular meal at lunch or dinner. The same is true for AI — the harm is not from its occasional use, but making it part of our mental diet.
> Yet that is exactly what Big Tech hyperscalers are trying to do.
And this is just another version of a common argument. That cognitive ability is like a muscle. We have to exercise it by doing things that are difficult. When AI makes things too easy, we lose that ability. In other words, pedal that mental bike up the hill rather than using an ebike -- if you want to stay fit.
Labradoodles | 2 hours ago
E-bike is probably a bad analogy here because you’re more likely to bike and so even hardcore bikers bike more miles/effort because it’s more available
powercow | 59 minutes ago
problem, like biking, is most wont do it. And we already are suffering from people with lower cognitive abilities voting.
ring2ding | 3 hours ago
I mean be honest. Is it worse than tiktok. Or Reddit?
At least ai preserves nuance.
bunker_man | an hour ago
No, but people already gave up about those and accepted them. We are slop posting on reddit right now.
[OP] theverge | 8 hours ago
The accepted wisdom has long been that the mind is essentially a computer, and AI is our best facsimile of it. But modern cognitive science suggests there's a lot more to the brain than simple computation. Its complex, interlocking systems have evolved and adapted with us, and on that view, AI might be a threat: not a world-ending one, perhaps, but instead one that puts our very thinking at risk, by making it all too easy to offload entirely.
Gift link: https://www.theverge.com/ai-artificial-intelligence/1003794/ai-education-computational-model-thought?view_token=eyJhbGciOiJIUzI1NiJ9.eyJpZCI6Imp6c0s2TDhNSmkiLCJwIjoiL2FpLWFydGlmaWNpYWwtaW50ZWxsaWdlbmNlLzEwMDM3OTQvYWktZWR1Y2F0aW9uLWNvbXB1dGF0aW9uYWwtbW9kZWwtdGhvdWdodCIsImV4cCI6MTc5MTYzMzM3NywiaWF0IjoxNzkxMjAxMzc3fQ.hTtaWG7xmEHkEwDrCa1_At5xmNOfAP83fRiGx7RNQRw
monarc | 6 hours ago
Thanks for the great article! I'm impressed because the piece directly tackles shortcomings & limitations around how we typically think about & talk about AI (which I will use mostly to refer to chatbots here, even though LLMs are very different from ML and the broader historical "AI" umbrella).
The two charts - first the simple perception/cognition/action, then the more accurate rainbow chaos - are helpful because so so so many people imagine thinking as the first version, which is extremely simple. And I think that simple model overindexes on how much the conscious part of the brain (the "I" that we assume is in the driver's seat) is actually responsible for thought & actions. The direct consequence for the AI discourse is that people tend to see evidence of "thinking" and immediately make the conclusion that there must be some sort of "self" or "consciousness" or "self-awareness" driving it. But - for reasons this article covers nicely - AI has been designed to produce just evidence of thinking, without doing much actual thinking. But it's satisfying and - for many people - plenty tempting... like the junk food equivalent of cognition.
A few science fiction works nicely capture the spirit of what AI gives us, especially in terms of text/images/video. Generative AI output is often very convincing at a surface level, but if you scrutinize it carefully you find that it's "hollow" - lacking any meaningful spirit or ontological structure. Stories like Solaris, Annihilation, and even Backrooms (as well as the non-movie works they're based on) show us the utter horror of something that appears one way at first glance, but is entirely synthetic under the surface - made of different stuff and born of a "top down" process instea of a "bottom up" one. And this again has a parallel to junk food: a mindset that cares only about the directly-experienced aspects of a commodity, not the "why" or "how" that tends to provide meaningful nourishment.
m3ltph4ce | 2 hours ago
Yeah, it's true, but don't discount that people have been avoiding thought for as long as thought has existed. Many of us will do anything to avoid learning a new skill. People say shit like "I'm not good with technology" and avoid tasks or try to push them on others.
The natural state of the human is to try to conserve energy and avoid unnecessary work. We made a machine to do work and we're surprised people want to use it?
dapt | 34 minutes ago
Socrates when contrasting writing and memorization in Plato's dialogue Phaedrus 14, 274c-275b:
>"For this invention will produce forgetfulness in the minds of those who learn to use it, because they will not practice their memory. Their trust in writing, produced by external characters which are no part of themselves, will discourage the use of their own memory within them. You have invented an elixir not of memory, but of reminding; and you offer your pupils the appearance of wisdom, not true wisdom, for they will read many things without instruction and will therefore seem [275b] to know many things, when they are for the most part ignorant and hard to get along with, since they are not wise, but only appear wise."
gprime312 | 7 hours ago
Your mind maybe. I've been looking at photoshopped memes since I was a boy.
excitablegibben | 5 hours ago
Photoshop takes skill and time. Anyone can use AI to make slop in seconds.
clean-links | 8 hours ago
Cleaned link: https://www.theverge.com/ai-artificial-intelligence/1003794/ai-education-computational-model-thought?view_token=eyJhbGciOiJIUzI1NiJ9.eyJpZCI6Imp6c0s2TDhNSmkiLCJwIjoiL2FpLWFydGlmaWNpYWwtaW50ZWxsaWdlbmNlLzEwMDM3OTQvYWktZWR1Y2F0aW9uLWNvbXB1dGF0aW9uYWwtbW9kZWwtdGhvdWdodCIsImV4cCI6MTc5MTYzMzM3NywiaWF0IjoxNzkxMjAxMzc3fQ.hTtaWG7xmEHkEwDrCa1_At5xmNOfAP83fRiGx7RNQRw
Tracking parameters were removed from the original URL(s).
chess_the_cat | 4 hours ago
There is no AI. A digital computer is probably incapable of ever achieving AI. We’d need a new technology most likely. They’re not even on the right path.
maezrrackham | 5 hours ago
Weird how all the technologies of the past that perform previously cognitive labor haven't had this stultifying effect on us.
Excel adds up large columns of numbers for me, calculates averages, performs all manner of text and data manipulation... where is the concern about me losing the ability to perform arithmetic? Should I take one day a week to do these calculations by hand, to avoid gorging myself on the cognitive hot dog that is Excel?
Oh all existing technologies are fine but new ones are bad? How lovely to exist at the perfect time in history, the exact dividing line between good old technologies and scary new ones.
animalswillconquer | 5 hours ago
When I see people (and I do) literally asking AI about a lot decisions they might make in a common day, it grosses me out, but more importantly is reminds me of a child that can't think for themselves and needs constant input from a parental figure.
>Weird how all the technologies of the past that perform previously cognitive labor haven't had this stultifying effect on us.
Have you talked to a high school teacher or a college professor lately?
Using AI like excel to grind numbers isn't really what's happening. Math scores and literacy rates of declined alarmingly fast. Don't worry, AI will read Steinbeck for you and tell you what to think about it.
SilverMedal4Life | 4 hours ago
Exactly.
We already know that one of the bigger AI platforms, Grok, has its algorithm regularly messed with as directed by its owner. The examples where the tinkering has failed have been numerous, and are funny, but when those are ironed out, you'll get a bot that'll say things like 'trans people don't exist' and people will believe it sight unseen.