I love this line of thinking but the framing comes across to me as a bit inflammatory and uncharitable. Just about anything involved in the exchanged of ideas can be viewed as a virus! That is because, when viewed from an evolutionary biology perspective, ideas are the cognitive equivalent of genes (in fact, this is where the term meme comes from).
There's a whole field, evolutionary memetics, dedicated to this. Cultural values, marketing, religion, social media, education, propaganda, etc. can all be viewed quite naturally through this lense. Even the terms we use in every day language (eg, going "viral") seem to intuitively grasp this.
Is my friend a virus for recommending me their favorite book? Well, yes I guess, but that's kind of just how ideas work. Just my two cents
Not sure why this is interesting; it seems like anything that becomes "popular" could be cast as a "virus", including things like "having a fridge". Why does this give particular insight?
And like a cult to those of us preferring not to engage, for whatever reason, with everyone and their dog telling us that we should (and sometimes getting quite offended, like our refusal to play that game is a personal attack on them for choosing to).
Leads me to think of a related topic, the Cost of Cognitive Debt. I think it is about time that we start to quantify what it means that we no longer grasp major parts of our systems and infrastructure. Once we quantify it, we can start to reason about how to address it.
I literally thought of this comparison last week. The problem I noticed was how my parents' generation lacked the cognitive "defenses" to separate AI social media content on YouTube, WhatsApp and the like from high quality information. The AI generated content does not trigger an "uncanny valley" reaction, and so people (I myself included) have greater difficulty immediately separating what information is genuine and what is synthetic, and then this spreads exponentially through social networks.
> The viral analogy does not imply that LLM-human interactions are intrinsically parasitic.
The analogy doesn't, but LLMs are trained on text, or content, generated by humans, with the aim of their output being acceptable to human interlocutors. That's one parasitic aspect. A second one is the nature and behavior of the organizational entities providing and promoting the popular LLM services: Microsoft, Google, Meta, Anthropic etc - their influence on the economy, e.g. in energy and water use and the drawing of capital/human efforts - also has a somewhat parasitic aspect.
> with the aim of their output being acceptable to human interlocutors.
The more I use LLMs I find that their output drifts toward being acceptable to other LLMs. I find myself quite often telling the model: "this is not human readable.". Especially Opus and Fable write in such a way that it is difficult to follow. The often put sentences out of context, the context itself is fragmented. Often start with what you would think is end of a story so when you start reading you are like wtf it is about.
If think the issue might be that LLMs read text all at once whereas humans are somewhat linear. So when LLM puts out text it understands, to human it is close to random gibberish that looks like English.
Have to take commentary by non tech folks speaking about tech with a grain of salt and perspective since the tech is changing so fast both in capability and application.
Research can be more of a lagging indicator than a leading indicator with AI.
I can't disagree enough with this. Not everything needs to be understood in a technical context. Technology intersects with society, and understanding its impacts on that society is vital.
They're specifically not using the term "virus" pejoratively, but rather distinguishing the social mechanism LLMs seem to have from something as simple as a "meme".
If this is interesting, message me about my “elves”
They are the cognitive construct I fortified in my mind and in my code before the models began producing their own training materials. My approach is based on two decades of coding without ai.
The immunization is becoming increasingly more valuable as people begin understanding the market in which I do business.
I love this line of thinking but the framing comes across to me as a bit inflammatory and uncharitable. Just about anything involved in the exchanged of ideas can be viewed as a virus! That is because, when viewed from an evolutionary biology perspective, ideas are the cognitive equivalent of genes (in fact, this is where the term meme comes from).
There's a whole field, evolutionary memetics, dedicated to this. Cultural values, marketing, religion, social media, education, propaganda, etc. can all be viewed quite naturally through this lense. Even the terms we use in every day language (eg, going "viral") seem to intuitively grasp this.
Is my friend a virus for recommending me their favorite book? Well, yes I guess, but that's kind of just how ideas work. Just my two cents
Memes are what we are talking about.
https://www.suzannetreister.net/HEXEN2/TAROT_COL/TAROT_WofF_...
Not sure why this is interesting; it seems like anything that becomes "popular" could be cast as a "virus", including things like "having a fridge". Why does this give particular insight?
LLMs are both a blessing and a curse, they magnificent tools for those who know how to employ them , but a crutch for those who rely on them.
And like a cult to those of us preferring not to engage, for whatever reason, with everyone and their dog telling us that we should (and sometimes getting quite offended, like our refusal to play that game is a personal attack on them for choosing to).
Leads me to think of a related topic, the Cost of Cognitive Debt. I think it is about time that we start to quantify what it means that we no longer grasp major parts of our systems and infrastructure. Once we quantify it, we can start to reason about how to address it.
I explored this concept in this post:
https://ljtn.github.io/epiq/blog/cost-of-cognitive-debt.html
I literally thought of this comparison last week. The problem I noticed was how my parents' generation lacked the cognitive "defenses" to separate AI social media content on YouTube, WhatsApp and the like from high quality information. The AI generated content does not trigger an "uncanny valley" reaction, and so people (I myself included) have greater difficulty immediately separating what information is genuine and what is synthetic, and then this spreads exponentially through social networks.
Snow crash?
Ok... but are automobiles also a virus? Because we've become dependent on them? Should we also facilitate reversibility in our automobile usage?
What about cooking food? We've become entirely dependent on cooking food.
> Should we also facilitate reversibility in our automobile usage
Yes?
That's why I added the part about food. Or heck, computers. Or telephones. Or any incredibly useful technology.
Toothbrushes! What will we do if we come to rely on toothbrushes???
Zero surprise this is coming from a European institution. The continent is destined to be left behind.
It is a collaboration across many institutions including those from Europe, the US, and Japan
> The viral analogy does not imply that LLM-human interactions are intrinsically parasitic.
The analogy doesn't, but LLMs are trained on text, or content, generated by humans, with the aim of their output being acceptable to human interlocutors. That's one parasitic aspect. A second one is the nature and behavior of the organizational entities providing and promoting the popular LLM services: Microsoft, Google, Meta, Anthropic etc - their influence on the economy, e.g. in energy and water use and the drawing of capital/human efforts - also has a somewhat parasitic aspect.
> with the aim of their output being acceptable to human interlocutors.
The more I use LLMs I find that their output drifts toward being acceptable to other LLMs. I find myself quite often telling the model: "this is not human readable.". Especially Opus and Fable write in such a way that it is difficult to follow. The often put sentences out of context, the context itself is fragmented. Often start with what you would think is end of a story so when you start reading you are like wtf it is about.
If think the issue might be that LLMs read text all at once whereas humans are somewhat linear. So when LLM puts out text it understands, to human it is close to random gibberish that looks like English.
My thesis advisor and team doing interesting work.
I mean, what else can we call a "cognitive virus"? Quite a lot, actually.
Have to take commentary by non tech folks speaking about tech with a grain of salt and perspective since the tech is changing so fast both in capability and application.
Research can be more of a lagging indicator than a leading indicator with AI.
I can't disagree enough with this. Not everything needs to be understood in a technical context. Technology intersects with society, and understanding its impacts on that society is vital.
Or in other words: the humanities are good.
It helps to have outside observers looking at the effects of a new technology and industry and not just depending on those within it to report for us.
They're specifically not using the term "virus" pejoratively, but rather distinguishing the social mechanism LLMs seem to have from something as simple as a "meme".
Contagion probably is more accurate.
Well, a meme is contagious, but it doesn't do the same work a virus can at altering the structure of its host.
STD outbreak maps better.
If this is interesting, message me about my “elves”
They are the cognitive construct I fortified in my mind and in my code before the models began producing their own training materials. My approach is based on two decades of coding without ai.
The immunization is becoming increasingly more valuable as people begin understanding the market in which I do business.