AI guru behind Vibe Coding says only RSS can save us from slop on internet
AI slop is now everywhere, and even the people building the future of AI are seemingly feeling overwhelmed by it. For AI guru Andrej Karpathy, known for his work at OpenAI and Tesla, who famously coined the term "vibe coding", the fix for cutting down the slop is not less AI but smarter consumption. RSS feeds are one way to do that, he says.
AI slop is growing. For those unaware, the term refers to content that is made – or generated – by AI. It could be text, audio, or video. There is just lots of it everywhere so much so that platforms like YouTube are now cracking down on channels using AI – and only AI – to flood the internet with so-called “AI slop”. The growth has been so rampant, even veterans are starting to take notice, acknowledge, and speak out on the issue. Latest to join the bandwagon is Andrej Karpathy, the man who has played key roles at OpenAI and Tesla – and who famously coined the term vibe coding last year. The AI guru has made the case for the return of RSS feeds saying he looks up to them for when he wants to read high-quality long-form content and not AI slop with a propensity to “provoke.”
Taking to X, the platform previously known as Twitter, Karpathy, who co-founded OpenAI and later served as Tesla’s AI director, argued that bringing RSS/Atom feeds back may help users regain control over the type of content they see online. He shared that he himself is returning to RSS and Atom feeds to escape what he calls the rising tide of online “slop”, saying they offer higher-quality long-form writing and far less content designed purely to provoke engagement.
“Finding myself going back to RSS/Atom feeds a lot more recently. There's a lot more higher quality longform and a lot less slop intended to provoke. Any product that happens to look a bit different today but that has fundamentally the same incentive structures will eventually converge to the same black hole at the centre of gravity,” he posted.
His comments come at a time when concerns around AI-generated content – the AI slop is everywhere. With AI models becoming smarter and allowing users to produce videos with text, images, at a near-zero cost, social media platforms are getting flooded with such kind of content optimised for clicks rather than insight. This is raising concerns over the fact that AI slop is making it harder for users to find thoughtful and original work amid the noise.
Karpathy explains that the issue is not AI itself, but the incentive structures behind modern platforms. In his view, any product that relies on the same engagement-driven mechanics, likes, shares, outrage, and endless scrolling, will eventually collapse into the same low-quality centre of gravity. Even platforms that appear different on the surface, he suggested, are destined to converge towards the same outcome if their incentives remain unchanged.
As an alternative, Karpathy strongly endorsed RSS, describing it as “open, pervasive, and hackable”. Unlike algorithmic social feeds, RSS puts users in control, allowing them to choose exactly which sources they want to follow instead of having content pushed by opaque recommendation systems designed to maximise time spent on a platform.
Karpathy is, of course, not against AI itself. He is the person who popularised the term “vibe coding”, a process where developers rely heavily on AI tools and intuition rather than writing every line of code manually. In an earlier post on X, he encouraged vibe coding and even shared how he is using AI to write the code based on his mood. “There's a new kind of coding I call vibe coding, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs are getting too good.”
So, his criticism of AI slop, therefore, comes not from rejecting AI, but from concern over how it is shaping the quality of content across the internet, and impacting the originally worthy human conversations.
AI slop is growing. For those unaware, the term refers to content that is made – or generated – by AI. It could be text, audio, or video. There is just lots of it everywhere so much so that platforms like YouTube are now cracking down on channels using AI – and only AI – to flood the internet with so-called “AI slop”. The growth has been so rampant, even veterans are starting to take notice, acknowledge, and speak out on the issue. Latest to join the bandwagon is Andrej Karpathy, the man who has played key roles at OpenAI and Tesla – and who famously coined the term vibe coding last year. The AI guru has made the case for the return of RSS feeds saying he looks up to them for when he wants to read high-quality long-form content and not AI slop with a propensity to “provoke.”
Taking to X, the platform previously known as Twitter, Karpathy, who co-founded OpenAI and later served as Tesla’s AI director, argued that bringing RSS/Atom feeds back may help users regain control over the type of content they see online. He shared that he himself is returning to RSS and Atom feeds to escape what he calls the rising tide of online “slop”, saying they offer higher-quality long-form writing and far less content designed purely to provoke engagement.
“Finding myself going back to RSS/Atom feeds a lot more recently. There's a lot more higher quality longform and a lot less slop intended to provoke. Any product that happens to look a bit different today but that has fundamentally the same incentive structures will eventually converge to the same black hole at the centre of gravity,” he posted.
His comments come at a time when concerns around AI-generated content – the AI slop is everywhere. With AI models becoming smarter and allowing users to produce videos with text, images, at a near-zero cost, social media platforms are getting flooded with such kind of content optimised for clicks rather than insight. This is raising concerns over the fact that AI slop is making it harder for users to find thoughtful and original work amid the noise.
Karpathy explains that the issue is not AI itself, but the incentive structures behind modern platforms. In his view, any product that relies on the same engagement-driven mechanics, likes, shares, outrage, and endless scrolling, will eventually collapse into the same low-quality centre of gravity. Even platforms that appear different on the surface, he suggested, are destined to converge towards the same outcome if their incentives remain unchanged.
As an alternative, Karpathy strongly endorsed RSS, describing it as “open, pervasive, and hackable”. Unlike algorithmic social feeds, RSS puts users in control, allowing them to choose exactly which sources they want to follow instead of having content pushed by opaque recommendation systems designed to maximise time spent on a platform.
Karpathy is, of course, not against AI itself. He is the person who popularised the term “vibe coding”, a process where developers rely heavily on AI tools and intuition rather than writing every line of code manually. In an earlier post on X, he encouraged vibe coding and even shared how he is using AI to write the code based on his mood. “There's a new kind of coding I call vibe coding, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs are getting too good.”
So, his criticism of AI slop, therefore, comes not from rejecting AI, but from concern over how it is shaping the quality of content across the internet, and impacting the originally worthy human conversations.