I've been saying this for years. It's curious that 90% of the top-level comments here are all dismissing it outright. And we have the usual themes: 1) "This h

I've been saying this for years. It's curious that 90% of the top-level comments here are all dismissing it outright. And we have the usual themes: 1) "This has always been possible before. AI brings nothing new." 2) "We haven't seen anything really bad yet, so it is a non-issue 3) "AI execs are pushing this narrative to make AI sound important and get more money" Nevermind the fact that many famous people behind the invention of AI, including Jeffrey Hinton (the "godfather of AI") and others quit their jobs and are spending their time loudly warning people, or signing major letters where they warn about human extinction or job loss... it's all a grift according to the vocal HN denizens. This is like the opposite of web3 where everyone piles on the other way. Well... swarms of AI agents take time to amass karma, pagerank, and other metrics, but when with the coming years, they will indeed be able to churn out content 24/7, create normal-looking influencer accounts and dominate the online discussion on every platform. Very likely, the percentage of human-generated content will trend to 0-1% of content on the internet, and it will become a dark forest, and this will be true on "siloed" ecosystems like HN as well: https://maggieappleton.com/ai-dark-forest/ https://maggieappleton.com/ai-dark-forest/ Certainly, saying "this was always possible before" misses the forest for the trees. No, it wasn't. List of attacks made possible by swarms of agents: 1) Edits to Wikipedia and publishing articles to push a certain narrative, as an Advanced Persistent Threat 2) Popular accounts on social media, videos on YouTube, Instagram and TikTok pushing AI-generated narratives, biased "news" and entertainment across many accounts growing in popularity (already happening) 3) Comments under articles, videos and shared posts that are either for or against the thing, and coordinated upvoting that bypasses voting ring detection (most social networks don't care about it as much as HN). Tactical piling on. 4) Sleeper accounts that appear normal for months or years and amass karma / points until they gradually start coordinating, either subtly or overtly. AI can play the long strategic game and outmaneuver groups people as well, including experts (see #7 for discrediting them). 5) Astroturfing attacks on people who disagree with the narrative. Maybe coordinating posts on HN that make it seem like it is an unpopular position. 6) Infiltrating and distracting opponents of a position, by getting them mired in constant defenses or explanations, where the interlocutors are either AI or friends / allies that have been "turned" or "flipped" by AI to question them 7) Reputational destruction, along the lines of NSA PRISM powerpoint slides (https://archive.org/details/NSA-PRISM-Slides https://archive.org/details/NSA-PRISM-Slides) but at scale and implacable. 8) Astroturfing support for wars, unrest, or whatever else, but at scale, along the lines of Mahachkala Protests in Russia, etc. These are just some of the early low-hanging fruit for 2026 and 2027.