Attention is all you have
Algorithms optimized for engagement are reshaping how people spend their attention online, turning feeds, news, and recommendations into systems that monetize distraction and encourage “doomscrolling.” Commenters contrast this with earlier, more intentional web use—bookmarking sites, reading books, joining forums—and describe how always-on, personalized feeds, ads, and AI-generated content erode focus, taste, and even mental health. Many advocate practical countermeasures such as ad blocking, disabling recommendations, using RSS, self-hosting, time limits, and meditation, arguing that deliberately directing one’s attention is becoming a core 21st‑century survival skill.
Title, expectations, and ML reference
- Many initially assumed the post was about machine learning due to the echo of “Attention is all you need.”
- Some see the title as clickbait; others view it as a deliberate nod tying ML “attention” to human attention hijacked by algorithms.
- A few think the feigned ignorance of the ML paper is itself a rhetorical move; others say a joke doesn’t need to be explained.
Algorithms, doomscrolling, and attention hijacking
- Broad agreement that modern recommendation systems optimize for engagement, not wellbeing, and easily turn into doomscrolling.
- Some argue this is just a continuation of older incentives (TV, portals, early portals like MSN/Yahoo, news media).
- Others stress personalization at scale and infinite feeds as a qualitative change that magnifies polarization, outrage, and distraction.
Books, old web, and enshittification
- Several praise books and older media as less hijacking, though others note low-effort pulp has always existed and that modern books can be written for shorter attention spans.
- Many miss “intentional” pre-algorithm internet: bookmarks, blogs, forums, IRC, RSS, and non-personalized front pages.
- There’s nostalgia for “peak internet” when going online was a deliberate act and dismay at today’s “television, but worse.”
Coping strategies and technical hacks
- Common tactics: blocking algorithmic feeds, using RSS, uBlock/AdGuard, feed-eradicating extensions, and curated YouTube subscriptions only.
- Some self-host, segment networks, or enforce offline periods (timers, focus modes, blocking apps).
- Tools like full‑text history search, web clippers, and read‑later apps are proposed as “personal algorithms” to replace platform curation.
Advertising, AI slop, and content quality
- Strong hostility toward advertising as shaping desires and self-image, not just informing.
- Concern that cheap AI-generated “slop” plus low consumer standards (analogy to bad coffee) will flood culture.
- Others counter that recommendation algorithms can be useful if users actively prune and understand their biases.
Mental health, intention, and practice
- Multiple anecdotes of cutting social media/news leading to better mood and more reading or hobbies.
- Some emphasize that mere choice of sites isn’t enough; presence, attention, and meditation-like practices are needed to resist default mind-wandering.
- Debate over whether doomscrolling is net harmful, with references to declining test scores vs intuitions that younger “iPad kids” seem sharper.