The future of everything is lies, I guess – Part 5: Annoyances
Large language models are seen by many here as an accelerant for existing problems in consumer capitalism: manipulative interfaces, obstructive customer service, opaque pricing, and the diffusion of responsibility behind “the computer says no.” Commenters argue that AI will likely be used to cut costs, widen class divides, and further erode trust in institutions and information, even as others note real convenience gains and potential for productivity and research. The tension centers on whether regulation, market choices, or new high‑trust communities can meaningfully steer these technologies toward public benefit rather than deeper enshittification of everyday life.
Access to the blog / UK blocking
- Some UK readers can’t access the site; others note the domain is reachable.
- Linked post by the author explains deliberate IP geoblocking of the UK, framed as a response to the Online Safety Act.
- Debate over whether this is meaningful “protest” or just a necessary way to avoid compliance/liability.
LLMs, manipulation, and capitalism
- Many see LLMs as another tool to widen class divides, optimize extraction, and intensify already‑existing manipulative practices (dynamic pricing, claims denial, engagement farming).
- Others argue manipulation has always been central to markets and media; AI just scales it.
- A minority highlight positive potential: cheaper software, research acceleration, and local models empowering individuals.
Customer service, “enshittification,” and AI
- Strong concern that AI support will become a wall between users and real remedies, further reducing accountability.
- Several note this trend predates LLMs: phone trees, offshore scripts, and “computer says no” cultures already aim to reduce ticket volume, not help.
- Some predict good human support becomes a premium differentiator, but others say monopoly/oligopoly power undermines that.
Optimism vs doomerism about AI
- One camp: society has adapted to cars, phones, GM crops; LLMs are “just tools” and not civilization-ending.
- Opposing camp: collapse/decline is plausible; AI could accelerate existing negative trajectories, and benefits so far feel incremental.
Trust, information quality, and communities
- Nostalgia for the “old internet” as a relatively high‑trust whalefall now being eaten by spam, ragebait, and AI slop.
- People expect retreat into high‑trust zones: in‑person ties, closed chats, brands with reputations.
- Proposals split between proof‑of‑human / real‑identity networks and rule‑based spaces where it doesn’t matter if participants are bots.
Regulation, incentives, and accountability
- Many argue the root problem is incentives under consumer capitalism, not AI per se.
- Suggestions include stronger consumer law, antitrust, and “anti‑Kafka” style rules requiring reachable human recourse.
- Others are pessimistic: regulatory capture, weak democracies, and bribery‑driven politics make effective regulation unlikely.
Cognition and discourse
- Several commenters report friends leaning on AI summaries, leading to shallower engagement and shorter attention spans.
- A cited study claims prolonged LLM use in writing tasks reduces cognitive effort and performance over time.
- Some intentionally move back to long‑form email/letters to preserve deeper thinking.