Thoughts on Slowing the Fuck Down

Bugs, burnout and AI blame: users say slow down, vets eye-roll, believers push harder

TLDR: The piece warns that rushing AI-built software is flooding users with bugs and odd outages. Commenters split: veterans say the code isn’t worse—our speed is—while others insist AI code can ship and some beg for slower, saner work, a big deal because these apps run everyday life.

After a year of “AI coding agents” promising apps on autopilot, one fed‑up writer says the quiet part loud: everything feels buggier, with “98% uptime” now normal and interfaces acting haunted. They cite a disputed rumor about an AI-related cloud outage, brag posts about 100% AI-written products, and a general vibe that we’ve traded craftsmanship for a speed high. Hobby projects? Fine. But for real users, the author claims teams let bots choose features nobody wants and coded themselves into a corner.

The comments go full cage match. A grizzled operator, 0xbadcafebee, insists “the software has not changed” — we just removed human guardrails and now flaws hit us faster. Others are exhausted by the panic cycle: badlibrarian says this is Ruby/PHP/VB déjà vu — use the tool sanely, period. ex-aws-dude predicts a self-correcting smack from maintenance hell once bills come due. Then ketzo storms in swinging: calling AI code “not production-ready” is “insane,” arguing top models can ship without humans reviewing every line. Meanwhile, the memes roar: “98% uptime is the new Uber 5‑star,” “Clippy’s back and it writes Windows,” and grandma’s bot opening a Shopify. The internet can’t decide whether to slam the brakes or floor it — but everyone agrees the ride is getting bumpy.

Key Points

  • The article reviews a year of AI coding agents and distinguishes earlier assistant-like tools (e.g., Aider, early Cursor) from newer agents that build full projects.
  • It says Anthropic and OpenAI holiday giveaways expanded exposure to agentic coding among developers.
  • The author alleges software reliability is worsening and cites reports involving AWS and Microsoft (AI-written code scale; perceived Windows quality issues).
  • It claims over-reliance on agents (no code review, delegated design, excess features) leads to brittle, low-quality software.
  • The piece criticizes orchestration of autonomous agent swarms and cites examples (agent-built C compiler, Cursor-led browser) as not production-ready without human oversight.

Hottest takes

"the software has not changed." — 0xbadcafebee
"Use the tool sanely." — badlibrarian
"It’s just completely insane to me" — ketzo
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.