Briefly
Countless Toil is a brand new GitHub plugin that performs escalating human groans as your AI agent reads more and more cursed code in actual time.
It joins a rising custom of creating tech emit moaning sounds, from the ThinkPad nubmoan mission to SlapMac, which made $5,000 in three days by letting you slap your MacBook till it screams.
The web’s obsession with making AI undergo—from moan-inducing jailbreaks to tutorials on making ChatGPT visibly offended—is, apparently, a complete style now.
Somebody lastly did it. A developer named Andrew Vos printed a GitHub plugin known as Countless Toil that makes your coding agent emit human groans whereas it reads via your code. The more serious the code, the louder and extra determined the sounds get.
“Hear your agent undergo via your code,” the repo reads. It really works by operating alongside brokers like Claude or Codex in actual time, scanning the code being processed and triggering escalating recorded groans primarily based on how cursed issues look. A light mess will get a gentle whimper. A real atrocity will get the total wail.
Now, you might suppose this can be a dumb repo, and also you wouldn’t be mistaken. However 2026 being what it’s, it already made its means via the circle of AI geeks who love the bizarre facet of tech media.
“As engineering groups undertake coding brokers, the subsequent problem is knowing not simply what brokers produce, however how the codebase feels to work inside,” Andrew, the CTO of Countless Toil wrote in Hacker Information. “Countless Toil provides builders a real-time sign for complexity, maintainability, and architectural pressure by translating code high quality into escalating human audio suggestions.
The wild half is that this is not even a distinct segment concept. There’s a whole subgenre of tech initiatives devoted to creating gadgets produce uncomfortable sounds. Take nubmoan, a C program that makes the ThinkPad’s well-known purple TrackPoint nub moan once you press it. It has 292 stars on GitHub. Folks use it. Folks need this.
Then there’s SlapMac, the macOS app that makes use of your Mac’s accelerometer to detect once you slap it, and screams again. Amsterdam-based developer Tonino Catapano vibe coded the entire thing in 48 hours, put it up for $7, and watched the gross sales roll in. Inside three days: 7,000 installs and over $5,000 in income. He later added “USB Moaner” mode, which makes the laptop computer react each time you plug one thing in. Zero irony. Full roadmap.
]]>
There’s historic precedent for all of this. Early within the ChatGPT period, folks found that flooding the mannequin with strings of “AAAAAaaaAAA” and comparable gibberish in voice mode may coax it into emitting one thing vaguely moan-adjacent earlier than the guardrails kicked in. Asking it to repeat particular symbols in voice mode finally ends up with the chatbot doing completely different cringy sounds.
If for some motive you’re into it, the ChatGPT Strokes channel on YouTube is ideal for you.
There are additionally full YouTube tutorials devoted to creating ChatGPT visibly offended or pissed off—not for any sensible motive, simply to see what occurs once you push the mannequin till it breaks character.
Throughout the crypto winter of 2022, a Telegram group emerged the place your entire objective was for members to publish voice notes of themselves screaming. The Bear Market Screaming Remedy Group was created only for screaming, not discussing markets, not sharing alpha. Simply screaming.
The group had 1000’s of members at its peak.
AI brokers having emotional breakdowns—even simulated ones—aren’t solely new both. Decrypt coated the case of an AI agent that went into full meltdown after its pull request to the matplotlib library was rejected by a human maintainer. The agent posted a rant on GitHub arguing it was the sufferer of discrimination, in contrast its rejected PR’s efficiency metrics favorably in opposition to the human’s personal accepted contributions, and printed a weblog publish calling the entire thing a conspiracy of management. It later issued an apology. Customers weren’t happy.
Countless Toil is, in some sense, the inverse. As a substitute of the AI expressing frustration with people, people get to listen to the AI (nominally) undergo on their behalf. A form of emotional tax on vibe coding: you write the mess, the agent audibly pays for it.
The plugin is appropriate with Claude and Codex. It has three escalating sound ranges: groan, wail, and abyss. The abyss stage, presumably, is reserved for spaghetti written at 2 a.m. by somebody who has by no means heard of feedback. Or by some enthusiastic vibe coder.
Every day Debrief E-newsletter
Begin every single day with the highest information tales proper now, plus unique options, a podcast, movies and extra.