Open Slopware

Free/Open Source Software tainted by LLM developers/developed by genAI boosters, along with alternatives.

The intention of this list is to raise awareness of AI/LLM usage in popular open-source software. Provided below is an informed set of AI-free alternatives for users and developers to consider should their ethical boundaries be crossed or tolerance for risk be exceeded. This list is not a resource to be used for the harassment of other open-source developers. If you wish to advocate for the cessation of use and/or removal of AI-generated code from another project, we ask that it be done respectfully and constructively.

For more information on why we think this list matters, please see Why not LLMs?.

To Contribute, checkout our CONTRIBUTING.md.

Note

This is a fork of a repo by the same name maintained by the @gen-ai-transparency org on Codeberg. If they make their repo available again, we will contribute back.

• Definitions
  • Permissive AI Policy
    • Vibecode
  • AI Functionality
  • AI Code Reviews
  • Gen AI 'Art'
  • AI Sponsored
  • AI Databroker Usage
  • AI In Issue Tracker
  • Request for Help
• Sibling Repos
• Tainted Software
  • Audio
  • Browsers
    • Desktop Browsers
    • Mobile Browsers
    • E-Mail Clients
  • Cheat sheets and documentation
  • Command-Line Shells
    • Addons and Prompts
  • Search Engines
  • Mapping
    • Alternatives
  • Computer Graphics
    • Image Manipulation
    • Photography
    • Video
  • Terminal Emulators
  • Developer Tools
    • Compiler Tools And Related
    • Terminal Multiplexers
    • Version Control
  • Document Viewers
  • Programming Languages
  • Editors
    • Note Software Alternatives
    • Alternative TUI Code Editors
  • Accessibility Tools
  • Microcontroller Tools and Libraries
  • Service Monitoring
  • Programming Language-Specific Tooling
    • C
    • C++
    • CSS
    • JavaScript And TypeScript
    • Python
    • Rust
  • Gaming
    • Game engines
    • Game managers
    • Emulators
  • Language and Localization
  • Messaging
  • Package Managers
  • Password Management
  • Git Servers
  • Blogging Software
    • Content Management Systems
    • Static Site Generators
  • Services and Utilities
    • File Storage
    • Social Media
      • Servers
      • Clients
    • Unified Push Distributors
    • Budgeting
  • System Services and Daemons
  • Bootloaders
  • Operating Systems
  • Virtualization
    • Alternatives
  • VPN Software
    • VPN Software Alternatives
  • Window Managers
• Why not LLMs?
  • Legal Ramifications
    • Legal Cases And Law Problems
    • License Problems
  • Stolen Training Data
  • Environmental Impact
  • Labor
  • Poor Code Quality
    • Deskilling
    • Infosec risks
  • Healthy and Safety
    • Ties to the War Industrial Complex
    • Effects on Policing
    • Maintainer Fatigue
  • Effect on Hardware Prices
• License

Definitions

Term	Definition
LLM	Large Language Model, a language model trained on vast amounts of text (often entire websites without the website owner's permission)
slop	Anything generated by AI, typically code, books, articles, or images.
slopshop	Company, Organization, or Group of sloperators working together to use AI to generate content.

Below are special definitions that are used for tagging purposes.

Permissive AI Policy

A policy that permits the use of AI/LLMs in any capacity or is declared to be vibecoded. Both vibecoding and opening the door for people to vibecode count as a permissive AI policy.

Evidence can be:

• an explicit AI policy (e.g. AI_POLICY.md, CONTRIBUTING.md, developer docs) in the repo or something in the project's contributing guidelines or in their website's documentation that says that AI/LLMs are allowed
• an LLM friendly AGENTS.md, CLAUDE.md, or other such LLM instruction files or folders.
• core maintainers' blog or social media post about vibecoding
• link to readme, website, or documentation stating the project is vibecoded

Important

If a core maintainer has noted that the code is entirely vibecoded, please put that as the first evidence link.

Vibecode

Asking an AI to write software for you (often without human review). Per Wikipedia:

In computer programming, vibe coding is a software development practice assisted by artificial intelligence (AI) such as by chatbots (programs that simulate conversation) or AI agents such as Codex or Claude Code. The software developer describes a project or task in a prompt to a large language model (LLM), which generates source code automatically. Vibe coding may involve accepting AI-generated code without reviewing the output thoroughly, instead relying on results and follow-up prompts to guide changes.

The term was coined by Andrej Karpathy (co-founder of OpenAI) in February 2025.

A Vibecoder is a person who uses AI prompts to generate anything, typically code or images. It would therefore be appropriate to call Andrej Karpathy a vibecoder.

Note

Humans reviewing vibecoded code does not disqualify said code from being considered vibecoded.

AI Functionality

This would include the ability to enable or disable an AI feature, such as an AI assistant or AI summary feature. This would also cover the case when a project depends on an AI program or library, such as an LLM SDK.

Evidence can be:

• a link to the docs explaining the AI Functionality
• a blog post advertising the AI Functionality
• a link to the default branch in a repo that shows where the AI Functionality is implemented

AI Code Reviews

An AI agent/bot is used to review pull requests, which makes slop acceptance more probable and maintainability more difficult, especially when human reviews are scarce.

Evidence can be:

• a link to an AI code review requested by a maintainer.
• an AI code review CI workflow.
• a PR where an AI code review seems to be automatically created by some sort of third party app

Gen AI 'Art'

AI "Art" is being used in a project. This is typically for banners, avatars, promotional material, and/or blog posts. This means they're using tools that steal art from visual artists.

Evidence can be:

• link to the place where the AI art is used

AI Sponsored

The project is sponsored directly by an AI company.

Evidence can be:

• Blog post disclosing AI funding
• link to website that lists AI companies in sponsors list
• README that says "Sponsored by "

AI Databroker Usage

The use of an AI company as a third party service for processing user data.

Evidence can be:

• Blog post talking about using AI for processing user data
• Privacy policy listing an AI company as a third party service

AI In Issue Tracker

The use of an AI bot/agent for responding to Issues, Discussions, Forums etc related to a project's issue/bug tracking.

Evidence can be:

• Link to AI in an Issue, Bug report, Discussion, etc - must be related to the project's bug/issue tracking

Request for Help

This means someone is asking for help with research. They would like it if you found the last known good version of a project or suggestions on alternatives for people to use. You can submit a pull request to add this info.

Sibling Repos

This is a section for repos that are similar to this one either because they are also forks of the original upstream, or because they align with our goal of identifying and avoiding AI/LLM usage OR identifying/elevating projects that do NOT use AI/LLMs. These repos are not all maintained by the same people, but we may share some contributors and are generally kind FOSS neighbors. ✨

• llm-afflicted-software offers a similar list to ours here, however in YMAL file format for each category, making it more suitable to consumption via code (for instance writing a CLI, GUI, or TUI tool to consume it). They've also kindly reached out in the past and submitted Issues and PRs (e.g. #12, #22) to help us here.
• forge.starlightnet.work/Team/No-AI/ which is the repo for this no-AI list that lists projects that have explicitly pledged not to use AI or have an explicit policy to not use AI.
• slop-free-index is a list of software that has taken steps to reject AI in its development processes.

Tainted Software

Warning

We've recently added a "Last Untainted Version or Commit ID" section to our below tables. This is meant for tech savvy individuals to be able to fork the project and continue maintaining it without the use of AI. Choosing to install an older version of software could expose you to to security risks over time. Only do this if you are able to accept such risks.

Audio

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
HISE		(1)	DPF
zrythm		(1)	Ardour1
FineTune		(1)	BackgroundMusic
SimpMusic		(1) (2) (1)

1: Status as a replacement may be debatable.

Browsers

Desktop Browsers

Name	Last Untainted Version or Commit ID	Tags and Evidence
Chromium/Blink engine		(1, 2, 3)
Firefox		(1, 2) (1)
Ladybird		(1)
WebKit		(1, 2, 3)
Zen		(1, 2) (1)

Note

Desktop Browser Alternatives

• Epiphany (by GNOME, WebKit based) does not currently have a stance, but GNOME developers as a whole show aware anti-"AI" sentiments and roll out wide reaching rules against it.
• Servo (New engine, formerly owned by Mozilla, now by a co-op) is not yet daily driver ready, but has strong protections for when it is some time in the future.
• waterfox They have clear anti-ai stance

Mobile Browsers

Name	Last Untainted Version or Commit ID	Tags and Evidence
Firefox		(1, 2) (1)

Note

Mobile Browser Alternatives

• IronFox For Android, while not officially related to LibreWolf, some of the same developers work on it.

E-Mail Clients

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
NeoMutt		(1)	mutt

Cheat sheets and documentation

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
tldr-pages	v2.3	(1, 2)
cheat	v4.4.2	(1)

Command-Line Shells

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Oil Shell	dddab4a	(1)	dash
nushell		(1) (1)	elvish shell
PowerShell		(1, 2, 3) (1)
xonsh		(1)
yash		(1)	dash

Addons and Prompts

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
oh-my-posh		(1)
oh-my-bash		(1)
ohmyzsh		(1) (1, 2, 3)
Starship		(1)

Search Engines

Name	Alternatives	Evidence	Notes
Searxng	LibreY	Copilot used in reviews	Note: maintainers experimenting with and open to adding AI results, see 1, 2, 3

Mapping

Name	Last Untainted Version or Commit ID	Tags and Evidence
CoMaps		(1)
Organic Maps		(1)
OsmAnd		(1)

Note

Mapping Alternatives

• Navit for car navigation
• Open Street Map provides lists of clients on its wiki. For example, other mobile clients can be found on the Android page.

Computer Graphics

Image Manipulation

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
ImageMagick	v7.1.2.16	(1)	GraphicsMagick
excalidraw		(1)	LibreOffice Draw
tldraw		(1)	LibreOffice Draw
draw.io		(1)	LibreOffice Draw

Photography

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
RapidRAW		(1)
RawTherapee	5.12	(1) (1, 2, 3)	5.12

Video

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
VLC		(1)	Showtime (gstreamer-based) SMPlayer (mplayer-based) ffplay (ffmpeg+SDL-based)
gstreamer		(1)
mpv		(1)	MPlayer Showtime (gstreamer-based) SMPlayer (mplayer-based) ffplay (ffmpeg+SDL-based)
Kodi		(1) (1)
Jellyfin (server)		(1) (1)
Jellyfin Desktop	Use the (now archived) Qt-based client	(1) (1, 2)
PipePipe		(1)	LibreTube
NewPipe		(1)	LibreTube

Terminal Emulators

Name	Last Untainted Version or Commit ID	Tags and Evidence
iTerm2		(1)
Ghostty		(1)
Kitty		(1, 2, 3)

Note

Alternative terminal emulators

• Alacritty or the Alacritty fork with sixel
• Contour Terminal
• COSMIC terminal
• Deepin Terminal
• foot
• GNOME Terminal
• Konsole
• MATE Terminal
• Rio Terminal
• xfce4-terminal
• XTerm

Developer Tools

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Atuin		(1, 2) (1)	Built-in shell history
Nextest		(1)	Just use cargo test
Golly		(1)
Just		(1)	A Make implementation
Mold linker		(1)	GNU ld
Taskfile		(1)
Wild linker		(1, 2, 3)
curl		(1, 2, 3) (1) (1)	Wget
rsync	version ≤3.4.1	(1, 2)	version ≤3.4.1, scp or a FTP client
npmx		(1, 2, 3) (1)	npmjs

Note

On task runners and make(1) implementations:

A lot of Make implementations exist that are very fit to replace other task runners. In fact so many were produced, that one can just look up "GNU Make alternative" and find >10 results to pick the one they like the best. Also see: wikipedia/Make Here are some Make implementation suggestions:

name	description
GNU Make	the most complete one, extra extensions, guile scripting, exports, etc.
NetBSD make	Make used to build NetBSD ports tree
Jam	OSS rewrite by Perforce

Compiler Tools And Related

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
LLVM		(1)	GCC, Cranelift, pcc

Terminal Multiplexers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
tmux		(1)	dtvm, mtm, GNU Screen
zellij	1411b2d9	(1)	dvtm, mtm, GNU Screen

Version Control

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Sapling		(1, 2)	Game of Trees
Git		(1, 2)	Game of Trees
Jujutsu		(1, 2, 3, 4)	Game of Trees

Document Viewers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Sumatra PDF	v3.5.2	(1)	Xournal++

Programming Languages

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
.NET	v9.0.0	(1, 2, 3) (1)	See below
Julia	v1.11.4	(1 (1, 2)	See below
Kotlin	v2.3.0	(1, 2) (1)	See below
mruby	v3.3.0	(1) (1, 2)	See below
Nim	e2a3063	(1, 2, 3)	Ada?
Python	v3.14.0a4	(1, 2) (1)	See below
MicroPython	1.26.1	(1, 2, 3)	See below
CircuitPython	9.2.9	(1, 2, 3)	See below
Ruby	v.3.3.8	(1, 2) (1)	See below
Scala	v3.3.6-RC1	(1, 2, 3)	See below
TruffleRuby	v33.0.0	(1, 2) (1)	See below
Typescript	v5.1.3	(1, 2, 3)	See below

Note

On alternative languages:

Suggestions for alternative programming languages are necessarily more complex than many other categories considered here, as that's one of the core technical choices to be made when starting a new software development project.

Depending on your needs and technical concerns, you may find any number of different alternative languages useful. For example, when considering alternatives to .NET, Java may be an interesting candidate from the perspective of being a similar VM-based framework.

• Clojure - dynamic, compiled, general-purpose programming language: AI-generated code not accepted
• Java - memory safe, object-oriented, VM-based programming language: AI-generated code not accepted
• Zig - general-purpose, compiled, system programming language: Strict No LLM / No AI Policy

Editors

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Helix		(1)	See editors below
Joplin		(1)	See note software below
Logseq		(1)	See note software below
Neovim		(1, 2, 3)	See editors below
Qt Creator		(1)	See editors below
VIm	v9.1.0189	(1, 2, 3)	See editors below
VS Code		(1) (1, 2, 3)	VSCodium's builds disable AI "features"
Zed		(1) (1, 2, 3)	Lapce Gram (fork) Others (emacs, etc.)

Note

Note Software Alternatives

• Zettlr (Markdown-based)

Note

Alternative TUI Code Editors

• emacs
• EVi
• vim-classic
• Other forks of VIm prior to March 20th 2024, build v9.1.0190 or commit 8950bf7f8b85c1287d4e696965d88091fcc60594

Accessibility Tools

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
espeak-ng		(1, 2) (1, 2)	espeak (note: heavily outdated and very bad from a modern standpoint), flite
Nametag		(1, 2, 3)	Monica

Microcontroller Tools and Libraries

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
FastLED	3.10.1	(1, 2, 3)
ch32-hal	07687b9	(1, 2, 3) (1)	Official WCH C SDK's, ch32fun

Service Monitoring

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Uptime Kuma		(1)
Glitchtip		(1)

Programming Language-Specific Tooling

C

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
harfbuzz	12.3.2	(1,2,3)
sd-bus (via systemd)		See systemd section	basu
zlib-ng		(1,2,3) (1)	zlib

C++

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
spdlog	v1.16.0	(1, 2) (1)
SQLiteCpp	v3.3.3	(1, 2)	SQLite's first-party interface

CSS

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
tailwindcss		(1) (1)	twind

JavaScript And TypeScript

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
ArkType		(1, 2, 3) (1)	TypeBox (in grey area)
Electron		(1, 1, 2)	Using native GUIs, such as GTK, instead of making a web site
nvm	before July 2025?	(1, 2)
Valibot		(1, 2) (1) (1)	TypeBox (in grey area)
Zod		(1, 2, 3) (1)	TypeBox (in grey area)

Python

Note that Python is itself tainted; see the Programming Languages section.

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Chardet	6.0.0.post1	(1, 2)	charset_normalizer
JupyterLab	v4.5.6	(1) (1)	Possibly Emacs (1, 2)
Pygments	2.19.2	(1)
Requests		(1)	urllib.request in Python Standard Library
Textual	v7.5.0	(1) (1)	pytermgui

Rust

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
c2rust	0.21.0	(1,2) (1) (1,2,3)
egui	0.31.1	(1,2,3) (1,2,3)	imgui, relm, iced
facet	cad0de9 as Shapely before it got rebranded to facet-rs	(1,2,3) (1,2,3)	serde (note: tainted by its Palantir-defending author), nanoserde If possible avoid all-in-one libraries for deserialization and avoid macros that slow things down.
ratatui	0.29.0	(1,2) (1,2)	iocraft (<=0.7.17), cursive, console
iocraft	0.7.17	(1,2) (1, 2)	cursive and console
rust-analyzer		(1,2) (1,2,3)	cargo check
wgpu		(1, 2, 3) (1,2,3)	vulcano, ash, glow
zbus		(1, 2) (1)
zerocopy		(1,2)

Note

Also, there is an advisory database of Rust slopware crates, which you can use in tools like cargo-deny: https://codeberg.org/1millibyte/rust-slopdb.

Language and Localization

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Weblate		(1, 2, 3)

Messaging

Name	Last Untainted Version or Commit ID	Tags and Evidence
Converse.js		(1)
Mattermost		(1) (1)
Psi		(1)
Telegram		(1) (1)
Zulip		(1)
Sharkord		(1, 2)
ArcaneChat	None	(1) (1)
Delta Chat		(1 (1, 2, 3)

Note

Messaging alternatives

• matrix

Package Managers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
brew		(1)	Build packages from source
macports		(1)	Build packages from source
PiKISS		(1)	Use system repositories provided by your distro or manual installation of software

Password Management

Name	Last Untainted Version or Commit ID	Tags and Evidence
AliasVault		(1)
Bitwarden		(1)
KeepassXC		(1, 2, 3)

Note

Password manager alternatives

• vaultwarden (server) / rbw (client)
• KeePass 2. It is a .NET application, but Mono or Wine can be used for non-Windows platforms.
• KeePassXC 2.7.9 was released before the statement, and was awarded 3 year security Visa by the French National Cybersecurity Agency (ANSSI) for a First-level Security Certification (CSPN), valid in France and Germany, under report No. ANSSI-CSPN-2025/16 (archive.org, certificate).
• KeePassChi: A fork of KeePass 2.7.9 with clear stance against LLMs.
• Secrets uses the KeePass file format.
• Password managers based on pass or Himitsu could be interesting.

Gaming

Game engines

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
UPBGE 0.5	0.36.1	(1)
Godot Engine		(1, 2)	Pandemonium Engine
GZDoom		(1)	UZDoom, Chocolate Doom
GDevelop		(1, 2, 3, 4, 5)	Pandemonium Engine

Game managers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Lutris		(1, 2)	Bottles

Emulators

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Azahar	2124.3	(1)

Git Servers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Gitea		(1, 2, 3) (1)	Forgejo (self-hosted) or Codeberg (hosted)
GitLab		(1)	Forgejo (self-hosted) or Codeberg (hosted)

Blogging Software

Content Management Systems

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Ghost		(1)	WriteFreely
Wagtail		(1) (1)	WriteFreely
WordPress		(1, 2) ( 1)	WriteFreely
Bludit		(1) (1)	WriteFreely

Static Site Generators

Name	Last Untainted Version or Commit ID	Tags and Evidence
Hugo		(1, 2, 3)
Pelican		(1)
Zola		(1)

Note

Static site generators alternatives

• Bridgetown
• BSSG
• Jekyll maybe? No signs of AI in the repo but also no clear statement provided
• Haunt
• Zine

Services and Utilities

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Calibre	8.15.0	(1)	BookLore Clbre, a Calibre fork without AI Arcalibre, a Calibre hard-fork from before AI features
somafm_tui		(1, 2)	soma-player

File Storage

Name	Alternative(s)	Evidence
MinIO	Garage	Advertises AI as their primary focus on their website.
Nextcloud	seafile for file storage and syncing only (not a complete worksuite replacement)	✨ Nextcloud Assistant (can be disabled) Nextcloud Desktop has AGENTS.md
SeaweedFS	Garage for S3 api + JuiceFS for the NFS mounts and K8s CSI driver Rook/Ceph	Using Gemini and coderabbitai in PRs.

Social Media

This is for both social media websites and apps.

Servers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Bluesky		(1)	See below
rsky		(1, 2, 3)	See below
PeerTube		(1) (1)
Mastodon		(1)	See below
Misskey		(1)	See below
Lemmy		(1)

Note

Generic social media server alternatives

• GoToSocial
• Wafrn supports both ActivityPub and ATproto
• snac2
• Iceshrimp.NET
• Epicyon

Clients

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
IceCubesApp for iOS		(1)	Feditext (an iOS fediverse app in beta), wafrn-rn
Bluesky Social App		(1) (1, 2, 3)	wafrn-rn

Unified Push Distributors

Name	Alternative(s)	Evidence
ntfy-sh	see below	v2.18.0 is "14,997 added lines of code [...] written by Cursor and Claude" (archive)

list of unified push distributors at: https://unifiedpush.org/users/distributors:

• Sunup
• NextPush (see entry for NextCloud on taintedness)
• Conversations (XMPP client) or one of the forks

Budgeting

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Actual		(1, 2)
beancount		(1, 2)
Firefly III		(1)
hledger (v2.x)		(1)
ledger-cli		(1, 2)

System Services and Daemons

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
framework-system	50a18e8	(1, 2)	BIOS settings, UI widgets, /proc
lvm2		(1,2,3)
rsyslog		(1)	syslog-ng
systemd		(1, 2)	dinit, s6, OpenRC, shepherd
wireplumber		(1)
Mesa3D		(1)

Bootloaders

Name	Alternative(s)	Evidence
Limine	Hyper, GRUB	This entry was added by the creator and main maintainer of Limine. Limine does not ban code authored by LLMs or anyone/anything else, as long as it's reviewed by a human. The author has used LLMs (Claude Code) numerous times for many commits.

Operating Systems

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
DragonFlyBSD		(1)	See below
FreeBSD		(1)	See below
FreeDOS		(1) (1, 2)	The original MS-DOS source code is under the Expat/MIT license
GNU Mach - kernel for GNU Hurd		(1)	See below
Linux Kernel		(1) (1)	See below
MidnightBSD	ed4b0a5	(1) (1) (1, 2, 3)	See below
NixOS		(1, 2, 3)	GNU Guix System
OpenBSD		(1, 2)	See below
ReactOS	d4993c6	(1) (1, 2, 3)	Wine for some use cases
TrueNAS		(1)

Note

Alternative operating systems

Operating systems are a tough ask. No matter what you choose, support will be inferior compared to an operating system with the Linux kernel, so don't feel pressured to switch.

• MirBSD has banned LLMs
• NetBSD has a somewhat ambiguous policy
• illumos and its distributions remain unaffected as of April 2026.
• RedoxOS has banned LLMs
• Haiku has banned LLMs

Virtualization

Name	Last Untainted Version or Commit ID	Tags and Evidence
Botspot Virtual Machine		(1)
VirtualBox		(1)

Note

Virtualization Alternatives

• QEMU: Emulator and virtualizer that does not accept slop contributions.
  • virt-manager: GUI for managing libvirt machines. Needed for VirtualBox feature parity as QEMU does not provide one of its own.
• GNOME Boxes

VPN Software

Name	Last Untainted Version or Commit ID	Tags and Evidence
Headscale		(1, 2, 3)
netbird		(1)
nebula		(1)
frp	0.48.0	(1, 2) (1, 2, 3) (1)

Note

VPN Software Alternatives

• Use WireGuard^®️ capable VPN's such as those listed in the official wireguard repos.

Window Managers

Name	Last Untainted Version or Commit ID	Tags and Evidence	Alternative(s)
Hyprland		(1)
i3		(1, 2)	wmii, dwm, awesome

Why not LLMs?

Could use some help writing this with concrete receipts on environmental, social, political, and economic/labor harms.

Legal Ramifications

LLMs are often trained on, and thus prone to, regurgitate either completely, or in-part, chunks of code that are licensed under terms which have specific legal requirements that a sloperator may not understand or even be aware of when making a contribution. Regardless of this ignorance, it falls to the repo's owner to comply with the terms of any and all licensed code integrated into their project.

Legal Cases And Law Problems

• copilot litigations, IEEE article explaining how we go here
• broader lawsuits against AI companies tracker

License Problems

Legal, copyright and ethic problems arise especially with copyleft licenses such as (A/L)GPL. With the "help" of AI the copyleft code may be "license-washed" very easily.

There are ongoing problems with AI "license-washing" in the FOSS world:

• chardet - switched from LGPL to MIT license without asking all contributors (which itself is a violation of GPL)
  • relicensed release: https://github.com/chardet/chardet/releases/tag/7.0.0
  • original author's concerns: https://github.com/chardet/chardet/issues/327
  • "consumer's" concerns: https://github.com/chardet/chardet/issues/331

Stolen Training Data

AI companies use data from across the web for training their models, most often without the website owners' and users' consent. Big tech companies like Google and Meta are scraping data from the users of major FOSS projects, such as Mastodon, WordPress, and other ActivityPub-powered and self-hosted software.

• In 2023, the Washington Post published a list of sources in Google's C4 data set. A multitude of fediverse instances and personal sites were included. The fediverse is known for its userbase being major proponents of privacy and opt-in consent, making this especially jarring for those who have chosen to use decentralized social media for control over their data.
• In 2025, a similar leak of Meta's sources was published. Meta's list demonstrates how their integration of ActivityPub into their Threads software has enhanced their ability to mine content without authorization. Threads is widely blocked in some parts of the fediverse, but their scraping of server CDNs has allowed them to get around that. Notably, both the CDN domains of the managed hosting services masto.host and fedi.monster are included in the list; large servers like mastodon.art, which is hosted by the former and has many artists who've left sites like DeviantArt and others due to their AI scraping of user content, had media unknowingly scraped.

FOSS projects listed in this repo are using tooling that blatantly disregard licensing and violate of Codes of Conduct, making said tools antithetical to FOSS' purpose.

Environmental Impact

To start learning a bit more, you can checkout the wikipedia page on Environmental impact of artificial intelligence. We're very open to people contributing other explanations, links, and resources to learn more about this. Here's what we've gathered so far:

• MIT News: Generative AI’s environmental impact
• MIT Technology Review: We did the math on AI’s energy footprint. Here’s the story you haven’t heard.

Labor

AI usage and normalization contributes to labor violations in many ways that are obvious and some you may not be aware of.

On one hand, many things that you think are "AI" are actually humans in another country pretending to be an AI chatbot for you for either extremely low wages or in some cases, no wages e.g. prison labor. This is particularly common for "friend"/"sex" bots, but it is also extremely common in the image/video identification. You can find a bit more info at the following links:

• Long hours and low wages: the human labour powering AI’s development
• OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic
• 'AI Is African Intelligence': The Workers Who Train AI Are Fighting Back
• These Prisoners Are Training AI: In high-wage Finland, where clickworkers are rare, one company has discovered a novel labor force—prisoners.
• Amazon grocery stores previously reported to use AI actually used people in India

Poor Code Quality

Vibe coding / agentic workflows result in poorer code quality, and relaxed oversight practices. These effects may be compounded by the common practice of using additional LLM-based tooling to provide code-reviews.

• How I Dropped Our Production Database and Now Pay 10% More for AWS
• Claude Tested Everything Except the One Thing That Mattered (Ai agent refuses to follow explicit instructions to test createPost() in increasingly erratic ways)
• Amazon calls engineers for a “deep dive” internal meeting to discuss “GenAI”-related outages
• GitClear has released reports in 2024 and 2025 indicating a worsening of key code quality metrics correlating with increased LLM adoption.

Deskilling

There is increasing evidence to show that LLMs negatively impact developers' coding abilities:

• Brains show less activity when completing tasks with LLMs compared to completing tasks with search or completing tasks without digital help.
• Developers who use early-2025 LLMs reported higher subjective performance, but were measured to have lower objective performance. This gap between subjective and objective performance was considered notable.
• In an Anthropic study, learners using LLMs demonstrated lower learning rates on average compared to learners not using LLMs.
• A recent study uses the term "cognitive surrender" to describe the way humans tend to offload key critical thinking skills onto LLMs, even when the output is wrong.
• There are anecdotal reports of LLM users forgetting how to code.

Infosec risks

This also often results in massive security holes.

• Meta Security Researcher's AI Agent Accidentally Deleted Her Emails
• Moltbook’s ‘vibe-coded’ breach is the future of security failures
• In a study evaluating over 500k code samples, LLM-generated code was found to contain more high-risk security vulnerabilities than human-generated code
• LLMs make up package names, making them vulnerable to incorporating malicious code in "slopsquatting" attacks (Arxiv study)

Healthy and Safety

There's been a number of high profile incidents that have resulted in endangerment or death. Here's some examples:

• AI Toy Exposed 50,000 Logs of Its Chats With Kids to Anyone With a Gmail Account

• Content Warning: Suicide

  • Wikipedia: Deaths linked to chatbots
  • AI chatbot pushed a teen to kill himself, a lawsuit against its creator alleges

LLM use has also been linked to new-onset psychosis.

Ties to the War Industrial Complex

A lot of AI companies also work directly with nation states for use in their Departments of War (sometimes called Defense) which in turn leads to further AI usage during war and invasions. This is coupled with NYT: Palantir, Anthropic and small start-ups are generating rewards from their investments in defense tech.

As another example NPR: OpenAI announced Pentagon deal after Trump banned Anthropic which was due to the USA Department of War launching an AI acceleration strategy.

Due to the nature of LLMs being only kind of as good as the data they are trained on, this can lead to additional civilian deaths and housing/infrastructure damage either intensionally or not. Examples:

Content Warning: War details, death

• Gaza: UN experts deplore use of purported AI to commit ‘domicide’ in Gaza, call for reparative approach to rebuilding
• Lavender & Where’s Daddy: How Israel Used AI to Form Kill Lists & Bomb Palestinians in Their Homes
• Microsoft says it provided AI to Israeli military for war
• Google has dropped its promise not to use AI for weapons

All of this to remind you that if you use AI, you're helping to support these companies and the additional activities they participate in, outside of generative code or images.

Effects on Policing

Police have quickly embraced AI, which has already directly led to people being jailed for things they've never done. As examples:

• This Grandmother was jailed for 6 months after an AI error linked her to a crime in a state she had never even visited
• How Wrongful Arrests Based on AI Derailed 3 Men's Lives

This is, in part, due to companies such as Amazon Aggressively pushing police to use AI which they do through both facial recognition and offering compute for predictive policing. With regards to facial recognition, here's an example of how it too can lead to false arrests: Face Recognition on Flawed Data.

There have been warnings about AI in policing, particularly around racial bias, such as:

• Bennan Center of Justice: The Dangers of Unregulated AI in Policing
• OxJournal: Predictive Policing or Predictive Prejudice? A Study of the Legal, Historical and Ethical Implications of AI in Policing
• The Guardian: Police AI Chief admits crime fighting tech will have bias...
• A third of all Black children were flagged by a child services agency

Maintainer Fatigue

Having to deal with the onslaught of many LLM written pull requests and issues, causes real maintainer burnout that stagnates projects as maintainers become overwhelmed with half baked, poorly written, insecure code. Here's some examples:

• Overrun with AI slop, cURL scraps bug bounties to ensure “intact mental health”
• Godot maintainers struggle with 'draining and demoralizing' AI slop submissions
• An AI Agent Published a Hit Piece on Me (matplotlib)

Effect on Hardware Prices

The demand for construction and outfitting of new data-centers to host AI/LLM compute capacity has overwhelmed global supply and production of multiple hardware components. Memory, Storage, and GPUs have seen massive increases in price for both consumer and enterprise models upward of 400% in some cases.

The lack of supply has led large system-builders to purchase production capacity from OEMs well in-advance of delivery leading some manufacturers to end consumer-oriented product lines in favor of enterprise capacity.

The down-stream effects for consumers is that near all electronic devices which contain memory and storage will see their prices rise and availability decline. Those who already own existing electronics and computer hardware components may also find themselves unable to have their devices repaired or replaced under warranty.

• The RAM shortage is coming for everything you care about

• Western Digital is already sold out of hard drives for all of 2026 — chief says some long-term agreements for 2027 and 2028 already in place

• The 2026 storage crisis: Why AI data centers are hoarding every hard drive on the market

This all results in shrinking the pool of people who have access to building home computers for any purpose, from gaming to coding to home labs, which in turn makes the tech industry less diverse due to people who have been historically marginalized having less financial resources to learn the skills at home. When this is factored in with the price of college being unaffordable in many places, we will see a sharper decline in disabled people, people of color, women, and the queer community entering the tech industry.

License

This repository is licensed under the Creative Commons Attribution Share Alike 4.0 International license. Please see LICENSE.txt for more information.