From 140ca0c786e0b8144514d9fceb552bdfbb0dd58f Mon Sep 17 00:00:00 2001 From: y9san9 / Alex Sokol Date: Tue, 9 Jun 2026 21:24:52 +0300 Subject: [PATCH] feat(ai): initial statement --- AI.md | 18 ++++++++++++++++++ README.md | 2 ++ 2 files changed, 20 insertions(+) create mode 100644 AI.md diff --git a/AI.md b/AI.md new file mode 100644 index 0000000..1bab6e0 --- /dev/null +++ b/AI.md @@ -0,0 +1,18 @@ + +![made-by-humans](https://img.shields.io/badge/made_by-humans-green) + +# AI + +To understand our view on LLMs, it is first important to understand our mission. Friendly is free and opensource software. We don't chase numbers or sells and therefore we have no need to win any kind competition. We create because we want, not because we have to to stay afloat. Therefore software that we produce is a creative impulse that we, humans, perform to serve other humans like us. + +## Allowed + +We encourage the use of LLMs while learning new things and testing different approaches. LLMs sole purpose is to look back at things retrospectively, which pretty useful to search for information about our world, including but not limited to frameworks APIs, usage examples, specific tutorials on a matter, etc. + +If you don't blindly copy-paste stuff from your chatbot or let AI agent write code for you, nothing really changes. You will learn faster than before thanks to immediate information lookup, and with time you will not need LLMs to write code. + +## Forbidden + +We forbid use of LLMs to produce final code that you will blindly trust without manual review. When you have an occurring task of having LLM to write some piece of code for you, we believe it's a code smell and you should automate that part in a different way. By reducing boilerplate, simplifying domain, etc. + +Code itself is not a contribution, it's a liability. LLMs produce mediocre code that is a summarization of everything average there is in the world with its good and bad. The same problem arises as when there are too many code owners. You lose the ability to put your vision in the work you are doing which leads to slower development speeds and overcomplication of the codebase. Use of LLMs encourages further use of LLMs. diff --git a/README.md b/README.md index 653aad3..32b7de5 100644 --- a/README.md +++ b/README.md @@ -9,8 +9,10 @@ Issues are not used, any decisions are stored in a separate `fed` directory as p ## Abstract - [Project Idea](IDEA.md) +- [Phases until 2027](PHASES.md) - [Quarter News](news/README.md) - [Friendly Evolution Decisions](fed/README.md) - [Development Process](LABS.md) - [Active Laboratories](DEVELOPERS.md) +- [AI Statement](AI.md) - [API Documentation](api/README.md)