{"body":"One of the main problems when artificial intelligence is the central brain of your company is giving it the best possible context. Decisions, conditions, rules, facts. All of this matters.\n\nToday I decided to create a system that asks questions. We put together a long list of questions about the narrow points in our system, where there may be contradictions, doubts or unclear areas. I answer those questions, work things out and record the decisions.\n\nThe decisions fall into two categories: permanent and temporary. Permanent decisions are the principles we follow. Temporary ones concern, for example, a particular piece of development that matters here and now, or an experiment.\n\nPermanent decisions will gradually accumulate. We will put them into a database and give the agents convenient access through an API. You cannot ignore this: the larger your project or company, the more of these things you need to describe.\n\nI have created a regular protocol for asking questions and making decisions. The narrowest and at the same time most powerful point in the interaction between a person and AI is the decision. People have to make decisions, and those decisions have to be good. When a person takes responsibility for them, they understand where they are heading and what result they want.\n\nThis is a very important part of it. No AI will ever decide this for you. And if it does, the result will already be something that is foreign to you.\n\nThere is a philosophy in this. Wherever artificial intelligence goes, the key decisions should stay with us. If someone decides how you should dress, what shoes to wear, what hairstyle, moustache or beard you should have, you will lose the desire to go outside and get dressed. Because you are deciding nothing.\n\nThat is why I think these decisions should stay with us. I am creating a database of such decisions, convenient access to it for agents, and I will gradually fill it. I think this could also work well as part of the digital brain of any company in the future.","content_language":"en","day_counter":{"label":"Day 386","number":386,"start_date":"2025-09-11"},"id":"2026-10-01-the-decisions-should-stay-with-us","project_date":"2026-10-01","published_at":"2026-10-01T15:24:27+00:00","retrospective":false,"schema_version":"crash_brief_public_blog_post_v0_1","slug":"the-decisions-should-stay-with-us","source_language":"ru","source_types":["telegram"],"status":"published","summary":"I am building a record of permanent and temporary decisions for AI agents, while keeping the key decisions and their responsibility with people.","title":"The Decisions Should Stay with Us","topic":"founder_post","type":"founder_post","updated_at":"2026-10-01T15:24:27+00:00","version":1,"visuals":[{"alt":"A person choosing a decision card beside an organized archive.","caption":"The decisions stay with people; accumulated context helps them make those decisions.","url":"/blog-media/2026-10-01/decision-context.png"}],"x_thread_url":"https://x.com/Amigov_AI/status/2105678637001576832"}