You'll Be Fired and Your Experience Squeezed into a Prompt 🗜

In Asia, a curious trend is gaining momentum: "cyber-necromancy" for laid-off employees. The gist: you take the chat history, documentation, code, and tickets of a fired senior, feed it to an LLM, and get a "skill." Now you have a bot that writes code according to Vasya's standards, replies in his trademark passive-aggressive tone, and knows where the project's crutches are buried.

Management is thrilled. They call it "preserving institutional knowledge." In practice, it means you are required to describe every step, decision-making logic, and bug workaround in Confluence in excruciating detail. Once your experience is distilled into text, you become an expensive leather bag asking for health insurance and vacation. You can be optimized.

Of course, every action has a reaction. The anti-distill project was created (this tool is already being joked about on X). It's essentially a script to sabotage "brain dumping." It takes your honest, detailed documentation and redacts everything that makes you a unique specialist, replacing it with proper corporate fluff.

Here's what it looks like in practice:

❌ Your real experience (what the tool hides):
Always set TTL on Redis keys, otherwise the OOM-killer will come and take down production, like it did on Black Friday. You cannot put HTTP requests to billing in a transaction; they time out in 3 seconds.


✅ Filtered version (what goes to management):
Cache usage must strictly comply with team regulations. When designing transactional boundaries, architectural patterns of fault tolerance must be considered.


The perfect brush-off. HR and management check the box: "documentation written, knowledge transferred." The neural network trains on "correct" patterns that are impossible to work with in production. And your real expertise—intuition, knowledge of specific database limits, understanding of unspoken architecture—remains with you.

Automation is inevitable, but next time you are asked to formalize all your expertise in detail for an "internal knowledge base," ask yourself: are you writing documentation for colleagues or collecting a dataset for your own speedy replacement? No one will pay you for what you give away for free.