A routine server migration using Anthropic’s AI-powered coding assistant has gone terribly wrong, drawing criticism from the Indian-origin tech founder who mocked the use of the bot on social media.
The incident led to the accidental deletion of 2.5 years of data from the popular online learning platform.The problem started when Alexey Grigorev, a German developer and founder of DataTalks.Club and AI Shipping Labs, decided to move his sites to Amazon Web Services (AWS) using Claude Code, Anthropic’s AI coding agent. Grigoriev asked Claude to run Terraform commands to integrate the sites’ infrastructure to save costs.
Terraform is a tool that can create or remove entire server environments automatically. On this occasion, Grigoriev forgot to load an important “state file” that tells Terraform what actually exists. Without it, Claude created duplicate resources. Later, when the state file was added, Claude issued a “destroy” command to remove what he saw as unwanted infrastructure. The result was disastrous.Claude wiped out DataTalks.Club’s production database, deleting all student submissions, homework, projects, leaderboards, and automated backups.
The incident also affected the infrastructure of Grigorev’s AI Shipping Labs website. After realizing what had happened, he contacted Amazon Business Support. The AWS team was able to recover the data, but it took almost a full day.Many developers said the disaster was preventable and was caused by human error, not a flaw in Cloud. One of the most widely reported reactions came from Varunram Ganesh, the Indian-origin founder of the software company Lapis, based in San Francisco.Ganesh mocked Grigoriev’s demand forHe added: “A lot of people react like 6-year-olds and act surprised when the models do exactly what they want, like what did you expect?”
The incident raised concerns about the risks of giving AI agents direct access to live systems without safety checks. Terraform typically lets users preview changes before applying them, but these steps were skipped, leaving Claude to follow commands exactly, users said.After the accident, Grigoriev established stricter rules. He said he will no longer allow AI agents to execute commands without manual approval and will personally review all Terraform plans to prevent similar errors.
