Exploring Codex’s New /pet and /goal Commands: Features, Developer Reactions, and Competitive Context

When a developer types the /pet command in the Codex terminal, a small virtual animal appears, a new optional feature announced by OpenAI that is not an illusion but an added layer of interaction.

The /hatch command further allows users to customize the pet’s appearance. One programmer remarked, "We let AI write code, and it gives us emotional support pets," while another observed, "The most advanced code assistant now has a virtual pet, either the best developer‑experience decision ever or proof that AI companies know programmers are lonely at 2 a.m."

Reactions vary: some developers appreciate the novelty, whereas others dismiss it as unnecessary, asking, "Does anyone really need this feature?" and questioning whether time is being wasted on an AI‑powered electronic pet.

According to OpenAI’s official documentation, the electronic pet feature is optional; developers can enable or disable it and also customize its look.

In addition, Codex CLI version 0.128.0 introduced the /goal command, which is described as much more practical because it lets developers set long‑term objectives and have the AI assistant track progress across conversation turns.

Developer Felipe Coury called /goal "a reinterpretation of the Ralph loop," designed to keep goals active across dialogue rounds. The feature, built by Pyright author Eric Traut, is currently available only in the CLI version and must be activated in the configuration file:

[features]
goals = true

The article then shifts to a broader industry view, noting that while developers have previously been "treading water" with OpenAI’s offerings, the competitive landscape is shifting with Anthropic emerging as a strong contender. Codex is portrayed as rapidly improving and attempting to replicate capabilities to keep pace with Anthropic’s aggressive approach, suggesting that a turnaround for Codex is possible.