Stop conda from auto-activating base in every new shell
One line — conda config --set auto_activate false — takes (base) out of your prompt. Plus which file it writes, the older key name, and why this is more than tidiness.
Every new terminal window opens with (base) sitting in front of the prompt. Everyone who has installed conda has seen it, and it is only a setting.
One line to turn it off
conda config --set auto_activate false
Open a new terminal and (base) is gone.
All this does is write auto_activate: false into ~/.condarc. The shell hook that conda init added to ~/.zshrc or ~/.bashrc stays where it is, so conda and conda activate keep working — nothing is pressing activate for you any more, that is all.
Activate when you actually want it:
conda activate base # or whichever environment you need
To check, or to change your mind:
conda config --show auto_activate # current value
conda config --set auto_activate true # back to auto-activating
Older versions use a different key
The setting used to be called auto_activate_base; newer conda renamed it to auto_activate (the old name still works but reports as deprecated). If the line above gives you an unknown-key error, you are on an older build — use:
conda config --set auto_activate_base false
When in doubt, conda config --describe auto_activate tells you whether the key exists in your version and what it defaults to.
Why it is worth turning off
The prompt is the least of it:
which pythonstops lying. While base is active itsbinsits first onPATH, so the interpreter you believe is the system one, or the project venv, belongs to conda. That class of bug is expensive to chase, because every step looks correct.- Fewer fights with venv, pyenv and uv. An already-active conda environment stacked under another virtual environment leaves both convinced they are in charge.
- Slightly faster shell startup. Not much per shell, noticeable across a few dozen tabs a day.
The cost is remembering to activate. For a conda-only setup that is one extra step; for a machine carrying several Python toolchains at once, it is the step that buys you knowing where you stand.