Public
- Public
- Network
- Groups
- Featured
- Popular
- People

Sora Minazuki @soraminazuki · 4 hours ago There is a risk that everything I type is being sent to someone else's server somewhere There is a risk that some LLM-generated random spew is typed into my shell and executed ?! (possibly as root) This is completely false. There is no invasion of privacy or risk of commands being executed without consent. Looking at the release note, wiki, and the actual code, for data to be sent to OpenAI, you have to purchase an API token, set it up, and explicitly type in a question in a dedicated text box. By default, only the question you type and the name of your Shell (e.g., zsh) and OS (Darwin) will be sent. https://gitlab.com/gnachman/iterm2/-/blob/master/sources/iTermPreferences.m#L200-202 The generated commands also won't be executed automatically unless you explicitly check the box "Run commands automatically" and ignore a popup warning against doing so. https://gitlab.com/gnachman/iterm2/-/blob/master/sources/ToolCodecierge.swift#L679-692 From the onboarding UI: UI As much as I don't personally care for AI features, that this is an security or privacy risk that people can accidentally run into is just not true. Edited 4 hours ago by Sora Minazuki

Download link

Sora Minazuki @soraminazuki · 4 hours ago There is a risk that everything I type is being sent to someone else's server somewhere There is a risk that some LLM-generated random spew is typed into my shell and executed ?! (possibly as root) This is completely false. There is no invasion of privacy or risk of commands being executed without consent. Looking at the release note, wiki, and the actual code, for data to be sent to OpenAI, you have to purchase an API token, set it up, and explicitly type in a question in a dedicated text box. By default, only the question you type and the name of your Shell (e.g., zsh) and OS (Darwin) will be sent. https://gitlab.com/gnachman/iterm2/-/blob/master/sources/iTermPreferences.m#L200-202 The generated commands also won't be executed automatically unless you explicitly check the box "Run commands automatically" and ignore a popup warning against doing so. https://gitlab.com/gnachman/iterm2/-/blob/master/sources/ToolCodecierge.swift#L679-692 From the onboarding UI: UI As much as I don't personally care for AI features, that this is an security or privacy risk that people can accidentally run into is just not true. Edited 4 hours ago by Sora Minazuki
https://files.mastodon.social/media_attachments/files/112/480/823/398/215/519/original/44b4b901f2f3ae42.png

Notices where this attachment appears

Embed this notice
Dan Luu (danluu@mastodon.social)'s status on Wednesday, 22-May-2024 08:16:28 JST Dan Luu

Wow, people really don't like iTerm2 adding an optional AI integration which requires you to enter your OpenAI key to use, calling it "no longer fit for purpose", etc.
Someone pointed out that this feature is optional and not only has to be enabled, but it requires you to enter a key to use. That user was, apparently, reported on gitlab and is now blocked.
https://gitlab.com/gnachman/iterm2/-/issues/11470
It reminds me a bit of how some people don't like "algorithms", as in https://x.com/danluu/status/983466330320965632.

In conversation about 6 months ago from mastodon.social permalink