Zed now predicts your next edit with Zeta, our new open model (zed.dev)

Article summary

Zed has introduced a new feature called edit prediction, powered by an open-source model called Zeta, which predicts the user's next edit and allows them to apply it by pressing tab. The model is derived from Qwen2.5-Coder-7B and is fully open source, including an open dataset. Zed's edit prediction feature is currently available for free during its public beta. The feature aims to make the editing experience feel instant by anticipating the user's next move.

Main themes

• Edit prediction
• Open-source model
• Zed features
• Code editing
• AI integration
• Latency optimization

What commenters say

• The use of the tab key for accepting edit predictions can be problematic, as it can conflict with existing uses of the tab key, such as indenting code.
• Some users prefer to manually indent code, even with the presence of automated formatting tools, and may encounter issues with AI-powered edit predictions.
• The open-sourcing of the Zeta model may seem contradictory to making money, but it can be beneficial for users who want to run the model locally or contribute to its development.
• The need for a modern GPU to run Zed smoothly is a limitation for some users, who may not have access to up-to-date hardware.
• The use of indentation to denote code structure in languages like Python can lead to errors and inconsistencies, but some users argue that this is not a significant issue in practice.
• The integration of AI-powered edit predictions into code editors can be useful, but it requires careful consideration of user interface and workflow to avoid conflicts and errors.