Anthropic Sets Claude Code’s Auto Mode as Default

Artificial‑intelligence startup Anthropic announced that its coding assistant Claude Code will now launch with auto mode turned on by default. The shift means developers will see the AI write and apply code without the need to manually approve each suggestion, unless they explicitly switch to manual mode.

In earlier releases, Claude Code offered two distinct operation styles. Auto mode let the model directly insert generated snippets into the project, while manual mode required a human review before any code was committed. With the latest update, every new session starts in auto mode; users who prefer a more hands‑on approach can toggle the setting in the preferences panel.

Anthropic frames the change as a step toward higher reliability and productivity. Internal benchmarks show the latest Claude Code version produces correct, runnable code in roughly 92 % of cases, a 15 % improvement over the previous generation’s error rate. Those numbers translate into faster prototyping, reduced boilerplate work, and less time spent on repetitive coding tasks.

Industry analysts, however, warn that defaulting to full automation carries its own set of risks. Fully AI‑generated code can sometimes embed security flaws or performance bottlenecks that escape casual review. Anthropic therefore advises users to still perform code inspections, especially for critical components, and to keep an eye on any unexpected behavior.

The move underscores how deeply AI is being woven into software development pipelines. By offloading routine coding to Claude Code, engineers can concentrate on architecture, design, and problem‑solving, while organizations benefit from quicker delivery cycles and lower development costs. Anthropic plans to roll out the default‑auto update to all existing customers over the next month and will iterate based on community feedback.

Source: TechCrunch

ETİKETLER