AI news for builders — what shipped, and why it matters.
Anthropic 3
The latest Anthropic news, each summarized in plain English with a take on why it matters for builders — pulled from across the labs, research, and press and updated every few hours.
Anthropic's AI model, Claude, has been updated to handle questions about its own training data and export controls. The update ensures Claude provides accurate information on these topics. This change is based on events that occurred after Claude's training data cutoff.
Why it matters This update allows developers building with Anthropic's models to rely on accurate information from Claude when handling sensitive topics like export controls, reducing the risk of incorrect or misleading answers.
Anthropic's Claude Code will automatically generate code without requiring manual input from developers, a feature previously available in 'auto mode'. This change applies to all users of the platform. The update is set to take effect shortly.
Why it matters Developers building with Anthropic's Claude Code can now automate more of their workflow, reducing the need for manual coding and potentially increasing productivity.
Anthropic is making auto mode the default setting for new sessions in most Claude Code plans starting August 14th. The change affects Pro, Max, and Team plans. Auto mode is intended to mitigate risks such as prompt injection and data exfiltration.
Why it matters This shift could reduce the risk of AI model exploitation by defaulting to a safer configuration for users who don't manually adjust settings.