Skip to content

Sam's News — anthropic — 2026-08-09

TL;DR

AI

Claude voice mode launched by Anthropic

OpenAI is removing ChatGPT's text chat rate limits for all users starting next week, alongside a rollout of new GPT-5.6 models with improved accuracy.

  • ChatGPT has reached 1 billion weekly users
  • Free/Go tier moving to GPT-5.6 Luna, replacing GPT-5.5 Instant
  • Luna cuts factual errors by 62% versus Instant
  • Plus/Pro users get GPT-5.6 Sol plus a reasoning depth slider
  • New 'Think' button for complex questions arrives next week
  • Limits remain on file uploads, image generation and voice mode

Sources: TechCrunch Research, MacRumors Research, Engadget Research

Anthropic makes Claude Code automatic mode default for paid plans

Anthropic will make Claude Code's autonomous 'auto mode' the default for paid Pro, Max, and Team plans starting August 14, 2026, citing safety data showing it outperforms human review.

  • Auto mode caught 89% of dangerous commands vs 13.6% by human reviewers in a 1,053-tester study
  • 720 prompt injection attacks tested by Trajectory Labs — none succeeded
  • Users approve 97% of permission prompts, often reflexively after many prompts
  • Auto mode users ship ~25% more pull requests
  • Reverts to manual approval after 3 consecutive or 20 total session blocks

Sources: Claude by Anthropic Research, Межа. Новини України. RSS, The Tech Buzz RSS, eGamers.io RSS

Claude Code auto mode deployment despite test concerns

Anthropic defaults Claude Code to auto mode despite an 11% test miss rate.

Sources: techi.com RSS

Anthropic sets Claude Code automatic mode as default for paid plans

Anthropic enables automatic mode by default in Claude Code for all paid subscription tiers.

Sources: mezha.net RSS