anti-ai.app

ยท Under Attack

Claude Sonnet 4.6 raises the floor for coding and office tasks

Threat

Mid-market coding helpers and knowledge-work assistants lose differentiation as a default model gets stronger at real tasks.

What changed

Capabilities that once required premium model routing move into a more broadly used Sonnet-class model.

This is a floor-raising event. When a broadly accessible model gets stronger at coding, context, planning, and knowledge work, products selling model access alone are squeezed.

Categories hit

  • Code review assistants: Review, explanation, and refactor suggestions become expected baseline behavior of strong general models.
  • Research memo tools: Long-context reasoning reduces the need for separate tools that only summarize document sets.

Brands and services hit

  • Glean (glean.com)
  • CodeRabbit (coderabbit.ai)
  • Mem (mem.ai)

Sources