Killed By Claude Report
← HomeCurrent verdict
Assessment
The website content shows broad, horizontal AI assistant usage: chat, coding, search, writing, planning, translation, analysis, and business workflows. Anthropic’s corpus shows Claude expanding across the exact same surface area: coding agents, office suite integrations, connectors, long context, computer use, managed agents, code review, memory, charts, and enterprise deployment. There is no niche insulation visible here. This is direct platform-to-platform collision.
Biggest historical hit
Introducing Claude Opus 4.6 (Anthropic, Feb 5, 2026) is the clearest historical hit: Anthropic explicitly claims state-of-the-art performance across coding, agentic search, knowledge work, office tasks, long-context reasoning, and autonomous workflows—the same broad categories visible on OpenAI’s own homepage.
What still protects them
Only scale, distribution, brand, and existing product surface. OpenAI still has a massive installed base, strong consumer mindshare, its own coding/search/research stack, and adjacent products like Sora. But that is not moat from overlap; it is just having enough mass to survive a direct war. No differentiated wedge is visible from the submitted site content because the site itself advertises a generic everything-assistant.
Signals
Why this is in the blast radius
OpenAI’s homepage presents a broad assistant for coding, research, analysis, writing, planning, and business tasks. Opus 4.6 directly targets those same categories and explicitly outperforms on coding, knowledge work, search, finance, long-context reasoning, and autonomous task execution. That is direct product overlap, not adjacent competition.
Sonnet 4.6 pushes Claude further into the mainstream default-assistant slot: coding, computer use, long-context reasoning, knowledge work, design, and office tasks. OpenAI’s site copy is exactly a mainstream default-assistant pitch. Claude is attacking the same front door with a cheaper, broadly deployable model.
Managed Agents expands Claude from chatbot into deployable agent infrastructure for teams and developers. OpenAI’s site highlights coding, business workflows, product discovery, and Codex. Anthropic’s move compresses the stack from model provider to hosted agent platform, directly pressuring any similar platform ambitions.
Cowork turns Claude into a persistent work assistant for multi-step business tasks across functions. OpenAI’s homepage examples heavily emphasize everyday work help, planning, writing, analysis, and knowledge work. Cowork lands squarely in that same productivity territory.
OpenAI’s site markets a general work assistant. Anthropic adding Outlook, OneDrive, and SharePoint connectors makes Claude more embedded in the daily enterprise workflow, increasing switching pressure for general AI assistants competing for workplace usage and retention.
OpenAI’s site points users to coding help, debugging, SQL, Python, database schema design, and Codex. Claude Code plus computer use turns Anthropic into a stronger end-to-end coding agent that can write, run, click through UIs, test, and verify. That directly attacks OpenAI’s coding-assistant surface.