Anthropic says the Claude app can now keep a chat going indefinitely, automatically summarizing earlier context when it hits its context window limit
Russell Brandom / TechCrunch : X: @danshipper , @shakeelhashim , and @claudeai Bluesky: @durumcrustulum.com and @chicagomike X: Dan Shipper / @danshipper : the thing that makes Opus 4.5 special is you can vibe code forever without it losing the plot. i vibe coded an iOS app this weekend that any frontier model could build in one shot. but the special thing about Opus is i kept adding features and fixing bugs and never oncej fell [video] Shakeel / @shakeelhashim : “You can keep the chat going indefinitely” is, I suspect, part of the reason why ChatGPT has seen so many more psychosis incidents than Claude. V curious to see how this plays out. Claude / @claudeai : We're also releasing new product features across our model line-up. For Claude app users, long conversations no longer hit a wall. Claude automatically summarizes earlier context as necessary, so you can keep the chat going indefinitely. Bluesky: @durumcrustulum.com : claude to be party to more ai psychosis cases [embedded post] Mike / @chicagomike : So if there's a hallucination or if the summary misses context are we just in uncontrolled flight at that point? Cross yo fingers. [embedded post]
Context & Ripple Effects
Claude has steadily expanded the amount of work a single session can hold, from Claude 2.1’s expanded context window to code-oriented tooling and thinking summaries alongside Claude Code. This update shifts the focus from fitting more text into a prompt to preserving continuity once a session outgrows its active window.
The change also extends Anthropic’s recent memory rollout beyond saved user information: it makes the conversation itself persist through automatic compression. That is important for users treating Claude as an ongoing work surface rather than a series of isolated prompts.
First-order effects
- Claude app users can continue long-running chats without manually starting a new thread when the context limit is reached.
- Anthropic makes automatically generated summaries part of the live conversation flow, reducing interruption for extended coding, drafting, and iterative problem-solving sessions.
Second-order effects
- Competing assistants face added pressure to make long-session continuity reliable, not merely advertise larger context windows.
- For users building work around a single assistant thread, summary quality becomes a practical differentiator: poor compression can undermine the continuity the feature is intended to provide.
Third-order effects
- The competitive unit for AI assistants may shift from the size of a model’s context window toward durable, managed conversation state across longer workflows.
- If these continuity features become standard, assistants will increasingly resemble persistent work environments, raising the importance of user control over what context is retained, summarized, and carried forward.
The trend: AI chatbots are evolving from turn-by-turn interfaces into persistent, workflow-native assistants that manage context over time.