Anthropic's Revamped Projects Feature Sparks User Backlash: Why Claude Power Users Fear Their Workflows Are at Risk

A User Revolt Over the Projects Redesign

Anthropic is expanding its revamped Projects feature for Claude into the Chat and Cowork experiences — but instead of applause, the change has triggered a fierce backlash from the community. A Max-tier subscriber took to Reddit with what amounted to a plea, stating bluntly that the update could completely destroy a workflow they rely on for several hours every day.

The controversy centers on an official statement from Anthropic: "We'll upgrade them as the rollout expands to Chat and Cowork." For heavy users who've built their routines around the existing Projects feature, that word — "upgrade" — sounds less like an improvement and more like a breaking change.

What Users Are Actually Protesting

Based on the user's post, the core tension is that the new Projects design appears to treat a project as a single, unified conversation — connecting multiple chats within a project so they share context and influence one another. The existing "project folders" model, by contrast, works precisely because conversations within a project remain isolated from each other.

The user put it plainly: "I don't want conversations within a project to talk to each other — that completely defeats the entire purpose of how I use projects." In other words, they treat Projects as a set of independent folders for organizing distinct tasks, with each conversation operating in its own silo. Once all chats are merged into a shared context, that clean task separation disappears.

For a Max-tier paying user who spends hours in Claude every day and considers project folders the cornerstone of their workflow, this isn't an optimization — it's a disruption of their entire working style. They went so far as to say, "I will genuinely cancel my subscription."

The "project folders" model referenced here is central to how Claude's Projects feature was designed when it launched in 2024. Users can create multiple independent conversations under a single Project, all sharing the same system prompt and uploaded files — but with each conversation's chat history kept fully separate. This structure lets users pursue multiple parallel threads within the same project topic. For example, under a single writing project, one conversation might be dedicated to brainstorming, another to editing, and a third to formatting — each maintaining its own clean context boundary. This "shared configuration, isolated memory" architecture is what sets Projects apart from ordinary conversation folders and is the foundation on which many power users build reusable workflows.

Isolation vs. Connectivity: Two Competing Product Philosophies

At its core, this debate is a collision between two distinct approaches to product design.

One school of thought holds that connecting conversations within a project enables stronger contextual continuity — Claude can remember prior discussions from the same project, reducing repetitive inputs and creating a more coherent collaborative experience. This aligns with the broader direction many AI companies are pursuing: longer memory and richer context.

The other school, represented by this user, emphasizes the value of context isolation. When you use Projects as a folder system to manage multiple parallel workstreams, keeping conversations from bleeding into each other is essential. Without that isolation, content from one topic "spills over" into another, creating noise and diluting the precision of each individual conversation.

Both needs are legitimate. The problem is whether Anthropic is forcing everyone onto one model with no alternative.

The tradeoff between "long-term cross-session memory" and "context isolation" is a recurring design challenge across the AI products space. When OpenAI introduced the Memory feature for ChatGPT, it sparked similar controversy — the system automatically retained information across sessions, prompting concerns from some users about privacy and unwanted information bleed between different tasks. Tools like Notion AI and Cursor have taken a more conservative approach, making cross-file or cross-session context sharing an explicit user action rather than a default behavior. This broader context shows that Anthropic's change is not an isolated incident — it's a microcosm of the ongoing industry tug-of-war between "smarter and more connected" and "more controllable and precise."

The Core Ask: Make It Optional

The solution this user repeatedly advocated is actually quite pragmatic, and it boils down to two points:

• Make it opt-in (off by default, activated by user choice), rather than force-overwriting everyone's existing experience.
• Or offer it as a standalone separate feature, letting users decide whether to adopt the new model.

This is a textbook example of a demand that comes up again and again in product development. When a tool has become deeply embedded in a user's daily workflow, any "upgrade" that's on by default can easily become a "breaking change." Respecting established user habits, providing a migration buffer, and offering genuine choice often matter more than how advanced the new feature itself is.

Opt-in and opt-out represent two fundamental default-state strategies in product design. Opt-in means a feature is off by default and must be actively enabled by the user; opt-out means it's on by default and users can choose to disable it. Research consistently shows that most users don't change default settings, meaning the choice of default effectively determines the experience for the vast majority of people. For existing users who've accumulated deep usage habits, a forced opt-out-style upgrade frequently causes experience disruption, while an opt-in strategy allows new and existing users to coexist with a smooth transition. Collaboration tools like Slack and Notion, when rolling out major interface or logic overhauls, typically give paid users weeks or even months to retain a "legacy mode" — a decision rooted in exactly this consideration.

What This Means for AI Product Development

This incident originates from a single user's post, but it reflects a tension that AI tool companies broadly face as they iterate rapidly. Paying users — especially high-tier subscribers like those on Max plans — are often a product's most intensive users. They've built elaborate workflows around existing features, and for them, stability and predictability can matter more than new capabilities.

It's worth noting that everything discussed above comes from a single Reddit user's description and emotional reaction. Anthropic has not yet provided a fuller explanation of the new Projects design — including whether a toggle or opt-in option will be available. The actual scope and impact of the changes should be assessed based on official announcements.

Regardless, the signal from this post is clear: for productivity tools where users have developed stable, ingrained habits, giving users the choice is always the lowest-risk path through a redesign. Forced product changes — even well-intentioned ones — can push loyal users to the edge of churning.