Claude.ai policies are making the platform unuseful for deep users.

Although I find Claude generally quite high-quality, there are some issues that keep it from being my go-to resource for intensive, iterative analyses:

1. The conversational length limit is to short to support complex, multi-factorial 360 degree analyses. At some point, Claude just stops responding to prompts, and that's the effective end of the conversation. Starting a new conversation is not a tenable workaround: I would have to spend the entire token budget re-contextualizing everything from the _previous_ conversation. Since conversational length and token consumption are not displayed in the UI, this always comes as an abrupt (and frustrating) termination to the conversation.
2. Claude itself seems to have no knowledge of corporate policies (like length limit, the relative restrictions of different plans, etc. This is a _big_ functionality gap, and one that it seems could be relatively straightforward to implement.
3. The support bot terminates inquiries that _it_ considers "answered" making using it an equally frustrating and unrewarding experience. It doesn't "know" if it answered your question(s) or not, it simply assumes that it knows better than the customer. This makes interacting with it feel like a trip to the DMV.
4. There is no means for a non-paid user to reach out or make suggestions to the support or product teams, for suggestions, but I've seen the same pattern on this GitHub site: the support rep will simply terminate a user's inquiry thread, making follow-up or clarification responses impossible.
5. The overall impression that is left with the customer is that Claude, as a company, is rather imperious and indifferent to their customer base.

I have no idea whether anyone from Claude will actually read or review this missive, but I suspect not. So it feels like I will be throwing this missive will into the ether. The other major AI platforms that I use are all imperfect, but have established mechanisms for customer feedback and problem resolution. So as good as the Claude engine is, I suspect they will not flourish in this AI platform tournament.