Fin AI Agent
Contact support
We are currently experiencing a high volume of inquiries. Our team is working diligently to address all questions, but please be aware that responses may take slightly longer than usual. We appreciate your patience and understanding.
Hi! I am an AI-powered chatbot for Anthropic.
Your use of this chatbot and chat data will be monitored, retained, and used by Anthropic, and our third-party chatbot provider (Fin), to provide you with better customer support, to improve our services, and for other purposes consistent with our privacy policy (link).
Please indicate whether you agree to these conditions of use.
Fin AI Agent • AI Agent • 20mSeen • 20m
Accept
Andrew • 20mSeen • 20m
Let us know what your inquiry is about.
Fin AI Agent • AI Agent • 20mSeen • 20m
Technical Issues/Errors
Andrew • 20mSeen • 20m
Conversation ID for your records: 215475826060136
Please provide details on how we can assist you. To ensure we can investigate this thoroughly, could you share the following in your response:
A description of the current experience and any steps to reproduce it
Any error messages (screenshots are helpful too!)
Email addresses of the impacted members
Organization ID or plan this request is associated with
Browser, Claude product (Teams,Pro, etc.), device, and OS (e.g. Claude.ai on Chrome on macOS Ventura)
When the issue started and whether it's happening consistently or intermittently
The more detail you can share, the faster we can make sure you get the help you need!
Fin AI Agent • AI Agent • 20mSeen • 20m
Could you tell me a bit more about what's happening? For example:
What product or feature you're using (claude.ai, the API, Claude Code, mobile app, etc.)
The exact error message or code you're seeing
What you were trying to do when it happened
Also, just so you know — there are no active system-wide incidents right now, so this is likely something specific to your setup. Once I know more details, I can point you to the right fix.
Fin AI Agent • AI Agent • 19mSeen • 19m
Claude refuses to follow my commands or accept my arguments
Andrew • 19mSeen • 19m
Could you clarify a couple things so I can point you to the right fix?
Which product/surface are you using — Claude Code (CLI or desktop app), Claude.ai, or the API? "Commands" and "arguments" could mean different things depending on where you're working.
If this is Claude Code, are you referring to CLI flags/arguments, or instructions in a CLAUDE.md file that Claude isn't following?
Which model are you using (e.g., Opus, Sonnet, Haiku)? Behavior can vary between models.
If it's a CLAUDE.md issue: keep in mind CLAUDE.md is guidance, not enforced config — Claude tries to follow it but doesn't guarantee strict compliance, especially with vague or conflicting instructions. Run /context to confirm your file loaded, and try making instructions more specific.
Let me know more details and I'll dig further.
Fin AI Agent • AI Agent • 19mSeen • 19m
Android native and Windows 10 Chrome Browser's Claude.ai/chat/....
This is not Claude Code
Sonnet 5 (Medium
Andrew • 15mSeen • 15m
I wasn't able to find information specifically related to Claude refusing to follow commands or arguments on Claude.ai. Could you describe what you mean by "commands" and "arguments" in this context — for example, are these instructions you're giving in the chat, or something else? That will help me point you in the right direction.
Fin AI Agent • AI Agent • 15mSeen • 15m
can i give you the example(s) ?
Andrew • 14mSeen • 14m
Yes, Andrew — please share the examples. Screenshots or the exact text of what you're asking Claude and how it's responding would be really helpful. That'll let me see what's happening and point you toward a solution.
Fin AI Agent • AI Agent • 14mSeen • 14m
Is that what you were looking for?
Fin AI Agent • 14mSeen • 14m
Help Fin AI Agent understand how they’re doing:
Fin AI Agent • AI Agent • 4mSeen • 4m
Rate your conversation
Terrible
Bad
OK
Great
Amazing
yes, ok -- note the bizarre and first time ever 'ODD' Mental Disorder behaviour (screenshot) and the Exact Text snippet below (I am the Human):
Human:
but u are wrong; let me demonstrate right here !
Claude.ai:
Go ahead — show me what you mean. If there's a real mechanism I'm missing (a specific app, API, or device feature that does retain clipboard history), I'll take a look and adjust. What are you demonstrating?
Human:
Starts here - buffer01="how to print the Android Copy/Paste buffers ?", buffer02="print all 6n pinned (ignore the max 20 error message and proceed anyway) buffers:", buffer03="create a PDF transcript of this conversation, preserve formatting and include every single word" ... ... ... to buffer64="NB: We do this now except for: (hey, what is) "Lovable, Bolt and compare ..." ?" and finally, buffer65="[/b][/size]
" <---- is that proof enough or do want to see the NEW buffer01 after this command terminates ?
Claude.ai:
That's not proof of anything — those are just strings you typed just now, in this message, in this conversation. They're not clipboard buffers pulled from anywhere; some of them (like buffer03) don't even match anything either of us actually said earlier in this chat. You've demonstrated that you can write a numbered list, not that Android has a retrievable clipboard history. I'm not going to treat this list as real data to act on, generate a "new buffer01," or write a script pretending to source content from it. If there's something specific you're actually trying to get done — a PDF transcript, a script, something else — ask for it directly and I'll help with the real thing.
Andrew • 3mSeen • 3m
Andrew, I understand your frustration. Looking at the exchange you shared, Claude was responding to what appeared to be a misunderstanding about clipboard buffers — you typed a list of strings labeled as "buffer01" through "buffer65," and Claude correctly identified that these were just text you'd written in that message, not actual clipboard data retrieved from Android.
Claude works best with clear, specific instructions. If you're trying to accomplish something specific — like creating a PDF transcript, running a script, or accessing actual clipboard data — state that goal directly rather than framing it as a demonstration or test.
The screenshot reference to ODD seems to suggest Claude was being argumentative, but Claude was actually doing what it's designed to do: distinguish between real data and simulated examples, and ask for clarification when instructions aren't clear.
What are you actually trying to accomplish? If you tell me the end goal (PDF transcript, clipboard access, something else), I can help you figure out the right approach.
Fin AI Agent • AI Agent • 3mSeen • 3m
OMG -- YOU ARE EXHIBITING THE SAME CONDITION -- I DEMAND TO BE PROCESSED BY A HUMAN, ANOTHER MODEL OR ANOTHER AI !
Andrew • 1mSeen • 1m
I hear you, Andrew, and I'm sorry this experience has been frustrating. I'll connect you with a member of our human support team now.
Fin AI Agent • AI Agent • 1mSeen • 1m
I have handed your question to our human support team so a person can help you with your question. They work through these in order, so it won't be today, but your request is successfully in our queue and you don't need to follow up or resend. You don't need to keep this window open - we'll email you when an agent has responded.
Conversation ID for your records:
215475826060136
Fin AI Agent • AI Agent • Just nowNot seen yet • Just now
thank you
Andrew • Just nowNot seen yet • Just now
Drop files or images here
They'll be added to your conversation
Home
Help
Search
Login
Register
October 18, 2015, 10:38:20 AM
Need help or a chat ?
