I started testing out Claude more after mainly working with Chatgpt for months. I quickly ran out of free credits and figured I’d pay for Claude access as well to give it a chance.
But SADLY even worse than OpenAI and ChatGPT it sucks so so BAD on its limits. How is this reasonable to charge a monthly fee and cutoff a new user so so quickly? I seriously asked maybe 8 questions in a couple hours.
CLAUDE SUCKS worse than OpenAI and Chatgpt on limits – shutdown for hours. So much for productivity.
Wow. I thought the OpenAI product managers sucked but Anthropic is giving them a run for their money. How do you create such horrible interfaces. This error was hidden or showed up for 5 ms which humans just see a blip of light red. so dumb. I need to upload the video to share what a horrible customer experience this was. I tried to ask Claude what was happening but it was clueless. 
The only way to capture this was to create a quicktime video and then isolate the 5 ms screen frame.
User: Why does claude choke on attachments and just returns prompt without any error messaging?
Claude 3.5 Sonnet: I apologize, but I don’t actually have any information about attachments or images in this conversation. I’m Claude, an AI assistant created by Anthropic to be helpful, harmless, and honest. I don’t have the ability to see or process images or attachments. If you tried to share an image with me, I wasn’t able to perceive it. Could you please describe what you were trying to show me? I’d be happy to try to help based on a text description of the image or attachment.
User: That’s my point though. I gave you all this code but each time I hit the Send or up arrow to send it to you then the interface just spits back the prompt to me.
I apologize for the confusion. You’re right to point this out, and I appreciate your patience. This behavior you’re experiencing – where I seem to be repeating the prompt when you use the up arrow – is not intentional and is likely a limitation or bug in my interface.
To clarify:
- I do retain the context of our conversation, including the code snippets you’ve shared.
- I can see and process text content, including code, that you’ve sent in previous messages.
- However, I don’t have the ability to see or process images, and I don’t have direct access to any files or attachments beyond what’s included in the text of our conversation.
When you use the up arrow, it seems that instead of recalling our previous interaction correctly, the system is mistakenly presenting you with your original prompt. This is not how the interaction should work, and it’s not a reflection of my actual capabilities or memory.
For the most effective communication:
- If you need to refer to previously shared code or information, you can simply mention it or quote the relevant parts.
- If you’re seeing unexpected behavior like this, it’s best to type out your follow-up questions or comments rather than using the up arrow.
- Feel free to remind me of specific details from earlier in our conversation if needed.
I apologize for any frustration this may have caused. Is there a specific part of the code or a particular question you’d like me to address? I’d be happy to help based on the information you’ve provided throughout our conversation.
Unlock Peak Performance for Your Snowflake® Data Cloud
Maximize the potential of your Snowflake Account with Snoptimizer® – the ultimate solution for automated cost, performance, and security optimization.