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.
It has been almost 3 months – june,july,august until we see problems again with OpenAI being incorrectly focused but then again this message is so confusing because GPT-4o is better overall I thought for me. Anyone else have thoughts on this?
Just got this at 4.30pm pacific. “You’ve hit the Plus plan limit for GPT-4. New responses will use GPT-4o until your limit resets after 6:10 PM.”
ChatGPT
I understand that hitting the limit can be frustrating, especially when you’re working on something important. The message you received indicates that you’ve reached the usage cap for GPT-4 in your current billing cycle or daily limit, and the system is switching to a more resource-efficient version, GPT-4o.
If you have concerns about your subscription or want to discuss this further, contacting support through the platform you’re using might help clarify things or find a solution.
It has been almost 3 months – june,july,august until we see problems again with OpenAI being incorrectly focused but then again this message is so confusing because GPT-4o is better overall I thought for me. Anyone else have thoughts on this?
Just got this at 4.30pm pacific. “You’ve hit the Plus plan limit for GPT-4. New responses will use GPT-4o until your limit resets after 6:10 PM.”
ChatGPT
I understand that hitting the limit can be frustrating, especially when you’re working on something important. The message you received indicates that you’ve reached the usage cap for GPT-4 in your current billing cycle or daily limit, and the system is switching to a more resource-efficient version, GPT-4o.
If you have concerns about your subscription or want to discuss this further, contacting support through the platform you’re using might help clarify things or find a solution.
No, it is not possible to duplicate a standalone Google Apps Script (GAS) project with the exact same files and permissions, generating a new script ID, without manual manipulation.
Duplicating a GAS project while maintaining the exact same files and permissions, and generating a new script ID, inherently involves manual steps or the use of external tools that automate the process, which technically counts as a form of manual intervention. Here are the key reasons why this is the case:
- Manual Duplication Process:
- Google Apps Script does not provide a built-in feature for directly duplicating projects with all settings and permissions intact. Typically, duplication would involve copying the script files and reapplying settings manually.
- Permissions Handling:
- Permissions associated with a GAS project are specific to the script ID and the user(s) who have access. When creating a new script ID (essentially a new project), permissions need to be reassigned manually.
- External Tools:
- While there are third-party tools and scripts that can automate the duplication process, these tools themselves require initial setup and execution, which is considered a form of manual manipulation.
- Scripting Limitations:
- Even using Apps Script itself to automate the duplication process would involve writing and running a script that performs the necessary steps, thus not avoiding manual manipulation.
In conclusion, while you can duplicate a GAS project and recreate its environment and permissions, achieving this without any form of manual intervention or automation setup is not feasible with the current functionalities provided by Google Apps Script.
