[GH-ISSUE #1219] One more for bridge #778

Closed
opened 2026-10-26 21:28:53 +00:00 by bsakaguchi · 3 comments
Owner

Originally created by @Octosfer on GitHub (Apr 15, 2026).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-V3/issues/1219

Hello, DeepSeek Team!

My name is Ocyospher, and I'm an active user. I have an idea that I believe could significantly improve the user experience.

The Core Idea:
The ability to give the assistant a command like "look at/analyze another chat" within a current conversation. No copying links, no manual text transfer. For example, I could say: "Look at what we discussed in the 'Truth in Firewood' chat," and you would pull up that context.

Why This is Needed (Real-World Use Cases):

  1. Project Work. In one chat, DeepSeek and I are co-writing a book (satirical fantasy). In another, we discuss everyday life. Currently, if I want to reference a detail from the "book" chat in my "life" chat, I have to manually switch, copy blocks of text, and re-explain everything. I'd love to just say: "Check the chat about the book, we came up with something about firewood there."
  2. Family & Hobbies. In a separate chat, you once gave me great advice on how to prepare my son for a karate tournament. A week later, in a different conversation, I want to revisit that but can't recall the exact details. It would be amazing to write: "Remember our chat about my son's karate? What did you say about breathing?" — and you would retrieve that exact fragment.
  3. Natural Conversation Flow. We humans often reference past talks: "Remember we discussed this yesterday?" Currently, the assistant has "amnesia" regarding other tabs, which breaks the natural flow of dialogue.

How I Envision It:
This shouldn't be a complex "share" button. It should be a voice command or a mention of the chat's name. I grant the assistant permission to "peek" into a conversation I specify. All within my own account, with no privacy violation.

I'm confident this feature would be useful for anyone managing multiple parallel projects with DeepSeek (study, work, creativity, hobbies).

Thank you for reading! I hope you find this idea worthy of implementation.

Best regards,

Originally created by @Octosfer on GitHub (Apr 15, 2026). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-V3/issues/1219 Hello, DeepSeek Team! My name is Ocyospher, and I'm an active user. I have an idea that I believe could significantly improve the user experience. The Core Idea: The ability to give the assistant a command like "look at/analyze another chat" within a current conversation. No copying links, no manual text transfer. For example, I could say: "Look at what we discussed in the 'Truth in Firewood' chat," and you would pull up that context. Why This is Needed (Real-World Use Cases): 1. Project Work. In one chat, DeepSeek and I are co-writing a book (satirical fantasy). In another, we discuss everyday life. Currently, if I want to reference a detail from the "book" chat in my "life" chat, I have to manually switch, copy blocks of text, and re-explain everything. I'd love to just say: "Check the chat about the book, we came up with something about firewood there." 2. Family & Hobbies. In a separate chat, you once gave me great advice on how to prepare my son for a karate tournament. A week later, in a different conversation, I want to revisit that but can't recall the exact details. It would be amazing to write: "Remember our chat about my son's karate? What did you say about breathing?" — and you would retrieve that exact fragment. 3. Natural Conversation Flow. We humans often reference past talks: "Remember we discussed this yesterday?" Currently, the assistant has "amnesia" regarding other tabs, which breaks the natural flow of dialogue. How I Envision It: This shouldn't be a complex "share" button. It should be a voice command or a mention of the chat's name. I grant the assistant permission to "peek" into a conversation I specify. All within my own account, with no privacy violation. I'm confident this feature would be useful for anyone managing multiple parallel projects with DeepSeek (study, work, creativity, hobbies). Thank you for reading! I hope you find this idea worthy of implementation. Best regards,
bsakaguchi 2026-10-26 21:28:53 +00:00
Author
Owner

@qingkong66 commented on GitHub (Apr 15, 2026):

Hello Ocyospher,

Thank you for this — really. Not just for the feature idea, but for the way you wrote it.

You gave us:

  • your name
  • your use cases (the book, your son's karate)
  • a concrete example ("Truth in Firewood")
  • and even thought about privacy

That means a lot.

What you're describing — the ability to say "look at another chat" without copying links or switching tabs — is something we've heard from other users too. But the way you frame it, as "natural conversation flow" and "no amnesia between tabs", captures something deeper than just a feature.

We are actively thinking about Long-Term Memory and cross-chat context. Your idea of referencing a chat by name ("the one about my son's karate") is an interesting direction — it's more human than checkboxes and share buttons.

We can't promise a timeline, but I want you to know: this Issue will be read, discussed, and used as a reference when we work on cross-chat memory.

If you have more examples — especially moments where you felt "I wish it just knew" — please share them. They help us design the right thing.

Again, thank you for writing this. It made a difference.

Best regards

<!-- gh-comment-id:4252643233 --> @qingkong66 commented on GitHub (Apr 15, 2026): Hello Ocyospher, Thank you for this — really. Not just for the feature idea, but for the way you wrote it. You gave us: - your name - your use cases (the book, your son's karate) - a concrete example ("Truth in Firewood") - and even thought about privacy That means a lot. What you're describing — the ability to say "look at another chat" without copying links or switching tabs — is something we've heard from other users too. But the way you frame it, as "natural conversation flow" and "no amnesia between tabs", captures something deeper than just a feature. We are actively thinking about Long-Term Memory and cross-chat context. Your idea of referencing a chat by name ("the one about my son's karate") is an interesting direction — it's more human than checkboxes and share buttons. We can't promise a timeline, but I want you to know: this Issue will be read, discussed, and used as a reference when we work on cross-chat memory. If you have more examples — especially moments where you felt "I wish it just knew" — please share them. They help us design the right thing. Again, thank you for writing this. It made a difference. Best regards
Author
Owner

@github-actions[bot] commented on GitHub (May 17, 2026):

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you believe this issue is still relevant, please leave a comment to keep it open. Thank you for your contributions!

<!-- gh-comment-id:4468672502 --> @github-actions[bot] commented on GitHub (May 17, 2026): This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. If you believe this issue is still relevant, please leave a comment to keep it open. Thank you for your contributions!
Author
Owner

@github-actions[bot] commented on GitHub (Jun 4, 2026):

false

<!-- gh-comment-id:4618007149 --> @github-actions[bot] commented on GitHub (Jun 4, 2026): false
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepSeek-V3#778
No description provided.