From Using GPT to Building With GPT
A User’s Proposal for the Next Level of ChatGPT
My name is Maryam Akhbari, and I am a ChatGPT user who has gradually moved from simply using GPT to actually building with it.
I started like many users do.
I asked questions.
I had conversations.
I explored ideas.
I experimented with content creation and image generation.
But over time, something changed.
I started building projects with GPT.
Not just individual pieces of content, but larger ideas that evolve over time, require continuity, contain multiple layers, and develop across many conversations.
I am currently working on several projects, including a long-term creative project called Sora Universe.
And through this experience, I have realized something:
I don’t see a limitation in GPT. I see a ceiling — and I believe it’s time to raise it.
GPT has already become a global tool for thinking, creating, learning, imagining, and exploring.
A tool with this kind of reach and potential deserves the opportunity to evolve to the next level.
And as a user who has grown alongside it, I can genuinely see that potential.
The Question I Would Like to Ask OpenAI
Do you want GPT to remain primarily a tool where people come, ask questions, generate text or images, create something, and leave?
Or do you want GPT to evolve into something much bigger:
An environment where people can actually design, develop, manage, and build large, complex, long-term projects with AI?
I believe GPT is capable of becoming the second.
But if that is the direction, I believe there is a fundamental capability that needs to evolve with it.
Conversations need to become connected and reusable sources of context.
The Problem I Have Reached as a User
My own use of GPT has evolved through several stages:
Questions → Answers → Ideas → Content → Images
Then:
Concepts → Creative Development → Systems → Projects
And now:
Long-Term Projects → Multiple Conversations → Shared Knowledge → Continuous Development
This is a completely different level of usage.
At this stage, the conversations themselves become part of the project’s intellectual infrastructure.
One conversation may contain the original vision.
Another may contain research.
Another may contain strategy.
Another may contain characters, world-building, philosophy, storytelling, visual development, branding, or product thinking.
Together, these conversations become the project.
But today, these conversations are still largely treated as separate islands.
And that creates a serious problem when a project becomes large.
The Feature I Am Proposing
I would love to see ChatGPT allow users to make one conversation a reusable source of context for another conversation.
For example, I should be able to take a previous ChatGPT conversation and simply provide its Share link, then say:
“This conversation is part of my project. Please read it and use it as background context for what we are doing here.”
ChatGPT would then be able to access that conversation, understand its relevant content, and use it in the current conversation.
The conversation itself would become a referenceable knowledge source.
Instead of copying and pasting everything manually, the user could simply connect the previous conversation.
Why This Matters
When working on a serious project, users currently have to do things like:
- Copy and paste long sections from previous conversations.
- Re-upload the same files.
- Take screenshots.
- Re-send images.
- Repeat decisions that have already been made.
- Re-explain the history of a project.
- Manually summarize previous conversations before continuing.
This becomes increasingly inefficient as the project grows.
And the more valuable the project becomes, the more information it contains.
The irony is that the information already exists inside ChatGPT — but it can still be difficult to carry that context naturally from one conversation into another.
A Real Example: Sora Universe
One of the projects I am currently developing is Sora Universe.
It has evolved through multiple ChatGPT conversations.
Different conversations contain different layers of the project:
- Ideas and original concepts
- World-building
- Characters
- Philosophy and narrative
- Visual development
- Brand thinking
- Strategy
- Research
- Decisions that need to remain consistent as the project evolves
If I start a new conversation today, I may want GPT to understand two or three of those previous conversations.
Instead of manually transferring everything, I would like to simply say:
“These conversations are part of my Sora Universe project. Read them and consider them existing project context.”
Then I could continue building.
That seemingly simple capability would remove a huge amount of friction from the creative process.
From Shared Links to Project Intelligence
I don’t think this should stop at simply allowing ChatGPT to read a shared conversation link.
I believe it could become the foundation of something much bigger.
Imagine if ChatGPT could understand:
Projects
→ Conversations
→ Knowledge
→ Files
→ Images
→ Decisions
→ Versions
as connected elements of one evolving system.
For example:
Sora Universe
Conversation A — Original Vision
Conversation B — World Building
Conversation C — Characters
Conversation D — Philosophy & Narrative
Conversation E — Visual Identity
Conversation F — Strategy
Conversation G — Research
Then, in a new conversation, I could say:
“Use Conversations A, C, and E as context for this task.”
Or eventually:
“Use the Sora Universe project context for this task.”
That would transform ChatGPT from a collection of separate conversations into a project-aware environment.
This Is Bigger Than a Convenience Feature
I don’t see this simply as a request to make ChatGPT easier to use.
I see it as part of the evolution of what ChatGPT can become.
There is a fundamental difference between:
AI as a content generator
and
AI as a long-term creative and intellectual collaborator.
If people are going to use GPT to build companies, products, brands, research projects, creative universes, stories, systems, and ideas that take months or years to develop, then continuity becomes fundamental.
People should not have to reconstruct the history of their own project every time they open a new conversation.
Privacy and User Control
Of course, this capability should be built with strong privacy controls.
The user should explicitly decide which conversations can be referenced.
Users should be able to:
- Add a conversation as a project reference.
- Remove a conversation from a project.
- Select specific conversations as context.
- Control whether files and images inside those conversations can also be referenced.
- See which sources ChatGPT is using.
- Revoke access at any time.
The principle should be simple:
The user controls the connections between their own project contexts.
The Bigger Vision
My request is therefore not simply:
“Please make ChatGPT able to open a shared link.”
The deeper request is:
Please think about how ChatGPT can become genuinely project-aware.
If GPT is going to evolve from a tool that generates individual outputs into a platform where people can build large, evolving projects, then conversations cannot remain isolated islands.
They need to become connected pieces of a larger intellectual workspace.
A shared conversation should be more than something another person can view.
It could become a building block of a project.
Why I Am Sharing This
I am sharing this because I have personally reached the point where the current structure affects what I can realistically build with GPT.
I started by using GPT.
Then I started creating with GPT.
Now I am building with GPT.
And I want GPT to grow into that next level with us.
I don’t see this as a criticism of what GPT is today.
Quite the opposite.
I am making this suggestion precisely because GPT has already become powerful enough that users are beginning to ask more of it.
I don’t see a limitation in GPT.
I see a ceiling.
A ceiling that has become visible because GPT has already brought us this far.
I believe GPT is ready for its next level.
I believe it’s time to raise the ceiling.