This new ChatGPT feature will change your work forever
Your ChatGPT projects just became 10x more powerful
I've been using Claude (since Sonnet/Opus 4) to write and plan most of my content.
But a couple days ago, OpenAI released an update that made me immediately switch back to ChatGPT.
Not because of some flashy new model or benchmark scores. But because they finally fixed the one thing that was making me rage-quit my writing sessions.
If you've ever lost your train of thought jumping between chats, or had to copy-paste context from one conversation to another, you know exactly what I'm talking about.
The feature I'm most excited about: memory across project chats
Here's the problem that's been driving me crazy:
You're working in a ChatGPT or Claude project. You write/plan a bunch of content in one chat. Then you start a new chat that has no idea what you just worked on.
It can't reference your draft. Can't build on your ideas. Can't even remember the tone you established.
So you end up copy-pasting. Re-explaining. Trying to attach it to the project knowledge.
It's like having a writing assistant with amnesia.
Well, that just changed.
ChatGPT projects now have memory across all chats within a project. Every conversation can reference every other conversation. Automatically.
This is way more than just convenient. It fundamentally changes how you can build an AI writing system.
4 other updates worth your time
1. Canvas downloads
Now you can download your canvas directly as:
PDF
Microsoft Word
Markdown
Small feature. Big time-saver.
2. Voice mode in projects
You can now use voice mode inside projects, not just in regular chats.
Why this matters: Sometimes the best ideas come when you're talking through them, not typing. Now you can think out loud and have it captured directly in your project context.
3. Deep research inside projects
Before, you had to run deep research in regular chats, then manually move everything into your project.
Now? Run it directly inside your project. All that research automatically becomes part of your project's knowledge base.
4. Drag-and-drop organization
Got a brilliant chat sitting in your regular conversations? Just drag it into any project.
No more recreating work. No more lost context.
What this means for your AI writing system
Most people use AI tools like they're starting fresh every time. New chat, new context, new problems.
But with proper memory across chats, you're building something different: a compound knowledge system that gets smarter with every conversation.
You're not just using a tool anymore. You're building a second brain that actually remembers what you taught it.
The tool-agnostic truth
Look, I jump between Claude and ChatGPT like a hyperactive squirrel. When Claude releases something amazing, I'm there. When ChatGPT drops an update like this, I switch back.
That's why I created that mini-course last week on building your tone of voice across different tools. Because the tool doesn't matter as much as the system you build with it.
→ Click here to watch the full walkthrough on building your tone across LLMs.
But right now? ChatGPT just made building that system significantly easier.
Will Claude respond with something even better next week? Maybe. That's how this game works.
But today, ChatGPT projects went from "nice to have" to "essential for serious AI writing."
If you're still treating each chat like an isolated conversation, you're leaving massive productivity gains on the table.
Build the system. Use the memory. Transform how you create.
And let me know what you build with it.
Alex
Founder of AI Disruptor
PS: The mini course on building your AI voice DNA gets locked behind the founding tier tonight. Paid subscribers—make sure to grab it from the Notion page before then. Not a paid subscriber yet? Upgrade now to get instant access before it moves to founding tier only.
Imagine being so bothered by a headline that you leave comment about it and think you’re “helping.”
Can we stop with the hyperbolic headlines? They added a few useful features to projects. Ones we all knew would be added at some point. I mean, I guess technically they will change my work forever, since these features allow me to work in different ways as before, but that's true of any feature for any app. I guess I'm just asking for a little more respect for your readers. And frankly, it devalues the content that is behind the headline. You have genuinely valuable content most of the time, but headlines like this put you in the category of hucksters and low-value AI slop-slingers. You can do better, and we - your readers - deserve better. Please and thank you.