OpenAI just published a guide to ChatGPT Projects—a feature that's been quietly available for months but criminally underused by most people who could benefit from it.
What Projects Actually Does
Think of Projects as folders for your AI work. Each project bundles related chats, uploaded files, and custom instructions into one workspace. You can have a "Marketing Q2" project with brand guidelines, a "Python Learning" project with code snippets, or a "Thesis Research" project with 20 PDFs.
The killer feature: ChatGPT remembers context across all chats within a project. Ask a follow-up question three days later, and it still knows what you're working on. No more "as I mentioned in our previous conversation" preambles that go nowhere.
Why This Matters Now
Most people use ChatGPT like a search engine—one-off questions with zero memory. Projects turn it into a persistent workspace. Upload a style guide once, set instructions like "always write in AP style," and every chat in that project follows the rules automatically.
OpenAI's timing is strategic. As ChatGPT moves from novelty to daily tool, organization becomes the bottleneck. Power users already have 200+ chats with names like "Untitled" and "New chat (47)." Projects fix that chaos.
What This Means for Learners
If you're building AI literacy, Projects teaches you a crucial skill: prompt architecture. Instead of cramming everything into one mega-prompt, you learn to structure context at the project level and keep individual prompts focused.
Try this: Create a "Learning AI" project. Upload articles you're reading, set instructions like "explain concepts using analogies from cooking," and use it as your personal AI tutor. Watch how much faster you retain information when the AI adapts to your learning style consistently.
The meta-lesson: AI tools get exponentially more useful when you invest 10 minutes in setup. Projects is that investment.