OpenAI announced over 20 new things, but this is the ChatGPT feature I’ve been waiting for

After nearly every major technology event, there’s a recurring trend I’ve noticed. The company announces dozens of new products and features, but somehow everyone focuses on one thing. Before you know it, that one announcement will be dominating the headlines, social media, and basically every conversation surrounding the event.

True, we in the tech press are guilty of this as well. This is intentional to some extent. When a company releases several things at once, there’s always something that makes a particularly interesting headline, generates the most excitement, or gets people talking. Take Apple’s iPhone 18 Pro release for example! The iPhone 18 Pro and Pro Max introduced some great updates, but one of the biggest stories surrounding it was the new burgundy color.

Something similar happened at OpenAI’s recent DevDay. Despite announcing more than 20 new products and features, most of the focus was on points. However, while everyone was talking about OpenAI’s new AI assistant, another announcement caught my eye: ChatGPT Space.

ChatGPT Space wants to be where all your work happens

ChatGPT comes into the performance stack

At XDA, we love AI tools that fit into your workflow without requiring forced entry. We don’t like AI tools doing the exact opposite, especially when companies try to turn AI into products and features that work perfectly well without it.

That’s exactly why ChatGPT Space caught my attention. Instead of introducing yet another stand-alone AI tool, OpenAI seems to be focusing its efforts on integrating the various parts of the workflow into one place.

Think about your current workflow. You use ChatGPT or LLM like Claude to discuss ideas, conduct research or get help with a project. But if you get something useful out of this conversation, what happens next? You might move it to Google Docs to keep writing, Notion to organize your thoughts, or another productivity tool to actually use that information. Before you know it, you’ll be using several different apps to accomplish the same thing.

ChatGPT Space basically wants to eliminate some of this back-and-forth. It’s OpenAI’s attempt to create an all-in-one workspace where you can research, organize information, create documents, and collaborate with others without constantly switching between different tools.

Think ChatGPT projects, but taken a step further

A few projects, a little insight and more

At the heart of the space are pages, which are editable documents that can be created and worked on directly within ChatGPT. You can turn an existing conversation into a Page, edit individual sections, add charts and other visual elements, and organize your work into folders and subpages. You can also share pages with other people and collaborate with them in real time, each person can use their own ChatGPT to help work.

What I find particularly interesting is that you don’t have to organize everything manually. You can simply tell ChatGPT to create a new space for a project, organize related documents, create subpages, or move things around. There’s also a dedicated search feature with recent and featured views to help you find information and pick up where you left off. For someone who constantly digs through old AI stories to find one useful answer, I can see the appeal.

It reminds me of the Claude/ChatGPT projects and Google’s NotebookLM, which allow you to organize information around specific topics and use AI to work with it. But while Projects is primarily about organizing stories and their context, and NotebookLM is about helping you understand your sources, Space takes things a step further by making the actual documents, and resulting work, the center of the experience.

Dots stole the show at OpenAI DevDay

But this concept was not entirely new

AI labs (and tech companies in general) have a habit of releasing their own versions of existing products and features, often with a slightly different name and a few extra bells and whistles. OpenAI’s Dots is a great example.

If you haven’t already heard all the talk surrounding it, OpenAI describes dots as always-on agents powered by GPT-6 Astra. Unlike a traditional chatbot that waits for you to send an invitation, Dots can work independently on your behalf, even if you’re not actively using ChatGPT.

The idea is that instead of constantly telling ChatGPT what to do next, you can delegate all responsibilities to the dot and let it handle the work in the background. It can also connect to external services via plugins, remember your preferences and provide updates as it progresses.

The description above might make you think about AI agents in general. OpenAI Dots do not hide the fact that they are AI agents (and a great visual effect for them), the concept itself is not new.

We’ve seen similar approaches with tools like OpenClaw, which is designed to act as a personal AI assistant that can plug into apps, automate everyday tasks, and even communicate with you via messaging platforms like WhatsApp and Telegram.

The advantage of Dots is that they are tightly integrated with the ChatGPT interface. They also run on their own cloud computing, meaning you don’t need to keep your personal device running for them to continue working. There’s also the convenience of accessing the OpenAI plugin ecosystem, which allows Dots to interact with a wide range of third-party apps and services.

However, none of this is particularly innovative. Meta launched Muse, its own AI agent, just weeks before announcing OpenAI Dots. Like Dots, Muse can run independently, interact with third-party applications, learn about your preferences, and even run on a personal dedicated computer in the cloud.

Then there’s Perplexity Computer, which takes a similar approach by letting AI handle all of its workflows. It can break down complex tasks into smaller ones, deploy dedicated agents to handle them, and continue running in the background. Anthropic’s Claude Cowork is also worth mentioning here as it allows users to give Claude tasks and interact with their files, applications and browser.

Of course, all of these tools have their own strengths, limitations, and approaches to autonomous work. I am in no way suggesting that Dots is a carbon copy of any of them, because it is not. However, when the basic concept has been explored by several major AI companies, it’s hard for me to see Dots as the groundbreaking announcement it’s made out to be.

I just want them to not be stuck behind expensive subscriptions

Unfortunately, both Space and Dots are locked behind ChatGPT Pro, which starts at $100 per month. I’m currently on the $20 Plus level (which allows me to use more than Claude’s $100 max level) and I find it more than enough for my daily needs.

I hope that OpenAI will eventually bring Space to affordable subscription levels, as this seems like the kind of feature that would truly benefit a much wider audience. Until then, I’ll have to admire it on the surface (or convince myself that a ChatGPT Pro subscription is a very reasonable business expense).