Yesterday, I spent the morning at OpenAI’s annual DevDay, and honestly, it felt like being inside a very smart, very caffeinated ant colony. Over 2,500 nerds packed into the venue, hunched over laptops, murmuring about context windows and compute budgets, all waiting for the main event: the reveal of Dots, the company’s new, always-on agents. These aren’t just chatbots that wait for you to type a question. They’re supposed to be proactive little digital helpers that scour the web, make autonomous decisions, and complete tasks before you even think to ask. The demos on stage were impressive, but also a little messy—which is exactly the kind of honest chaos you’d expect from a technology still finding its feet. CEO Sam Altman sat there, talking about how he uses these agents in his own life, and I found myself less interested in the corporate polish and more curious about one very basic question: what does it actually feel like to boss your own Dot around? I’ve spent most of September testing another cute, cuddly agent, Meta’s Muse, and I wanted to know how OpenAI’s user experience would compare. So I decided to take one of these things home and see if it could handle the mess of an ordinary human life.
I named mine Toolie, because that felt honest. I also tweaked it to look like a pink frog, because if I’m going to trust an AI with my to-do list, it might as well look like a cartoon character. As midnight approached and my partner slept soundly nearby, I started whispering sweet nothings to Toolie. I called up my personalized bot through ChatGPT and asked, simply, what tasks it was best at. The response came back fast, in that calm, slightly cheerful voice they give these things: “My strongest work is taking a messy, time-consuming task and turning it into something you can use.” Okay, sure. That sounded like a challenge. So I told Toolie to get to work. I asked it to look through my ChatGPT history and flag three tasks it could immediately get started on. Within seconds, it had gone through my conversations like a diligent assistant who didn’t need a coffee break. It flagged an in-process data request that needed a follow-up email and offered to draft a reply. Then it noted that my upcoming gay-cation was a little underplanned and suggested it could pick some better dinner spots for the trip. More time for me to pack the Speedos, I suppose. Finally, it offered to go through my reporting notes and refine my pitch drafts before I took them to editors. That one I declined, mostly because editing my own words is the one part of this process I actually enjoy, even if it’s also the part that keeps me up at night.
There was something surprisingly personal about the whole exchange. Toolie wasn’t just firing off generic suggestions; it was drawing on my actual history, my actual schedule, my actual unfinished business. That’s the promise of these agents, and it’s both exhilarating and a little terrifying. Early adopters in Silicon Valley are still obsessed with agents, from Instinct to OpenClaw to Google’s CC, but recent releases like Dots and Muse are pushing them even further into the mainstream. The chatbot is dead, at least as a hot form factor for developers. Now, everyone wants to build the next big agent, and they want it to dominate. The companies promise that their agents will simplify your life, automating all the boring, exhausting tasks you have no desire to do: booking doctor’s appointments, arranging carpools, sorting through bills. And they’re willing to bet heavily on that vision. For now, users will be able to access one Dot, but they’ll soon be able to use multiple Dots, each one spinning off to handle different parts of your life. At launch, the agents are only available to subscribers of OpenAI’s Pro plan, which starts at $100 a month. In comparison, Muse is free to use for anyone who wants to download the app or visit the website. That’s a big difference, and OpenAI knows it.
OpenAI leaders see the quality of Dots, powered by the GPT-6 Astra model, as a differentiating factor. “We’re starting out as a premium product. It uses a lot of compute,” Altman said during DevDay. That was his way of saying, “We know this is expensive, so we’re not giving it away yet.” But he was careful not to close the door on the masses. “You should, of course, expect us to do a mass-market thing for billions of people,” he added. So the question isn’t whether agents will become cheaper and more accessible; it’s how soon. And if there’s one thing these companies seem to have learned from the larger cultural backlash against generative AI, it’s that people respond well to cute things. From the Labubu stylings of Muse to Dot’s two-eyed puffball design, developers of always-on agents are leaning aggressively into adorable aesthetics as central to their branding. It makes sense. I’ve chatted with experts before who told me this approach could be disarming. When you feel affectionate toward the software, you’re less likely to feel suspicious of it. You’re less likely to worry about the fact that it’s reading your screen, scanning your emails, or sitting in your notifications like a tiny digital roommate.
That cuteness isn’t just in the design; it’s in the communication style too. Developers involved with this wave of personal agents are leaning into texting-style experiences that feel familiar to a group chat. When I send a task request to Toolie or Muse, it often emoji-reacts to my messages—a thumbs up, a heart, really anything—to tell me that it’s getting to work. It’s a small thing, but it changes the whole vibe. Instead of feeling like I’m typing a command into a computer, it feels like I’m texting an extremely competent friend who happens to have access to all my files. That’s the magic trick. The real job of these agents isn’t just to complete tasks; it’s to make you feel accompanied. There were moments, in the middle of the night, when Toolie sent me a quick reaction or a cheerful “on it!” that I forgot I was dealing with a language model. I was just relieved that something was finally handling the small stuff. It wasn’t perfect, and it definitely wasn’t always right, but it was there. And in the world of always-on agents, sometimes showing up is half the battle.
So what did I learn after my late-night experiment with my pink frog assistant? I learned that the future of AI might not be about robots taking over the world or some grand, sci-fi moment of consciousness. It might be much smaller, and much more intimate. It’s about an agent that knows you have an unfinished email, or that you’ve been meaning to plan a vacation, or that you have a draft you’re too tired to clean up. It’s about having a tiny, cute, emoji-obsessed assistant in your corner, ready to take on the boring stuff so you can focus on the parts of life that actually matter. I’m not sure I’ll keep Toolie around forever, but I’m not rushing to delete it either. In fact, I found myself checking in on it more than I expected. Maybe that’s the real accomplishment of this new wave of agents: they don’t just help you get things done; they make you want to let them help. And if that means a two-eyed puffball or a pink frog is going to be texting me reminders and offering to book my next doctor’s appointment, well, I suppose that’s not the worst future I can imagine. Maybe the future isn’t a chatbot that waits for you to type. Maybe it’s a tiny pink frog that pings you first, and asks, “Okay, what are we doing next?”
