🚨 OpenAI just unveiled Dots at DevDay, and I’ve been using them since Friday: I was part of the Dev Early Access Community testing.
Here’s what it’s been like.
A Dot is an agent that’s always on, with its own computer in the cloud. You connect it to your email, calendar, and other tools, and it doesn’t wait for you to message it: it checks on its own what you might need and suggests solutions that you often just have to approve.
The conversation feels more like WhatsApp than a ChatGPT chat.
You give it a name and avatar (mine is called Keet), and if you reply with a 👍, it understands that you approve and keeps going.
Here’s what has surprised me the most these past few days:
1️⃣ This Friday I have a training session followed by an event, with only half an hour in between. It spotted this from an event email, warned me about the timing, and asked how I planned to get there. I told it I was driving, so it found the cheapest nearby parking, opened the reservation, and filled in the details for me. I made the payment myself.
2️⃣ It started reviewing an assignment submitted by an AI Expert student. To watch the video, I asked it to install Whisper on its machine. It transcribed the video, reviewed the repo through the GitHub connector, and gave me an initial review. I’m no longer starting from zero.
3️⃣ Another student had asked a question, and Nino had already answered it. It noticed this and suggested adjusting my draft to add only what was missing.
It also joins Slack like another member of the team, and you can call it by voice just like a regular phone call.
For those small tasks that keep getting pushed back because you never find the time, I’m finding it really useful.
Are you going to try it?
Your dot is ready to meet you.