Autonomous has created a new version of its AI Intern, a plug-and-play AI super-assistant that physically sits on your desk. It uses multiple AI models to work on every task you can imagine, using OpenClaw and Hermes, and selecting from models like Opus, Sonnet, Deep Seek and Qwen.
It promises to handle jobs like social media posts and everyday scheduling and admin – but will even work on more involved tasks like coding. You can leave it to run on its own schedule – handy when you're away or working to another time zone overnight. And it works locally, rather than in the cloud. To communicate with it, you can text or talk.
I should state from the off that this will be problematic to creatives worrying about jobs in the AI age (especially at its marketing of the difference between a co-worker and the device being it "works unlimited hours" – I think we can all agree an actual human would bring more to any role than a computer). But its description will likely entice anyone feeling they're at capacity without the means to hire an actual intern or small team.
I'm going to press on describing the Intern 2 with a pin in the ethical concerns, because I'm actually impressed by the sound of the device's capability from a technical perspective. Of course, if you are able to hire a human, they would add more – but if you can't, this seems like a very close second.
The latest upgrade gives the Intern 2 "a mind and a voice" and takes on the admin load of your workflow, which would free up your time for creativity. The key features are below:
- Multi-model intelligence: Rather than locking users into a single AI ecosystem, the Intern 2 dynamically swaps between top-tier large language models depending on the task. It routes reasoning to GPT-4/Opus, everyday tasks to Claude Sonnet, deep knowledge requests to DeepSeek, and multilingual work to Qwen.
- Always-on autonomous workflows: Unlike browser-based AI companions, this hardware runs 24/7. It is designed to track a schedule and triage emails, monitor GitHub issues, handle DMs, and run morning digests autonomously while the main computer is completely asleep.
- Plug-and-play chat integration: It integrates with communication apps right out of the box. Users can text or talk to the hardware directly, or assign tasks via team chat platforms like Slack, Discord, and Telegram.
- Strict local data privacy: For creators and developers worried about intellectual property leaks, the Intern 2 keeps API keys, project context, team vocabulary, and style guidelines stored entirely as local files. No third-party cloud servers are used for storage.
As with the best applications of AI, Autonomous has created a device that promises to handle the busy work of your day. This appears to be where most genuinely useful AI is heading, as our AI in Focus series showed. We also found out during that series that lots of creatives and developers prefer the idea of a local AI rather than using the cloud (find out here what that can do for you).
Autonomous promises each device has been tested end-to-end before it gets to you, and the design itself is sleek – with a futuristic light on the top. I'm hopefully getting one in to test so I'll report back with my findings – and decide how I feel about the ethics of it all after actually having used one.
Would you try out a device like this? Let us know in the comments.
