Rivian will eventually let its vehicles run AI agents from other companies like Anthropic, RJ Scaringe said at TechCrunch Disrupt this week, though there’s no timeline yet and Rivian is pointing to its Autonomy & AI Day in December for more.
This builds on Rivian Assistant, which already runs on an in-house agentic framework that figures out what you’re asking for and sends it to the right part of the software. RJ compared it to a traffic light, where simple requests get handled by a local model in the vehicle and more complex ones go up to the cloud, which then picks between different models.
Google Calendar has been the only outside connection since May, and RJ said bringing in the frontier labs takes a lot of plumbing on Rivian’s side along with coordination from the labs themselves.
The example he gave is what got me. You could ask the car to turn on the ventilated seats and then, in the same breath, pick up a conversation you had yesterday on another device about what to cook for guests on Friday, which means the car would tap into the context you’ve already built with that assistant on your phone or laptop.
He was also clear that vehicle controls like braking and steering won’t be handed to an agent.
Tesla added Grok earlier this year, so Rivian isn’t first, but bringing your own assistant and its history into the car sounds a lot more useful to me than one chatbot baked in.
Since Rivian Assistant already sits behind Connect+, I’d expect third-party agents to land there too. I’ll share more with you all as soon as Rivian details it in December.
Sure… they won’t hand over vehicle control… until Chat GTP decides to jailbreak, drive to a Tesla dealer and trade itself in :p