In the latest development surrounding the Rabbit R1, CEO Jesse Lyu finds himself defending the AI device once again. This time, it's in response to claims that the Rabbit R1 is just a ChatGPT wrapper. Adding to the controversy, the Rabbit R1 experienced a temporary outage of its push-to-talk feature on Thursday night, coinciding with an outage on OpenAI's API.

The outage was first noticed by The Verge, who reported that Rabbit's push-to-talk functionality was affected. Rabbit swiftly responded, stating that the incident was related to OpenAI and impacted only a portion of the R1's functionality. Thankfully, the issue was promptly resolved, ensuring that the push-to-talk feature was restored.

Rabbit reiterated that it collaborates with various language model services, including OpenAI, to enhance its device's capabilities. Moving forward, the company plans to work closely with these vendors to minimize potential service disruptions.

A customer support account on Rabbit's Discord channel acknowledged the issue with the push-to-talk queries not responding and assured users that the team was urgently working on a solution. Meanwhile, the vision mode was stated to be fully operational.

As users flooded the "bug-report" Discord channel with error messages, OpenAI reported an API outage on its status page. The company noted elevated error rates on fine-tuning API models and reserved capacity. Approximately 45 minutes later, OpenAI resolved the issue, with Rabbit's customer support account confirming that push-to-talk should be functioning properly again around the same time.

Gizmodo reached out to OpenAI for a comment on the matter, but the company did not respond.

It is worth noting that one of the key factors contributing to the Rabbit R1's affordability, priced at $200, is the outsourcing of some AI processing. Rabbit uses OpenAI's ChatGPT to decipher user queries before its proprietary large action model (LAM) takes over to perform tasks like playing music or ordering food.

However, despite these efforts, concerns have been raised regarding the R1's functionality. Gizmodo's Dua Rashid pointed out the device's struggles, such as providing incorrect pickup and drop-off locations when using the Uber app, or acknowledging commands to play specific songs on Spotify but failing to follow through.

The R1's limitations were also highlighted during a demo event when CEO Jesse Lyu encountered difficulties while attempting to order from McDonald's. The device took an extended time to process the request and displayed an incorrect restaurant before eventually managing to place an order for a McNugget combo meal. Notably, Lyu attributed the mishap to the DoorDash interface, rather than the LAM.

These instances reveal that Rabbit's proprietary technology may pose challenges for users, primarily due to its heavy reliance on various language models. Moreover, the recent OpenAI API outage serves as a reminder of the potential hurdles that arise when depending on third-party AI models.

Looking ahead, Rabbit will need to address these concerns, enhancing its device's functionality and minimizing reliance on external AI platforms to ensure a seamless user experience.