Browser AI that runs on your own laptop is a good idea for privacy, and I want to see how it performs on a regular machine before I get excited. I have tested many local models using LM Studio on both my Mac and a variety of PCs (recently, I have been testing newly released hardware as part of a sponsored initiative with MSI). The bottom line from my ongoing local AI tests is that you REALLY need good hardware to be productive.
I applaud Opera’s initiative, but I honestly think running AI locally in a browser may be limited to very simple tasks. A chatbot may be okay for this, but for coding or deep research, I worry it could quickly become a resource hog (something browsers are already notoriously bad at).
That said, I like how companies are starting to think beyond the all-in-one AI app. This would be something compelling to test (with the right hardware, of course), and I know that there will be some creative individuals or companies that can build something compelling that integrates with Browser AI. Stay tuned!