I would like to cancel my donation.

The goals "Building the world’s most trusted software company" and "It will evolve into a modern AI browser and support a portfolio of new and trusted software additions." are intrinsically contradictory.

LLMs cannot ever be trusted.

They are impossible to audit, and can be trained to have whatever bias the training organization would like to inject, and it is impossible for a third party to predict those biases.

Additionally adding a bunch of new code directly into the browser engine increases bugs that negatively impact the user experience and increase the risks for new security vulnerabilities.

Agentic LLMs are also intrinsically unsafe for security as there is currently no way to separate trusted commands from parsing application data. There have a continuous stream of ways for malicious actors to get their hostile commands to be interpreted by an LLM agent and executed.

Straight up we should not use agentic browsers.
https://www.theregister.com/2025/12/08/gartner_recommends_ai_browser_ban/

The only way to be protect against this attack is do not implement it.

If you ever come to your senses and remove the LLM integrations from Firefox I will consider renewing my donation.