The Rabbit R1 has been logging users' chats – with no way to wipe them

The linked story where the company claims an employee leaked keys to a “hacktivist” group and then something about sending defamatory emails to employees is very strange: https://www.rabbit.tech/security-investigation-062524

Nothing about this feels like a well-run engineering team. I understand it’s a startup, but all of this is just weird.

Feels like a case study in _why not_ to trust internet randos without hardware experience with a device that's meant to handle a bunch of private data.

The article says that an incoming update is addressing this by giving people the option to delete local data, so it looks like this soon won’t be an issue.

But stepping back, it is not at all surprising to hear about this type of flaw. MOST startups have these kinds of flaws. They spend as little as possible on things like security or privacy. Enough to meet some checklist minimally but not enough to actually respect customers and their data. In some ways it is understandable since they have to do whatever they can to survive, and the chances of that are low to begin with - so saving their energy and time for other things is what happens.

I bet most customers would be frightened if they knew exactly how cavalier most startups truly are.

And this is why user data should be stored encrypted. Not just at rest but zero access encryption or - where possible - client side, end-to-end encryption.

I'll make a short shameless plug: If you would like to use generative AI (OpenAI/Google/Anthropic/Open source) but don't want to run everything yourself, Cognos[0] stores your conversations encrypted so there are no risks of hacks, leaks or your data being used for training.

[0] https://app.cognos.io/

I remember watching Coffezilla's vid on Rabbit R1, and it looks both shady, and 10 years away from what it is promised to be.