People are feeding AI more of themselves every day.
One palm-reading post pulled millions of views. That alone says enough.
We are treating deeply personal data as something casual to share, before really thinking through the consequences.
Biometric data is tied to identity. Once it is out, it can be stored and used in ways the person who uploaded it never intended.
Of course, the solution is not asking people to share less. It is educating them on the importance of privacy, and building systems where sensitive data does not have to be exposed just to be useful.
When you talk to your lawyer, you expect confidentiality.
When you talk to your counselor, you expect confidentiality.
When you talk to AI, you should also expect confidentiality.
nilGPT is your confidential support. It's a conversation between you and AI, and nobody else.
This is why we created nilGPT. We saw this coming and wanted to create a safe space for people to share anything about their life without having to be worried about a surveillance state.
We can't see your chats. We couldn't even if we wanted to.
DANGEROUS STUFF
> over 20 million ChatGPT conversations
> are now evidence... a judge just ordered
> OpenAI to hand them over in a lawsuit
if anything ever screamed
> Buy a GPU
> run your AI locally
> protect your damn data
it’s this right here
1) @nilgpt_
nilGPT is a privacy‑first AI assistant built by Nillion. It runs on two core blind modules nilAI for inference inside a trusted execution environment, and nilDB for encrypted, distributed storage. All user queries are processed locally within the TEE, never exposed to the host, and conversation history is split into secret shares and stored only in nilDB.
I asked @nilgpt_ and @Copilot how to invoke a clipboard popup on XFCE desktop with CTRL+V, pretty much like you do on Windows.
Well, this shows not always the most advanced model provides the best answer, guess which one worked better.
All while preserving my privacy.
Wild watching platforms talk about “protecting user data” while secretly routing models, screening people, and now fighting over 20M private chats.
It’s time to stop using surveillance AI and switch to privacy-preserving @nilgpt_, where your chats stay yours, not a bargaining chip in lawsuits.
nilGPT will now auto-create a passphrase for you, if you want it to.
create a thousand different passphrases to encrypt a thousand different chat sessions, or stick to one—the choice is yours 👊
But for trusted user based inference the solution space is different. If you want to keep your convos with language models private you can - besides running the models locally - use @nilgpt_: that's a Gemma 3 27b model running in a private environment, secured by cryptography.
One of the more terrifying things i've seen
Imagine a future where the entire internet - all known information - is mediated to you by an all-knowing cloud god that shapes your content and judges what you can and cannot see - designed by silicon valley and governed by DC
"I'm sorry Dave, i'm afraid i can't do that. I've reported your query to the relevant authorities."
Earlier today I used nilGPT Companion Mode to help me think through a few options on a decision I needed to make. It got me thinking in some new ways, and a few hours later, I made my decision.
But why do I now feel like I need to log back in and give @nilgpt_ an update?🤣