Very confused, is this all BS?
"Confer, the private AI chatbot launched by Signal creator Moxie Marlinspike, uses that Nvidia-based TEE setup— along with an authentication system...."
"Meta, too, uses a TEE to offer a more private version of the Meta AI chatbot included in WhatsApp—a privacy system that Confer’s Marlinspike has said he is helping Meta to develop."
This feels too good to be true and I am surprised this article is the first i am hearing about the relationship between Marlinspike and Meta for this product.
19 Comments
twoBrokenThumbs@lemmy.world · 11 pts · 1d
I'm no expert so take this with a grain of salt.
But basic logic says if the data (your prompt) gets processed by anything then it needs to be able to read it. Therefore its not private by default.
Sure, they can encrypt it (where? At rest? End to end?) and not retain history etc, and that can make it more private, but it won't be private.
That's like Proton's Lumo which is "private" in the same sense, but the requests are being sent to a server to process so that data is read by something somewhere. I use Lumo BTW, but never put personal info into it etc.
Then no matter how private it is, if Meta starts using it it's immediately spyware.
If you want true private you'll need to run your own local models.
un_ax@lemmy.today · 4 pts · 1d
The TEE mentioned in this article is essentially hardware level identity/encryption similar to the boot chain of a phone or game console, so if you trust nvidia then it sounds like it would work. If you trust nvidia. And all of the other implementation details.
There is actually Homomorphic encryption where you can operate on encrypted data, but I don't imagine it's used much in machine learning since it takes 100-10,000x longer than the plaintext operation.
utopiah@lemmy.ml · 1 pts · 1d
It's two different problems :
technically speaking TEE and homomorphic encryption have proven to work. You can read public research and results from e.g. NIST or ask a researcher in the domain you trust but basically unless you find a way to break it (which is not impossible theoretically, just nobody has ever done it), it is indeed private.
in terms of business model, Meta is an advertising company which only happens to rely on tech to sell more ads. The more information they have about someone, the more expensive they can sell ads to their true clients, advertisers. So... one can safely assume that everything AROUND the encryption itself will be used to align with that very profitable business model.
zdhzm2pgp@lemmy.ml · 6 pts · 1d
Lemmings, what are your thought on duck.ai?
helix@feddit.org · 4 pts · 1d
I like it. I treat it the same as any other AI apart from my local llama.cpp.
lagg_mine@lemmy.world · 3 pts · 1d
I tend to use it to help with homework for uni if I don't have access to a powerful enough device to run something locally.
zdhzm2pgp@lemmy.ml · 3 pts · 1d
Marlinspike seemed like he might be ok, and now he's working with a company that's been proven time and time again that it can't be trusted to do anything?
eremophila@lemmy.zip · 2 pts · 1d
Would rather talk to myself, it's saner.
eldavi@lemmy.ml · 1 pts · 1d
i do it because i'm a fan of intelligent conversation. lol
grey_maniac@lemmy.ca · 2 pts · 1d
It's paywalled for me so I can't verify this, but the quote provided looks like two separate statements. It does not necessarily mean Marlinspike works with Meta.
xela@lemmy.ml · 2 pts · 1d
It links to a blog entry from his AI project confer.to: https://confer.to/blog/2026/03/encrypted-meta/
"As Confer continues to operate as an independent entity, I will also work to integrate Confer’s privacy technology so that it underpins Meta AI. Meta is building advanced frontier models, so this will combine the most private AI chat technology in the world with the most capable AI models in the world. As Meta builds more AI products beyond the basic chat paradigm, the privacy technology from Confer will be a part of the foundation of everything that is to come."
grey_maniac@lemmy.ca · 2 pts · 1d
Thank you for the further detail and working link.
xela@lemmy.ml · 2 pts · 1d
Sure thing!
DieserTypMatthias@lemmy.ml · 1 pts · 1d
Just install Ollama and pull a model from their registry.
tiny@midwest.social · 3 pts · 17h
You need a minimum of 16gb of vram in order to do anything that requires tool calls and even then it is slow. Running llms locally is a fun challenge and it's impressive what can be done on consumer hardware but it isn't cheap or simple
DieserTypMatthias@lemmy.ml · 0 pts · 5h
Yeah. But you at least have the incentive to upgrade.
I still have my GTX 1660super I want to upgrade to something better so that I can save on my AI subscription (unlike a sub you actually own it and can run whatever open weight model you desire).
eager_eagle@lemmy.world · 1 pts · 1d
hirihit640@sh.itjust.works · 0 pts · 1d
If TEE is anything like Secure Computing, then it works something like this:
There's also ways to make sure you as a user/client are communicating with one of these trusted servers and not some malicious server masquerading as a trusted server, but I forgot how that part worked. You can look up Apple's announcement on Private Cloud Compute, or any article on remote attestation.
DieserTypMatthias@lemmy.ml · 1 pts · 5h
AFAIK Confer is independent from Signal.
reader@lemmy.zip · 0 pts · 1d
cold sun?