JetpackCat
My feedback
29 results found
-
61 votesJetpackCat
supported this idea
·
-
42 votesJetpackCat
supported this idea
·
-
67 votesJetpackCat
supported this idea
·
-
67 votesJetpackCat
supported this idea
·
-
59 votesJetpackCat
supported this idea
·
-
84 votesJetpackCat
supported this idea
·
-
80 votesJetpackCat
supported this idea
·
-
64 votesJetpackCat
supported this idea
·
-
143 votesJetpackCat
supported this idea
·
-
135 votesJetpackCat
supported this idea
·
-
175 votesJetpackCat
supported this idea
·
-
431 votesJetpackCat
supported this idea
·
-
3 votes
An error occurred while saving the comment -
2 votesJetpackCat
supported this idea
·
-
3 votesJetpackCat
supported this idea
·
-
1 vote
An error occurred while saving the comment JetpackCat
commented
Please keep in mind that Proton does not train their own AI models and they do not dedicate a large amount of resources to research and development in this area. They are a provider and host for open weights AI models that were published by Z.ai, Alibaba, etc. but with their own privacy focused architecture.
An AI chat bot makes sense for Proton because AI has become a core tool for many Proton users, and Proton aims to offer a private alternative to an otherwise privacy nightmare at other providers. An AI platform knows more about you than anyone else, and I hope we can rely on Proton to be that private alternative where we don't have to worry about surveillance.
Proton can easily capitalize on AI and make a good service whilst maintaining high standards in their other services. I don't see any lack of development for other tools.
-
1 vote
An error occurred while saving the comment JetpackCat
commented
An open weights image gen model via Proton does not go against privacy, and this one in particular is not going to be doing any of the things you described. People that generate "AI art" use specialized platforms that are a lot more performant for AI image gen. This addition is nothing but a bonus feature and I think most would agree that it's welcome.
While I agree that making and presenting AI generated images as art is not okay, this has basically nothing to do with Proton, and it really is hypocritical to say that AI image gen is stealing while not caring about text generation and actively using it to write this feedback.
LLMs generating text are even worse for the whole 'stealing IP' topic. If you wanna boycott generative content, you should stop using chat bots altogether and find one that is trained only on content in the public domain & specially licensed content.
I am sure that Proton understands this. Personally I love everything about this new 2.0 update and I hope they can keep improving the performance and usability of their platform for serious work while maintaining the promised privacy layer.
P.S. none of the models used in Lumo are made by Proton. They are just the host and provider for these open models listed in the Lumo Privacy Policy. Just wanted to say that in case you were unaware.
-
2 votesJetpackCat
shared this idea
·
-
35 votesJetpackCat
supported this idea
·
-
11 votesJetpackCat
supported this idea
·
This would just mean expanding the proton_info tool that Lumo already has access to.
The issue is that this would require a huge amount of documentation internally that would have to keep up with proton updates, etc. Proton can't just "train" their AI to know this. They use open weights models that they would then simply have to supply with the right information through a tool.