this post was submitted on 04 Apr 2024
91 points (85.3% liked)

Privacy

31609 readers
264 users here now

A place to discuss privacy and freedom in the digital world.

Privacy has become a very important issue in modern society, with companies and governments constantly abusing their power, more and more people are waking up to the importance of digital privacy.

In this community everyone is welcome to post links and discuss topics related to privacy.

Some Rules

Related communities

Chat rooms

much thanks to @gary_host_laptop for the logo design :)

founded 5 years ago
MODERATORS
91
DuckDuckAssitant is back! (media.kbin.social)
submitted 7 months ago* (last edited 7 months ago) by [email protected] to c/[email protected]
 
all 34 comments
sorted by: hot top controversial new old
[–] [email protected] 59 points 7 months ago (1 children)

Beware that it just forwards your queries to openai

We have agreements with model providers to further protect your privacy.

As noted above, we call model providers on your behalf so your personal information (for example, IP address) is not exposed to them. In addition, we have agreements in place with all model providers that further limit how they can use data from these anonymous requests that includes not using Prompts and Outputs to develop or improve their models as well as deleting all information received within 30 days.

[–] [email protected] 35 points 7 months ago

Important question: exactly what does it forward to them? After reviewing their privacy policy, it looks like they may be forwarding something in addition to the query itself.

When it comes to adding a layer of privacy between your queries and the OpenAI results, I believe them, but I really want to know exactly how much privacy there is and isn't.

[–] [email protected] 39 points 7 months ago* (last edited 7 months ago) (1 children)

FYI: You don't have to choose OpenAI. You can also choose Claude 2.1. Its corpus is more ethically sourced & it's more private with user data, but in return it is naturally less feature-complete. Still pretty awesome though!

[–] [email protected] 9 points 7 months ago* (last edited 7 months ago) (3 children)

ChatGPT begs to differ

(ETA: I'm pretty sure ChatGPT is wrong on this one, but it was amusing at least)

[–] [email protected] 17 points 7 months ago (1 children)
[–] [email protected] 12 points 7 months ago (1 children)

ChatGPT: it tells you what you want to know!

(And sometimes what it tells you is even true!)

[–] [email protected] 1 points 7 months ago

This. I guess you can save the permanent prompt to "output an error if the certainty of a result is below 50%" or something

I use that prompt to remove annoying talking, give me a single command and not "open nano here, copy this".

[–] venoft 10 points 7 months ago

That's super wrong. Typical ai hallucination since it's not in training data (Claude is quite new).

[–] [email protected] 10 points 7 months ago* (last edited 7 months ago) (1 children)

ChatGPT is incorrect. I asked Claude about it and this is what it has to say.

(Skip to bottom for tl;dr.)

 

 


 

tl;dr

Claude's end response:

when I said earlier that "Claude acts as a privacy layer provided by DuckDuckGo", that was an incorrect statement. To clarify:

  • DuckDuckGo provides the privacy layer functionality through this chat interface.

  • I am Claude, an AI model created by Anthropic to converse.

  • When a user talks to me (Claude) through this interface, DuckDuckGo's privacy layer protects their data and anonymity, while I focus on the conversational aspects.

  • I do not act as or provide the privacy layer - that is solely DuckDuckGo's role.

Thank you for catching my mistaken statement earlier. To summarize correctly:

DuckDuckGo enables the privacy layer protections. I am Claude, having conversations through that protected interface, but I am not involved in providing the privacy layer itself.

I appreciate you taking the time to ensure I understand this properly. Please let me know if any part of the explanation remains unclear!

[–] [email protected] 2 points 7 months ago (1 children)

That thing needs to summerize it's own shit.

[–] [email protected] 1 points 7 months ago

I fucking AGREE.

[–] [email protected] 15 points 7 months ago
[–] [email protected] 11 points 7 months ago* (last edited 7 months ago)
[–] [email protected] 10 points 7 months ago (1 children)

Damn they have ClaudeAI! Its way better than GPT

[–] [email protected] 11 points 7 months ago (1 children)

Unfortunately, they are using the old Claude 1.2 instant model instead of the newer Claude 3 Haiku

[–] [email protected] 3 points 7 months ago* (last edited 7 months ago)

I also couldnt find their website at all, how do you access that chat?

Edit: tab on the search results page

[–] [email protected] 8 points 7 months ago

Meh. I'll stick with local llama and firejailed lmstudio.

[–] [email protected] 7 points 7 months ago

Better alternatives are always welcomed, even if they're not universally needed or perfect. I already use their search engine quite regularly so I can imagine I will try this occasionally, though I think trust issues with AI chat will always be difficult to overcome for me since I've spent the majority of my life doing this stuff for myself.

[–] [email protected] 5 points 7 months ago

It's funny how fads have moved on from fashion and toys and onto the internet

[–] [email protected] 4 points 7 months ago

They really think their target audience will like it?

[–] [email protected] 4 points 7 months ago (1 children)

What is the link? It doesn't appear on main ddg page for me

[–] [email protected] 4 points 7 months ago* (last edited 7 months ago) (1 children)
[–] [email protected] 2 points 7 months ago (1 children)

Thank you! I've used anonchatgpt before but this is better, no captcha and two models to choose from

[–] [email protected] 1 points 7 months ago

Its strange that you need to search something first to access it. And the Claude model is really outdated. The new claude model is better than ChatGPT4 but the old one isnt.

[–] [email protected] 1 points 7 months ago (1 children)

I'll just use Jan locally on my machine and Microsoft Copilot for images.

[–] [email protected] 2 points 7 months ago (1 children)

Which models do you run on jan? I often get very weird responses and/or never ending feedback loops.

That being said I still don’t have much use for text-based ai, it doesn’t really fill any gaps in my workflow

[–] [email protected] 2 points 7 months ago

Stealth 7B mostly