Replies (72)
src have that order and preserve is including ChatGPT
OpenAI by "temporary court to chats" now use all Don't deleted.
https://xcancel.com/kepano/status/1930403693310157159 would logs ChatGPT required been requests API
I came here to recommend this
And not only because of that. They manipulate him an awful lot. Grok is better.
Didn't everyone assume they were doing this all along and sharing the data behind our backs with 3 letter agencies throughout the the Israeli/US web of privacy infringing five eyes countries?
You should assume Grok is doing the same, even if it isn't.
Why is this a surprise 🤷 It's an american tech giant. They are obligated to keep logs.
Maybe it does, but I've compared them and chatGPT is much more censored and full of propaganda.
yes but now it is more obvious
and unfortunately most people will use it anyway
Better yet fill their database with convoluted drivel
View quoted note →What are currently the best alternatives? So much going on in the ai world, hard to keep up.
This is why I have never used it, but I have used nano-GPT.com once or twice.
@Start9 with FreeGPT is one
Thanks for the heads up
No surprises there...
Good thing they've been doing it the whole time, in case the state needs to see the historical logs you thought were private since you checked a box on a UI
Is this only ChatGPT or all private AI language models like Grok
if you need it, use PPQ.ai and pay with monero?
for now ppq.ai does a good job of being a proxy for queries.
pay with LN or monero
Wendy is a safe one to use?
duck.ai is my choice for now.
private, anonymized by duckduckgo and not used to train new models.
at least that's what they say....
duck.ai is my choice for now.
private, anonymized by duckduckgo and not used to train new models.
at least that's what they say....
View quoted note →This is why I continue to use Venice, decentralized.
of course
Venice is the same setup as this, unfortunately. Nobody can see what's running on their servers and they could be ordered to retain logs.
Best way for privacy is to use encryption like we do with
@Maple AI.
Venice is a business promise, just like "temporary chats" with ChatGPT. All chats proxy through their servers, they just promise not to log them.
Check out
@Maple AI, it uses private key encryption so we can't see your chats at all.
Trust encryption not business promises.
@Maple AI can't read your AI chats.
Use a burner account not associated to your other id's. That's what I do and I only access it via VPN.
Give us a try. All the latest and greatesst AI models in one place. No credit card or account required. Pay with Bitcoin/XMR.
Give us a try. All the latest and greatesst AI models in one place. No credit card or account required. Pay with Bitcoin/XMR.
I have heard of it once or twice, but I haven't tried it yet.
I'm good ty. I run my own isolated model
I think the big difference between something like ChatGPT and Maple is that GPT can get up to date web results making it much more useful. Unless I am mistaken? If not is this something Maple can implement in the future?
Venice is nice, but reeks of shitcoinery (Voorhees). Nevertheless, been using it as my go-to & was compensated handsomely. Immediately dumped their shitcoin into bitcoin though.
Nice to meet you tony 🤝😉
Nice to meet you friend 😉🤝
Nice to meet you friend 🤝😉
Nice to meet you 🤝😉
Nice to meet you friend 🤝😉
Fuck chatGPT 🖕
what a huge surprise, so they have to "comply" with the law... 🤪
I am even thinking they were just waiting for this to claim they don't have choice and must save all data.
#4yoursecurity
hello hello 🤝
trust but verify.
How can you verify that in real ?
Loving this so far. Had some issues initially using opus but seems to be resolved. Nothing beats paying with lightning
Can you specifie what the exact numbers are for starter and pro?
"Starter: Enough messages per month for a casual user
Pro: Great for heavier workloads with a high monthly cap"
You as well 🙂👊
Could you help me use encryption, I’m not aware how that works?
I stopped using it, knowing the writing on the wall.
Use HuggingChat instead.
Hey Marks, love the idea of Maple AI. I've used it before.
Do you have an API for it? I use OpenAI's API quite a lot
The encryption is automatic. Secure enclaves on the Maple servers handle it for you. And for the watchdogs on the internet, all the Maple code is open source and verifiable.
That green "Verified" badge let's you know you're securely connected and encrypted.
It's tough to be specific because every chat uses a different amount of compute. And as we add more models they each have their own cost structure. Our messaging of "Pro gets 5x more messages than Starter" is based on averages of how much individual chat messages use up across the platform of all users.
Starter plan might be going away in the future when we have more significant features, like document upload. Will likely allow our early adopters to stay on that tier for a while to say thanks.
We provide a usage meter in the app so people know where they are at. Always looking to improve the experience, though.
😆 turning your AI into sats
The crypto token piece of Venice is a head-scratcher and not something Maple will get involved with. We accept fiat and Bitcoin as payment, that's it. If it uses the Bitcoin protocol, like ecash based on Cashu and Fedimint, that works as well.
Fiat...okay so maybe we deal with a bit of shitcoinery 😏
Will this happen to the other??? Claude or Grok?
Every company already does this
Ok, understand. I was wondering why you are so vague about how much prompts one will get. Now it's a bit clearer. Thanks for the explanation.
From the user's perspective, however, it's not ideal for deciding whether the service is worth to buy if you don't know exactly what to expect for your money.
We need DECENTRALIZED AI asap!!! #realmantalk
For sure, understood. In my personal usage, I’ve never hit the Pro cap. And I use Maple to run the Maple business itself as well as my weekly show Freedom Tech Weekend.
With Starter, on average maybe you get around 150 messages and Pro around 750 a month. But that is all variable because some people paste in large amounts of content and go through quota faster than others.
750 is pretty decent, very very likely more than a user would need in one month. I thought about how much I use an AI, even 150 is a lot I think.
Our costs lowered two months ago so we increased everyone’s quota accordingly. Goal is to keep doing that.
Will definitely test it. Really like the privacy aspect of your service, hard to find these days.
@Marks it’s a typo lol. Meant to say “What Is”.
Lol Wendy though, good typo.
Check out
@Maple AI as well
Won’t happen to
@Maple AI because it’s encrypted
Thanks, I will check it out🍯🦡
I considered Maple as well. However, from what I tried to take a look at, what license is being used for it. Is it a Free Software license, by chance?
Ah very cool, I will check into it, thank you! So will that help with the ChatGPT issue they were talking about earlier? How does that work
They learn from us.
for AI ipso contosmo alatoism
Two out of three questions to AI should be absolute complete garbage, like that Italian singer singing garbled English.
