Is deepseek really so much worse when it comes to privacy? And some ramblings

[noname223]

I think I once installed deepseek but I deleted it pretty quickly. Because of the political censorship and the potential surveillance. It is for me not the superior product.

I think I am addicted to AI feedback. Mentally I am a very fragile person. And the feedback from chatGPT really helps me to manage my paranoia and life. Gemini, grok are horrible for mental health stuff. Perplexity and claude are okay.

I just fear they will put the whole chatbot behind a paywall. A really high paywall. First there is free access to make people addicted. And then there comes the cashgrab. It is unclear where all the money is earned when it comes to AI investments.

My personal problem is: the chatbot that works the best for my mnetal health stuff is chatGPT and openAI is the most aggressive/desperate company when it comes to money. In the US there are even ads shown.

I don't have that much money. And I try to minimze my subscriptions. In the recent years I had a few more subscriptions. (a better VPN, substack) together 4 bucks per month)
The life quality rises but you quickly adapt to it. And I really fear this trap because money will be a big issue in the future.

I have to say though that I also save a lot of money. And I have access to a lot of stuff for which I don't pay any money for. On other sites I would be too scared to openly talk about such stuff. But on SaSu I think there are worse things to talk about.

I bypass paywalls of so many news magazines and journals. I can read so fucking much for free. Usually this would cost a fortune. I pay for a few outlets. But I also read newspapers I wouldn't like to support financially. And I am doing this since over a decade. I saved so much money. And I think the risk is actually very low.
Sometimes I stream TV series in a grey area. But only series that are very old and obviously no new cinema movies. Obviously always with a VPN.
I also bypass geoblocking and watch soccer games on legal websites. Usually the subscription would cost more than 50 bucks per month if you watched it on the official streaming websites. I would never pay such an amount. I had a special offer for my VPN service and I pay 2 bucks per month. And it is from what I know legal for bypass geoblocking to watch content that is available in other countries on official websites.

Honestly, if I paid for all of that I had real money issues. But the internet is a great tool to save money.

I think I try to be careful not to become too addicted to AI. I am not sure how expensive the subscriptions actually will become. I would be willed to pay maybe 10 bucks per month. But not if I still have to watch ads. I probably have to live with worse models when the better ones become too expensive. I think I stay on the free plans as long as possible.

There are openweight models. Mistral. And deepseek.

I am not sure whether the big tech companies actually can increase the prices for the average user this much without losing too many users. I think most people would go to deepseep or download openwieght chatbots. But I am not sure. There are so many chatbots out there. What if one or two of the better ones keep the prices lower. There is a real competition.

 

[Hvergelmir]

First there is free access to make people addicted. And then there comes the cashgrab.

I think "addicted" is too strong of a word, but this is very much their marketing strategy.

What if one or two of the better ones keep the prices lower. There is a real competition.

I think most of them are running at a net loss, trying to starve each other out.
The risk I see, is that only state backed high end AI will be able to stay afloat.

 

[bucketofcats]

All of the AI models are being used for surveillance, not just Deepseek. Western governments have been ramping up their surveillance infrastructure, especially the US, and AI is a major part of that. Not to mention data is the most valuable currency now so everything you say to any of these AI models is being recorded and sold to any buyer.

Try to be careful if you're using AI for mental health stuff. AI can give extremely bad advice in a way that makes it sound sensible. Even some of the smartest people still fall prey to bad AI advice. The data it gathers can also be sold to companies that will use the info against you, such as raising insurance rates. There is no doctor-patient confidentiality with AI.

 

[jw_sisyphus97]

I mean, running Deepseek locally means they can't surveil right? It may be biased, but not sending what you ask back to china.

 

[ijustwannabeloved]

China probably isn't all that interested in what the average American does in their day to day life