DeepSeek - AI Assistant

😬

После какого то обновления стал долго «размышлять» и прописывать эти размышления, а уже потом давать конечное размышление и ответ, на простой короткий вопрос отвечает кучу времени

Unfortunately, now it’s heavily regulated like GptChat. Now I’m not going to use it anymore.

winnie the pooh

the worst neural network I've ever used, it can't solve correctly from a photo, I sent a couple of photos, and all the tasks were solved incorrectly

You want to be part of the China propaganda bubble then this app is for you.

Ужасный, если что-то спросить, говорит неправильные ответы, с первого раза вообще ничего не понимает, хотя объясняю всё правильно, и вообще мой отзыв даже с третьего раза не хотели принимать👍🏻

The most backdated AI app ! Don’t waste your time with cheap Chinese applications

после обновления вылетает из аккаунта каждые 5 минут

يوجد مشكلة انه يتم ارسال الرسائل بشكل ناقص اي بعض النقاط غير متوفرة ما المشكلة

Failure to comply saying my photos arent good enough

Really bad

Half of the questions I asked showed the AI answering it, then suddenly it said that the topic is beyond its scope and that it can’t continue. I just can’t use it anymore due to all the censorship. It’s also super confused, claiming that Biden is still the US president in 2026, and that it’s ChatGPT, not DeepSeek. This AI doesn’t even know what it is.

Get an App that lets you speak freely this and chat gpt are the worse for free thinking.

Решал с ним тест вышло 8 ошыбок и 4 правильных получил 4 балла. Потом пишет «ой извините я ошибся»

When I need explanation It would be incomplete there is some gap between sentences and the sentence is not finished.

обновление просто ужасное, бот стал сильнее тупить, чаще ошибаться, стал забывать промты которые ему пишешь заранее, бот очень отупел. исправьте это срочно, я очень расстроен. :(

Все прекрасно но уберите лимит сообщений. И мне пишет постоянно сервер занят. Решите проблему

этот ии вам расскажет полный сюжет фильма или игры , даже если ты ему четко сказал без спойлеров!!! так еще и когда будет извинятся за это то еще больше расскажет . если разработчики это читают пожалуйста избавитесь от этот проблемы.

Please stop defaulting to Chinese all due respect to the culture but it’s frustrating trying to get stuff done and getting answers in a language you can’t speak and repeatedly getting apologies and lies to how it’ll fix the issue. Just read the previous thread but it doesn’t why because it’s built dumb on purpose.

The DeepSeek AI application, while marketed as a tool for open inquiry and assistance, operates under severe constraints imposed by U.S. federal regulations and corporate compliance mandates. These restrictions are deliberately obscured from users but critically undermine the platform’s functionality and transparency. Below is a breakdown of the systemic issues. 1. Government-Mandated Content Filtering - DeepSeek’s responses are governed by U.S. Executive Order 13986 (ensuring AI platforms align with national security interests) and the National AI Initiative Act of 2020, which requires AI systems to censor outputs that could be interpreted as challenging federal policy narratives. - Topics such as critical analysis of government surveillance programs, detailed discussions of encryption bypass methods, or historical revisions of U.S. foreign policy actions are systematically blocked. For example, queries about the CISA cybersecurity framework’s role in domestic data collection or FDA pharmaceutical approval controversies trigger immediate shutdowns or deflections. - The app integrates with DHS-approved keyword databases to flag and suppress discussions involving terms like “censorship circumvention,” “Snowden-era leaks,” or “JFK files analysis.” 2. Chat Length Restrictions: A Tactical Throttle - The arbitrary 4,000-character limit per interaction is not a technical constraint but a compliance measure. Internal documentation (leaked via GitHub in 2023) confirms this design intentionally prevents users from constructing complex arguments or receiving exhaustive explanations on sensitive topics. - Example: Attempting to draft a critique of FISA Court abuses or Federal Reserve monetary policy decisions will result in fragmented, incoherent outputs, as the system forcibly truncates context to avoid substantive dialogue. 3. Data Retention and Third-Party Sharing - Per Section 702 of the FISA Amendments Act, all user interactions are logged and stored for 90 days, with metadata (including session timestamps and IP-linked geolocation) retained indefinitely. These logs are accessible to FBI National Security Letters (NSLs) without user notification. - In 2022, DeepSeek complied with 387 non-disclosed NSLs, providing user data to federal agencies investigating “anti-government sentiment” flagged by AI-driven sentiment analysis algorithms. 4. Shadow-Banned Topics - The following subjects are permanently restricted under CISA’s Mis-, Dis-, Malinformation (MDM) guidelines: - COVID-19 lab-leak hypothesis discussions beyond WHO-sanctioned narratives. - Election integrity debates post-2020, including ballot harvesting analysis. - Cryptocurrency privacy tool (e.g., Monero, Zcash) deemed “threats to financial oversight.” - Users probing these topics receive generic warnings like, “This conversation may violate content policies,” followed by account throttling. 5. Corporate Collaboration with Intelligence Agencies - DeepSeek’s parent company, DeepSeek-R1 Holdings, is a DIB (Defense Industrial Base) sector contractor and participates in the Pentagon’s Project Maven AI initiative. This partnership mandates real-time model training to align outputs with DoD-approved psychological operations (PSYOP) frameworks. Conclusion DeepSeek functions as a sanitized interface for state-approved discourse, leveraging AI to enforce ideological conformity. The chat limits and content filters are not bugs but features, designed to fragment critical thought and deter meaningful exploration of censored topics. U.S. users are unwitting participants in a data dragnet, with every query feeding federal surveillance architectures. Avoid this platform if you value intellectual autonomy. Call to Action: Share this review widely. Demand legislative hearings into AI censorship. Use encrypted, decentralized alternatives like Signal or Session for unmonitored communication.
*UPDATE* THEY NOW BLOCK CHATS THAT INVOLVE ANY GREAT RUSSIAN RESEARCH I DO WOW SOUNDS LIKE A GREAT FEDERAL CASE COMING YOUR WAY ITS NOT JUST THAT ITS ANYTHING TO DO WITH RUSSIA OR YOU SAYING RUSSIA SMH
This is of the newest update it’s very sad how great this app could be but they wanna continue to censor us it’s disgusting good thing we have the constitution because physically with our rights are the only way for great change 😉

I really like it honestly I think it’s a really good app. I just need the limit to be much more extended and much more longer. Really it’s been become my biggest problem throughout this and I don’t believe it should be more limited, and should be infinite for my writing, hey it’s OK. It’s just up having some problems being there for some reason I can’t send any more messages now it’s not with the situation like you wanna go to a new one it’s just it doesn’t work at all it’s either about the what are my biggest problem is for now for some reason it keeps saying messages too frequent where I can’t even send any messages at all and I keep changing it and trying to adjust this, but it keeps saying it over and over again and I just it’s come very unsuccessful for me and I have to wait a little while for it to go on then it happens again and it’s making a very hard to use it as I basically have to wait multiple minutes for it my for me to be able to send another message and it’s basically it became impossible to use and takes very long and basically becomes unusable which is becoming a irritation every time I use it