DeepSeek Primer

    
Plugin by: PHP Freelancer
This entry was posted in Editorial. Bookmark the permalink.
0 0 votes
Article Rating
Subscribe
Notify of
guest

4 Comments
Oldest
Newest
Inline Feedbacks
View all comments
Al Buckner
Al Buckner
1 day ago

Really, just $5 Million cost and you believe China? Those numbers are not adding up through the whole layout of though! I don’t think so. Anyway competition is good, Nvidia pull back is what I was looking for entry. Thank You DeepSeek.

MrLiberty
MrLiberty
1 day ago

Far too much HYPE around all of this. Feeling like the dotcom bubble. When you try and pretend that human programming will somehow over the limitations of our workforce (created mostly by the failed government schools), you are deluding yourself. There is always a place for improvement, but it seems like this is all about people making something out of nothing in hopes of getting super rich somehow.

Big Ruckus D
Big Ruckus D
1 day ago
Reply to  MrLiberty

Absolutely. The hype has been obscene, nobody has made a viable business case for any of this garbage they are trying to force on us (and thus far, the actual market is ambivalent at best about the use of AI, as it’s results mostly suck). This is being pushed as a top down thing by central planners, as is typical of stupid and dangerous ideas.

But now comes China with what appears to be a legit competitor, and has taken a whip to the techbro’s, suddenly threatening to disrupt the singularity envisioned by Larry Ellison and his ilk. I hope this Chinese upstart utterly destroys the US centric AI biz, it may be the best hope for humanity to derail what is being planned for us with this tech.

Hammers Thor
1 day ago

A couple things.

  1. I would never, EVER download an app AND sign up to use the app on my phone for any software that comes from China. Please see the accompanying image. I have tested out the DeepSeek OFFLINE Large Language Model, and it is actually pretty impressive. Having said that, I have since deleted it and have no intention of using it or any future iterations of the model.
  2. AI’s are a tool. Nothing more. They are not alive, they are not sentient, and they are quite capable of hallucinating and making errors. Nevertheless, they are enormously powerful tools that can, in many cases, save you hours or days of research time to finish projects faster, or answer questions that an ordinary internet search may fail at.
  3. AI is here to stay. The potential risks are every bit as terrible as the benefits are amazing. To marginalize these risks is stupid, and demonstrate a tremendous failure to understand how less-than-scrupulous people can use it for any number of possible evil purposes for surveillance, tracking, and scamming.

Imagine, if you will, you are using the Deepseek app. In order to use it, you have given it permission to know your location, access your camera, microphone, and files. Every bit of that data will become an entry on a Chinese server, and every inquiry you make will become part of that entry. Let’s say you are not thinking, and you ask it something about land that has been purchased by Chinese interests near U.S. Military bases. Or perhaps you say something unflattering about President Xi. Moments later, a Chinese drone flies overhead, releases a small package, and you are dead. You don’t think that is possible? Not only is it possible, it’s stupidly easy. Likely? Probably not [yet]. But Chinese surveillance is a thing, and will make our own NSA look like amateurs.

I refuse to use any ONLINE AI tools, with the exception of those that automatically answer questions in the search engine (like Brave Search).

I have tested, and use several offline models on my system in a sandboxed environment (LM Studio) for security. On March 1, 2025, Mike Adams is releasing his own model known as Enoch, and it promises to be a powerful and useful tool for anyone who wants to learn about any number of millions of topics, including health, medicine, prepping, survival, and just general knowledge without a WOKE leaning, which nearly every other offline AND online AI model exhibit. I have used his previous model Neo, which is good, but significantly less powerful than the upcoming model. I will try it out and provide a review once I have had an opportunity to thoroughly test its capabilities.

65253aa214738c69dc46a6a5c3229d6c_768x0