What Is Deepseek?
Deepseek is a Chinese LLM (large language model) that aims to rival OpenAI’s ChatGPT-4 and the like, developed in China.
The buzz in a nutshell
Deepseek made headlines last week, claiming performance on par with ChatGPT-4 at a fraction of the training cost. This brutally drove down the stock prices of major tech companies and chipmakers, such as Nvidia, which lost 16% of its value in a single day.
But the hype was short-lived and shouldn’t compel us to switch to Deepseek, not even try it or use it along our favorite LLMs.
- Chinese government access: It seems DeepSeek coding has the capability to transfer users’ data directly to the Chinese government. See this article from abcnews.
- Deepseek training cost: The low training cost announced by Deepseek is what made the buzz, not its performance. The training cost announced, unverified, ignores the actual investments in hardware.
- Deepseek vs. ChatGPT: While Deepseek may be advanced, there’s nothing you would do with it you could not achieve with more trusted LLMs.
Better (and Safer) AI Alternatives
If security matters to you, here are some LLMs with better privacy policies:
- ChatGPT-4 by OpenAI – The leading AI model for safe and reliable answers. ChatGPT
- Claude by Anthropic – A privacy-focused AI chatbot. Claude AI
- Mistral AI – A good AI with extensive information access. Le Chat (French humor)
Want to Stay Offline?
If you know how to open a terminal on your computer, you can run LLMs locally. I am using Ollama, which is easy to install and configure. You can download and use offline a lot of different open-source models, as the powerful Llama3 or… Deepseek.
Final Thoughts: Stay Secure
The Deepseek buzz, which has nothing to do with the model’s capabilities, makes us, curious animals, want to give it a try. Just don’t, red flags everywhere. If you want to test it, try it offline using Ollama.
Reading advice about LLMs
Note that as an Amazon Associate, I earn from qualifying purchases—but this does not affect the price you pay.
- Basic and practical: Understanding Generative AI for beginners gives a practical introduction to using LLMs, with optimization advice and prompt examples.
- Mastering the Art of the Prompt: Prompt Engineering for Generative AI: Future-Proof Inputs for Reliable AI Outputs
- Build One by Yourself, the best way to understand things, with Build a Large Language Model (from scratch)
- Science Fiction that is not fiction anymore: Have you read Neuromancer from William Gibson? This author invented the term “Cyberspace” in this book in 1985 and is clearly an inspiration of the Matrix movies. His books include LLMs that blur the human-machine frontier.
And as always
Stay sharp, stay fearless, and keep being an Everyday Hero. Want to discuss today’s topics? Email me at [email protected]
For Everyday Heroes is a media outlet where I share expert insights to support fellow everyday heroes.
✔ Verified information, trusted recommendations.
❌ No bias, no hidden agenda, no subscriptions.
Come, use, go.
I earn compensation from the carefully selected products listed above—products I use or would personally recommend.
More information on foreverydayheroes.com.