DeepSeek R-1 on Your Mac: 4 Surprisingly Simple Local Setup Tricks

Author(s): Krupesh Raikar

Originally published on Towards AI.

How to use the power of AI with absolute data privacy

This member-only story is on us. Upgrade to access all of Medium.

Everyone and their mother (and even their 91-year-old grandpa living off-grid in the Andalusian mountains) has heard about DeepSeek!

(Seriously, if you haven’t, maybe you are from another planet. So here is a quick overview for you anyway)

DeepSeek is a Chinese artificial intelligence (AI) company founded in 2023 by Liang Wenfeng, a former hedge fund manager.

In January 2025, DeepSeek released its R-1 model. Technically, however, the R-1 stems from a large base model called DeepSeek V3.

The lab ingeniously refined this V3 model through supervised fine-tuning (SFT) on high-quality human-labeled data and reinforcement learning (RL).

The result was R-1, a chatbot that could handle intricate prompts, reveal the reasoning steps behind complex questions (sometimes more transparently than other models), and even render code in the chat interface for quick testing!

The surprise was, that all of it was trained using approximately 2,000 Nvidia H800 chips at about $5.6 million.

Naturally, this knowledge caused a massive market avalanche quaking off a trillion dollars overnight! And the DeepSeek app toppled ChatGPT off the top spot in the Play Store!

Then came all the critical questions:

Is the model safe to use?Is my data… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI