😺 🎙️ PODCAST: Can you run DeepSeek on your laptop?

Welcome, humans.

A Chinese hedge fund just built an AI that's almost as good as ChatGPT's best model.

When it first came out, it crashed NVIDIA's stock by $600 billion in a single day. Jensen Huang personally lost $20 billion before lunch.

The company, DeepSeek, proved they can build an AI that rivals ChatGPT's best models—and they're giving it away for free (because it’s open-source).

So how do you use an open source AI model? Does this mean you can just run a ChatGPT-level AI on your computer? Not exactly… 

In our latest podcast episode, Grant and Corey break down the reality of running open-source AI:

• (4:00) The day that broke Wall Street: How DeepSeek crashed NVIDIA by $600 billion in one morning.

• (8:21) The hardware reality check: Why frontier models need 1.5 terabytes of VRAM (and what that actually costs).

• (8:37) Open source vs. actually usable: Understanding AI weights, parameters, and why "free" has hidden requirements.

• (19:15) The trade-offs you need to know: Security considerations when choosing between cloud and local models.

• (27:05) Practical options for experimenting with open-source AI on the hardware you do have.

Bottom line: You probably can't run GPT-4 level AI on your laptop yet, but Sam Altman thinks that might change by 2025.

Listen now on YouTube | Spotify | Apple Podcasts

Dive deeper with these resources:

• How to pick the best AI model for your needs

• Compare AI models and hardware requirements

• Our DeepSeek coverage and demos

• U.S. military concerns about DeepSeek

• ChatGPT privacy changes you should know about

• Tools for running models locally: LM Studio (our pick) | Ollama

Stay curious,
The Neuron Team

P.S. Grant is writing a step-by-step guide on how to install and run your first local AI model with LM Studio. Perfect for anyone who wants to experiment safely on their own hardware. 🖥️