Learning Timeline
Key Insights
No Terminal Required
Unlike older versions, the latest Ollama includes a graphical user interface (GUI). You no longer need to type commands in the Terminal to run AI models.
Automatic Download Trigger
AI models aren't actually stored on your computer yet, even if they appear in the dropdown. The download only triggers automatically once you send your first message.
The Advantage of Web Search
The web search feature is a unique capability in Ollama for GPT-OSS models, allowing the AI to access real-time information that might not be available in standard local GPT models.
Local Feature Limitations
Currently, running models locally via Ollama has some limitations compared to the official ChatGPT, such as the lack of a 'file upload' function and other advanced features.
Step by Step
How to Install and Run GPT-OSS Locally Using Ollama
- Download and install the Ollama application from the official website for your operating system (Mac or PC).
- Open the Ollama application once the installation is complete.
- Click the dropdown menu at the top of the interface to view the available models.
- Select the open-weight GPT model listed at the top (you can also choose other models like DeepSeek or Llama).
- Type any message or question in the chat box to begin the session.
- Wait for the application to automatically download the model (this only happens the first time you select a specific model).
- Use the 'web search' function in the model menu if you need the AI to find information online.
- Start interacting with the AI entirely locally on your machine without an internet connection once the download is finished.
