For those that were in the LIVE Session yesterday, I successfully downloaded Ollama onto my Windows laptop.
For my non-technical members, Ollama is an LLM like ChatGPT that you run local on your computer - with no internet. It is what people use that are concerned about privacy and the frontier models using your information as training data.
Claude Cowork was guiding me through the process, step by step.
After a few diagnostic tests, the screen said:
>>> Send a message (/? for help)
I began what is effectively the local AI version of "Hello, world".
Enter:
“Summarize this in one sentence: The quick brown fox jumps over the lazy dog.”
Ollama responded!
Wow.
Side note: It is wrong I am as excited as a kid in the candy store.
My own AI model.
Running locally.
On my laptop.
After a few more simple tests, Cowork suggested I verify that Ollama was truly working offline.
The instructions were clear.
☑️ Turn off Wi-Fi.
☑️ Enter prompt.
☑️ Wait for response?!?
Check. Check and Check!
Holy Smokes! It actually works.
Mind you, I am using Cowork to guide me through what feels like my first few steps on Mars.
So... Now I decide to really test it with a complex prompt:
"Can you write a brief story about two 12 year old boys throwing a football in the front yard, dreaming about playing in the NFL?"
It produces a very nice short story.
Offline.
On my laptop.
I am impressed.
I am proud.
Queue the tears...
So naturally, I copy the story and paste it back into Claude Cowork to await my atta’ boy!
Instead ⚠️⚠️⚠️⚠️
“We couldn’t connect to Claude. Please check your network connection and try again.”
OMG!
The F*&king WI-FI is off.
LMAO!
The local AI passed the offline test with flying colors.
The AI helping me test the local AI did not.
Now that is AI in Real Life.