CHARBONNEAU – An AI all of my own: Yes, Dave, we can definitely do that

“I’m sorry, Dave, I’m afraid I can’t do that.”

— HAL 9000 AI computer, in movie 2001: A Space Odyssey.

I INSTALLED a personal AI model on my desktop.  I was curious  to find if a locally installed AI would relieve the use of data centres.

I’m calling my AI DAVE. DAVE didn’t arrive fully formed.

I asked: “Does this model run on my local computer without an internet connection?” DAVE didn’t know what I was talking about and replied: “I don’t have any information about a specific model. Could you provide more context or details about the model you’re referring to?”

Then I realised that DAVE didn’t know who he was so I had to tell him:

“You are Meta AI’s Llama language model running locally via Ollama. Does the Llama AI model require an internet connection to operate after it is downloaded?”

DAVE got it: “Hello! I’m happy to help you with your question.

“The Llama AI model, like other large language models, typically requires periodic updates to its training data and fine-tuning to maintain its performance. However, once the model is downloaded and installed locally, it does not necessarily require an internet connection to operate.”

It seemed to me that he should have a name, so I entered:  “Your name is DAVE.” and he cheerily replied: “That’s me! I’m Dave. How can I help you today?”

The Large Language Model I’m using is called Llama (Large Language Model Meta AI).

I wanted to explore my own AI after reading an article in the Globe and Mail by Professor Robert Diab  at Thompson Rivers University. He says that small, local, models like mine could reduce the need for all the monstrous data centres being proposed:

“AI data centres may turn out to be overbuilt, or even obsolete, by the time they are finished, because more of our AI processing is shifting from the cloud to our personal devices. The language models behind popular tools like ChatGPT, Claude, and Gemini vary widely in size. And smaller, open-source models you can run for free on your laptop, with no connection to the internet, are growing more powerful by the day.”

I wondered what DAVE had to say so I asked:  “Will small models like you reduce the number of data centers to be built?” He replied:.

“Small models like myself, also known as artificial intelligence (AI) models, can potentially reduce the need for large data centers in several ways.”

Once you get DAVE wound up, he can go on so I’ll summarize.

By processing data closer to where it’s generated, small models can reduce the amount of data that needs to be transmitted to centralized data centers. This can lead to faster processing times,

Small models typically require less data storage and processing power than large models, which can reduce the need for large data centers to store and process vast amounts of data.

These models can optimize resource allocation, reduce energy consumption, and improve overall efficiency, which can lead to reduced energy costs and a lower environmental impact.

If you want to reduce the number of data centres being built, get a personal AI. They have their limitations but save on the amount of traffic to data centers and possibly reduce the need for future builds.

David Charbonneau is a retired TRU electronics instructor who hosts a blog at http://closetproductions.ca/EyeView/eyeviewcover.htm