History of Building a local 'assistant'
<p>An article in The Register prompted me to concider building my own personal assistant. <a href="https://www.theregister.com/2025/08/07/run_openai_gpt_oss_locally/">How to run OpenAI's new gpt-oss-20b LLM on your computer</a> would suggest that I could be running a copy on the server which has 64Gb of memory and space left for another 64Gb. The GL710 graphics card is probably not up to the task but I don't need anything more powerful for graphics, so a dedicated 'AI' co-processor? Then I looked at the price ... but nothing stopping me having a play.</p>
<p>The one thing that I do need is something that monitors my daily activities WITHOUT relaying it all to some third party advertising engine! A question to Mistral flagged up a few options but also that perhaps it's time to break out the Raspberry Pi stash and look at home automation yet again? I've even got a touch screen on one of the packs which would be a useful adjunct and keep me away from the desktop for the personal stuff.</p>
<p>The one thing that I do need is something that monitors my daily activities WITHOUT relaying it all to some third party advertising engine! A question to Mistral flagged up a few options but also that perhaps it's time to break out the Raspberry Pi stash and look at home automation yet again? I've even got a touch screen on one of the packs which would be a useful adjunct and keep me away from the desktop for the personal stuff.</p>