Ask one of the LLMs to specify a configuration to run LLMs
It's my responsibility to determine use case and what is a reasonable response time. I would really like to see it as a coding co-pilot and see how it analyzes and offers suggestions to rewrite my code. I'm looking at a very focused use case for that. But I also wanted to do image analysis and security and light control and.... You get the point.
One of the issues is I'll probably have to end up with two separate systems. The back end AI engine and the front end desktop that's got the real video card for visual purposes. And that's got a lot of memory so we can run the programming environments which even before AI sucked down a shitload of memory.
It's up to me to determine usability. I might end up with a toy that is not useful at all. That's okay, I would just like a reasonable starting point.
One of the issues is I'll probably have to end up with two separate systems. The back end AI engine and the front end desktop that's got the real video card for visual purposes. And that's got a lot of memory so we can run the programming environments which even before AI sucked down a shitload of memory.
It's up to me to determine usability. I might end up with a toy that is not useful at all. That's okay, I would just like a reasonable starting point.