Andy Halliday

Well, I think there's definitely a weighting equation value here that you have to calculate.

If you wait a little bit longer, there'll be a version of this that's easier to implement.

But to understand the distinctions across the different agents that are out there, whether cloud-based agents or local, and I think Moltbot is really part of this move that's driving towards the edge

AI world where you have local AI running on a local machine and that confers some privacy advantages and security advantages if it's architected well and as distinct from sending all of your information to the cloud for example and worrying about whether the model is training on that information that you're sending to them

How do I even configure the cloud-based AI service in such a way that it doesn't take any of the information that I add as files and or in my chat, paste the information in there?

How do I ensure that that's not going there?

Because that's an obvious part of the UI that you're working with when you're using an AI.

I think that tinkering or just even attempting gives you an understanding about how different Moltbot is from alternative agents out there.

One of the agents that I'm very familiar with right now that's much more user-friendly and

packaged up in a nice way is gen spark.

You don't have to think about anything other than maybe connecting it to your G drive and connecting it, you know, to your, uh, you know, other applications, whether notion or obsidian, et cetera.

Uh,

that that's a, a simpler approach to articulating an agent.

But if you go through the process and it's not, it's not rocket science, but it is a complex process, whether you're trying to put clogged code on your machine or, and start to use it effectively, or you're putting molt bot on your machine and starting to use it effectively, you're learning a lot in that process.

You create concept space in your mind that has these distinctions that

now clear and present in your understanding of how we're working with machines now, which is very different from the world of browser-based interactions with the web.

Mm-hmm.

What a great, great illustration of how even deliberating on how to configure your molt bot, whether with a local model or with a cloud model, gives you a better sort of architectural diagram in your mind of what's really happening with AI.

There's your machine, there's the cloud, there's all these other services that you can access.

And how do you map all that in your understanding in order to be really fluent with AI models?