Journey to Agentic AI, Part 3: External Services and Git Powers

Author(s): JV Roig

Originally published on Towards AI.

Journey to Agentic AI, Starting From First PrinciplesCover generated by Qwen2.5-Max

This is Part 3 in the series “”

In Part 1 (From Next-Token to Tool-Use: How to Give LLMs the Ability to Use Tools) we first learned the fundamentals of giving LLMs tools.

Then, in Part 2 (Giving Control of Our Computer to an LLM), we expaded our tools to give our LLM basically full access to our computer, giving it free reign over our filesystem.

Of course, the potential of LLMs isn’t just limited to our own computer and filesystem. Today, in Part 3, we’ll explore connecting to external services by making our LLM be able to do operations on a GitHub repo — and of course, that means we probably also want to give our LLM git powers too!

LLMs are just next-token generators, so it kinda looks like they wouldn’t be able to connect to external services, right?

But that’s just rehashing what we’ve learned from part 1 — essentially:

LLMs can’t directly use tools, but we can wrap programs around them so that they indirectly use tools, with the end result being indistinguishable from actual direct tool use.

So, given that, it’s actually pretty straightforward giving our LLM friends external service access:

If we wrap a program around them that has… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI