As part of my job, I had the opportunity to test Augment. During those professional tests, I found it very relevant, almost “proactive”. By proactive, I mean: not only answering the request, but also anticipating unexpressed problems, proposing alternative solutions, and drawing attention to critical points… you get the idea. I was honestly impressed.
Augment offers a 7-day free trial for its coding agent, which was perfect timing: I was planning to run my experiments over the weekend anyway.
Starting again from the same scenario as Copilot
As with Copilot, I asked it to:
- start from Olivier P’s repository
- use the Markdown file generated by Copilot
- analyze the situation and find a solution to the name display bug in DCMOTO
And then… incredible: Augment found the solution immediately.
The miracle: the name finally displayed 📟✨
Augment generated MO5 code for me that:
- captures the name
- displays the characters in real time (echo works!)
- and as soon as I press Enter, displays: “Hello ”
And honestly… you can’t imagine the joy. I genuinely smiled when the name finally appeared on screen in DCMOTO. After hours spent with Copilot going in circles around the problem, it felt amazing to finally see the program working.
Pushing the tests further
Carried by that success, I wanted to keep going.
New objectives
- test moving a character on screen using the keyboard
- test graphical display
And once again… everything worked perfectly (after a few iterations, of course). Augment didn’t just execute instructions: it even proposed variants, possible improvements, and useful checks.
Building tools for what comes next
After these tests, I asked it to create:
- assembly libraries to manage:
- text display
- graphical components
- a technical foundation that I will reuse for future developments (C logic + hardware interaction in assembly)
Augment generated all of this, cleanly organized.
Then I asked it to produce several Markdown files explaining:
- how to develop on MO5
- key points to watch out for
- potential pitfalls
- strategies to work around them
- and lessons learned from my experiments
These files will serve as the foundation for my future RAG server dedicated to MO5 development.
Why a RAG for MO5?
My parallel goal, in addition to developing a small MO5 application, is to build a documentation server for Thomson developers who want to use a coding agent.
Because — and everyone who uses a coding agent knows this — the quality of the provided context is fundamental.
An agent, even an excellent one, does not invent the technical documentation of an 80s machine. It cannot guess the subtleties of:
- 6809 assembly
- the MO5 ROM
- memory-mapped addresses
- system routines
- emulator limitations
- differences between a real MO5 and an emulated MO5
For an agent to reason correctly, you need to give it the complete playground:
- working examples
- low-level explanations
- technical constraints
- compatible tools
- validated code patterns
- and a clear project structure
That is exactly what I want to build: augmented MO5 documentation (no pun intended), directly usable by an agent.
Wrap-up
After the mixed experience with Copilot, Augment was a real breath of fresh air. Relevant, fast, efficient, and genuinely proactive.
This day finally allowed me to cross a technical threshold: I can now correctly interact with the MO5 in DCMOTO, handle input, display output, and even start experimenting with graphics.
And that… changes everything for what comes next.
To be continued 🚀