MI1
Official Website
MI1 was first proposed in early to mid 2006.
The above is the earliest mention that I have saved.
The origin of the method comes from email discussions that I used to have with Gilles Roux about creating new methods. One idea he suggested to me in early 2006 was a method with a Triangular Francisco like first step (before Triangular Francisco existed). I didn't like the look-ahead for all of the remaining edges, so I doubled the layers of the shape to make what is now MI1. After that step, I completed the first two layers then the last layer. When I showed the updated method to Gilles Roux, his response to me was something like "Sure. But it's just another way to solve the first two layers", which was a good point. I then played around with many variants to make the method more unique and to take advantage of MI1's shape. I eventually started communicating with Johannes Laire about the method and he helped find better algorithms for an L5EP variant.
The method was placed on a website, but only the main page remains archived.
https://web.archive.org/web/20110224050936/http://athefre.110mb.com/
I re-developed the method in 2020 and placed it on my Google site.
In July 2021 Cubing Forever proposed an advanced version of the method that ends in LXS
The steps are:
FB
dBr square
EO + DF or DR edge
LXS
LL
This proposal was inspired by Nautilus which has the LXS step in some variants. This was after APB had been proposed and during its development. Many in the development community knew about APB, but the method hadn't yet been fully developed and released to the entire community. So Cubing Forever's proposal probably wasn't inspired by APB, but instead by Nautilus. The above screenshot comes from the Nautilus server on Discord.