Meanwhile, I will trust my education and years of experience as a radio operator over a website that anybody can edit.
Dipole antennas are decidedly not omnidirectional, but you can get a largely-omnidirectional radiation pattern using the strategy I explained. That is a fairly common antenna arrangement for 802.11 access points with multiple internal antennas per band because the orthogonal radiation patterns help make the access points orientation-insensitive. It also typically makes the multiple spatial streams the radios support take separate physical paths through the environment, which can improve speed and reliability of the link.
The biggest downside is that it takes up a lot of space inside an access point. As a result, some use a Y arrangement, which gives a passable approximation of omnidirectional radiation. It gets the nulls out of coincidence, but you wind up with a bit more of an oblate spheroid (think an M&M) instead of a sphere. It's usually workable, but a little more orientation-sensitive than three dipoles all at 90 degrees from each other.
All of this discussion actually makes me really curious what antenna arrangement the Almond+ uses internally. I'm betting a Y oriented in the plane of the LCD, but it's deep enough there could be an antenna running the depth of the router on two corners.