Many are AI. Especially the ones who reply with automatically generated text and pasta. You can see it by constant usage if the exact same text.
One such example is tge muhjew spam chatbot that gets called "r2d2" or " rumba".
There are also the bots that post muhjew images with no text or provocation. I once replied to 30+ of its posts, just to see the reaction. The reaction was that it went to a frenzy state and finished the bread with 80+ image posts.
a theory on EBOT
From carefull observation it look like ebot is a sniffer thread, used for either automatic or manual flagging of posts nitable to shills, and it is also probably used as a way to pass coded messages.
How does it red flag the posts?
The bot system is one big multi-threaded app. Each bot is a separate thread (an operating system concept) and ot communicates through a distinct ip and socket. Also each bot thread allows a human operator to insert manual script, for whenever a human intervention is necessary. EBOT is also such a thread.
Whenever a good shill refuting/bashing post is present, ebot flags it by linking it in a post, together with a coded message and other unrelated posts. This can be done either automatically or by a human operator who watches the breads and does it manually. I have seen this pattern whenever ebot is present.
Why do they need this monitoring? Simple, to keep the AI bot behavior to be up to date so it wont feel like yesterdays' news.
As for the coded messages, i think that the bullshit meme and text, either serve as a coded description of the flagged and linked important post, or as a cided message between bad actors that use the board as an open communication channel.