My last troll with somebody who seems to be a libertarian with Trump Derangement Syndrome timed out with a question:
"Are you a bot? Your answers look like a bot"
Answer- to a newbie I'm not surprised I look like a bot. I'm autistic and I focus on subcategories of any given conversation that most human beings wouldn't. Also, I'm a veteran of the late 1980s Fidonet Flame Wars (when you could easily wait a weak or more for QWK packets to be passed across the continent) and the 1990s Usenet/Listserv flame wars (in that I was attacked by one of the first bots, Sedar Argic, a bot programmed by a Young Turk to try to convince everybody that the Aramaic Genocide really happened in reverse, killing Turks rather than Aramaic Christians).
From that, I developed the following four rules for online debates that make me extremely pedantic and frustrating to debate:
1. Ignore all personal insults. Personal insults are irrelevant to the conversation. Corollary to 1: Few people do this and thus personal insults are very useful to throwing your debating partner off topic, as long as you stay on topic yourself.
2. Take it Meta. Most subjects are far deeper than neurotypical non compos mentis realize. It can be as little as "I became pro-life because I noticed autistics like me are being targeted for genetic deletion"- I recently had a pro-choice feminist tell me she never forgot that, even after a couple of decades. So dig deeper, eventually you're going to find a point of agreement.
3. The conversation is not for the person you're debating. Your real audience is the 8-10 people reading the thread that will never comment. They deserve to hear both sides at their best, so be at your best. Don't sink to the level of the person you're debating with.
4. Have set principles and never change them unless absolutely disproven. Be a reverse skeptic on this, and be skeptical as possible. Force the other guy into trying to prove a negative.
I suspect as time goes on and AI gets better at these 4 rules, AI will win more debates.