The good thing about twitter is people of expertise chiming in to deconstruct or fortify an argument. The whole process is like having a peek into a writer's draft of a book and get a feel for the process (word choices, flow, rewrites) rather than just reading the final finished "perfect" chapter. In blog scenario, the writer is embodying all these distinct positions and forging a path from premises to conclusions to make a case. Being a fallible human, he will prefer a path which makes his pre-decided conclusions stand out best.
On twitter, the replies that deconstruct or fortify an argument are interspersed with a lot of junk. I've built https://en.howtruthful.com/ as a tool specifically for deconstructing and fortifying arguments. I see a lot of potential to succeed where twitter failed.
"The test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function" — F. Scott Fitzgerald.
Infact we can generalize this to N ideas cases.
Take two ideas. A and B. Give weight ∆ and 1-∆ to them. Multiple the positions of those ideas with the weights. This gives you a new interpolated position. Now repeat this with new position and idea C. And so on. You will quickly cover the area of convex hull of these ideas.
I think with future advanced LLMs the above scenario is achievable.