This discussion analyzes a "weave" diagram illustrating Donald Trump's speaking patterns, with participants offering a range of opinions on its accuracy, the nature of political discourse, the potential application of AI, and the technical presentation of the diagram itself.
Characterization of Trump's Speaking Style
A central theme is the unique and often criticized way Donald Trump speaks. Some users describe it as rambling, incoherent, or a "monologue" rather than a conversation.
- huflungdung describes the piece as a "Silly 'hit piece' for lack of a better term on Trump."
- bootsmann states, "This isn’t a conversation, he is holding a monologue."
- Adding to this, mnewme observes, "Not to that extent. Most public speakers try to convey one coherent thought they were working on before. Trump just rumbles like my granddad."
- AceJohnny2 expresses a strong aversion, saying, "Four minutes of the weave was about all I could handle... I wouldn't want to torture the author by force-exposing him more to Trump's inanities. That's already 5x longer than I can stand before the urge to tear my ears off becomes unmanageable."
- TheAceOfHearts goes further, characterizing Trump's speech as a "psychic attack" and a "cognitive hazard" designed to "hijack the listener's brain."
Comparison to Other Political Discourse and General Speech
Several participants debate whether Trump's speaking style is truly unique or representative of political discourse in general, or even common human speech patterns.
- atombender argues for a comparative approach: "This wasn't a conversation between ordinary people like you and me, but a statement delivered to reporters by a sitting US president. So the fair comparison would be to other such statement made by other sitting US presidents."
- keepamovin suggests extending such analysis: "That's funny. I'd like to see you do this more, on more of them. More diagrams! A good source is the introduction he gives at cabinet meetings. Or the way he answers questions with the press. You could compare with his campaign speeches and improvisations."
- Xmd5a offers a linguistic perspective: "Funny but unfair, we all speak like this, you have to sit down in a (psycho)linguistics class to figure it out."
- This is immediately countered by IshKebab: "We do not all talk like Trump."
- stevage provides a concise retort to the idea that "we all speak like this": "If we all spoke like this, noone would remark on Trump speaking like this."
- AlecSchueler raises a specific comparison: "Do you reckon you could take some of Obama's speeches and do the same thing as this with comparable results?"
Potential Applications of AI
The discussion frequently touches upon the idea of using AI to analyze such discourse, with some seeing it as a beneficial tool for handling tedious tasks or even a more accurate representation of certain cognitive states.
- MPSimmons suggests, "I hate to be like, 'Let's use AI for this' on everything, but this actually does seem kind of like a good thing for AI to do, rather than subject a human to it."
- However, AdeptusAquinas finds this idea problematic: "Seems really unfair to the AI."
- Drawing a parallel to literature, robaato quotes Douglas Adams: "video recorders watched tedious television for you, thus saving you the bother of looking at it yourself."
- ivape speculates on the accuracy of smaller AI models: "We act like small models are inconsistent and incoherent, but we rarely point out that it actually matches certain mental states and capacity accurately. We may need to actually see, how would a 0.5B model handle the presidency, because… it could be accurate. Having a super large model simulate these things would not be authentic."
- ivape further posits an interview test: "Might be the best interview test, you must generate hand written code that’s better than a small model (or show better judgement)."
- AceJohnny2 humorously downplays the benchmark for AI, stating, "For the presidency, the current benchmark to beat is GPT2 it seems." To which iamflimflam1 agrees about "word salad answers" and AceJohnny2 retorts, "You're generous. I would've gone with Eliza, but on second thought that's an insult to Joseph Weizenbaum."
Technical Discussion and Diagram Rendering
A significant portion of the thread is dedicated to the technical aspects of the generated diagram, specifically its rendering across different platforms and potential improvements for clarity.
- ggm suggests a longitudinal analysis: "Change over time. Get this done for 10y or more back non teleprompted lower edit recordings, chart and compare for some metric."
- The technical limitations of the embedded diagram are highlighted by several users. nilsherzig reports, "Does not render properly on iOS mobile (iOS 26 beta)."
- grues-dinner points out a similar issue: "Also not on Firefox." TristanDaCunha corroborates this, stating, "Same result on Firefox mobile."
- nilsherzig later clarifies, "only got the working on chrome desktop."
- ehnto describes a similar "weave" pattern in human speech: "Someone I know speaks in a reverse tree of sorts which actually does resemble a 'weave', they start with various statements about the topic at hand without ever mentioning the topic, and eventually arrive at stating the topic near the end (hopefully). Sometimes I have no idea what they are talking about because they forgot to mention it until the very end when they have merged all their branches."
- grues-dinner draws an artistic comparison to the structure: "Reminds me of a flash game (branch) or tool, art or general 'thing' perhaps (branch) I suppose it depends how you see it(return)(return) that I cannot not find (branch)such is the state of the modern web(return) (branch) something in the flavour of a Jared Tarbell piece (return) where you could type and the text would be displayed in a spiral in 3D space (branch) there was a way to make a branch a thought (branch) like this (return) and then return to the parent level. but even that wouldn't handle the weave (branch) not sure if that's a good thing or not (return)."
- Finally, shrx offers concrete feedback for improving the diagram's readability: "I think the diagram would be more comprehensible if the branch (topic) name would be shown next to the 'New Topic' label, not only at merges. I had to read it from the bottom up the first time to understand what's going on."