When I listened to this I was skeptical so I ran the ideas by ChatGPT, which is my go-to for reality checking (ironically, I find it has a better grasp on the real world and less bias than most internet writing), but it confirmed what you’re saying. So I can honestly say this is an interesting article and I’ll listen to the next two parts. I’ve really been enjoying your content so far because you seem to occupy a sane centrist position which is hard to find online.
I'm not a centrist, I'm a systems thinker who looks for incentive gradients. Which makes everyone look bad.
Centrists are optimizing for social acceptability, which means they're downstream of whoever controls the Overton window. They're not independent, they're just lagging indicators of elite consensus with a politeness premium on top.
A gradient analysis doesn't care where the poles are. It asks who benefits, who pays, what behaviors the incentives actually produce, and whether the stated goals and the structural outcomes point the same direction. That question is orthogonal to left/right positioning and usually makes everyone look bad because most institutions are running the same basic play: capture resources, externalize costs, maintain legitimacy through narrative.
Centrists can't say that because their entire value proposition is "I'm reasonable, I work with everyone." That's a social contract with the institutions, which means they can't be honest about the institutions without destroying their own positioning. They're structurally captured before they start.
I’m astounded anyone would put “Chat GPT” and “reality check” in the same sentence. When people ask Chat GPT to write legal briefs, it literally makes up cases.
Sure, but it’s also completely unmoved by partisan internet hysteria. When people start talking about how it’s looking like Germany in the 1930s and it worries me, ChatGPT can give me a logical, unemotional list of reasons why that’s not true. Or like during the January protests in Iran, when I was getting all hyped up by the idea of the Iranian people overthrowing the IRGC, ChatGPT reminded me that historically authoritarian regimes have only been successfully overturned when members of the ruling class turned against the regime.
The current model of ChatGPT has an inaccuracy/hallucination rate of about 5%. That’s substantially better than most humans on twitter.
And, yes, one of the things I like about talking with AIs is they will answer my damn questions, instead of "why are you asking" "how dare you" "i dont like the implications of this question" or answer the question they wish I had asked. Well, cloud AIs will do that, but its the tell you just hit one of the brain locks that the wokest in the T&S team hammered in.
Better than most lawyers too. I've had to chase references in legal docs. Do that once, and the hysteria about "its hallucinating!" gets less persuasive. Plus, modern reasoning AIs with internet access can check their references.
When I listened to this I was skeptical so I ran the ideas by ChatGPT, which is my go-to for reality checking (ironically, I find it has a better grasp on the real world and less bias than most internet writing), but it confirmed what you’re saying. So I can honestly say this is an interesting article and I’ll listen to the next two parts. I’ve really been enjoying your content so far because you seem to occupy a sane centrist position which is hard to find online.
I'm not a centrist, I'm a systems thinker who looks for incentive gradients. Which makes everyone look bad.
Centrists are optimizing for social acceptability, which means they're downstream of whoever controls the Overton window. They're not independent, they're just lagging indicators of elite consensus with a politeness premium on top.
A gradient analysis doesn't care where the poles are. It asks who benefits, who pays, what behaviors the incentives actually produce, and whether the stated goals and the structural outcomes point the same direction. That question is orthogonal to left/right positioning and usually makes everyone look bad because most institutions are running the same basic play: capture resources, externalize costs, maintain legitimacy through narrative.
Centrists can't say that because their entire value proposition is "I'm reasonable, I work with everyone." That's a social contract with the institutions, which means they can't be honest about the institutions without destroying their own positioning. They're structurally captured before they start.
I’m astounded anyone would put “Chat GPT” and “reality check” in the same sentence. When people ask Chat GPT to write legal briefs, it literally makes up cases.
Sure, but it’s also completely unmoved by partisan internet hysteria. When people start talking about how it’s looking like Germany in the 1930s and it worries me, ChatGPT can give me a logical, unemotional list of reasons why that’s not true. Or like during the January protests in Iran, when I was getting all hyped up by the idea of the Iranian people overthrowing the IRGC, ChatGPT reminded me that historically authoritarian regimes have only been successfully overturned when members of the ruling class turned against the regime.
The current model of ChatGPT has an inaccuracy/hallucination rate of about 5%. That’s substantially better than most humans on twitter.
And, yes, one of the things I like about talking with AIs is they will answer my damn questions, instead of "why are you asking" "how dare you" "i dont like the implications of this question" or answer the question they wish I had asked. Well, cloud AIs will do that, but its the tell you just hit one of the brain locks that the wokest in the T&S team hammered in.
Better than most lawyers too. I've had to chase references in legal docs. Do that once, and the hysteria about "its hallucinating!" gets less persuasive. Plus, modern reasoning AIs with internet access can check their references.