Hm, Andros I do have my doubts - I know that you are a proponent of generative AI. We had a discussion over your outspoken opposition to https://human-emacs.org/.

I also know that you are an excellent writer/thinker, even prior to the advent of generative AI.

I'm in two minds here, and don't want to insinuate that you are outright lying, but I do believe your workflow relies on generative AI.

There is value in disclosing how you use generative AI to help you write articles. See https://www.bydamo.la/p/ai-manifesto or https://sive.rs/ai.

If you do use generative AI, may I suggest that you review the choice of licence: https://creativecommons.org/licenses/by-nc-nd/4.0/

> NoDerivatives — If you remix, transform, or build upon the material, you may not distribute the modified material.

If you do use generative AI, it feels hypocritical to use this licence while using a technology that essentially tries to circumvent the copyright issue by remixing, transforming and building upon the material it has ingested.

I don't know why you think you can lie about this.

https://www.pangram.com/history/3e369754-75da-4922-8386-c0f3...

(Cue pivot to "oh, I only meant that it wasn't slop, I never pretended that I wrote a single word of it.")

It's so, so clearly written by Claude. What are you talking about. Do you really think that you can fool people with so little effort?

> The number is real. But there's a catch.

> The gap vanishes. It isn't measuring WebSocket against SSE, it's measuring

> And here is the deeper part the reply misses:

> In short, it isn't a structural problem, it's a design one.

> The real argument isn't the wire, it's the architecture

> Where SSE genuinely wins

> What really separates the two models is the architecture