Comment by usefulposter
1 day ago
>while Yegge made lots of his own ornate, zoopmorphic [sic] diagrams of Gas Town’s architecture and workflows, they are unhelpful. Primarily because they were made entirely by Gemini’s Nano Banana. And while Nano Banana is state-of-the-art at making diagrams, generative AI systems are still really shit at making illustrative diagrams. They are very hard to decipher, filled with cluttered details, have arrows pointing the wrong direction, and are often missing key information.
So true! Not to mention the garbled text and inconsistent visuals across the diagrams———an insult to the reader's intelligence. How do people tolerate this visual embodiment of slurred speech?
Yeah I couldn’t figure out if they were just intended as illustrations and gave up trying to read them after a while.
Which is unfortunate as it would have been really helpful to have actually legible architecture diagrams, given the prose was so difficult for me to untangle due to the manic “fun” irreverent style (and it’s fine to write with a distinctive voice to make it more interesting, but still … confusing).
Plus the dozens of new unique names and connections introduced every few paragraphs to try to keep in my head…
I first asked Gemini 3 Pro to condense it to a boring technical overview and it produced a single page outline and Mermaid diagrams that were nearly as unintelligible as the original post so even AI has issues digesting it apparently…
If you disentangle the prose, it doesn't get better. It reduces down to IDK, but it makes output. That is all it can reduce down to. I love using Claude on my code, I love refactoring via broad sweeping statements to an agent, but I don't love trying to refute charlatans. Yeggae can go fuck himself with all the dildos his token can now buy (for a limited time only!).
I generally am a fan of polished writing, but I do believe that there's room for quickly fired experimental stuff, and quite enjoyed this piece. With the speed he was going, I wouldn't be surprised if the system architecture actually changed in between subsequent sections of the post. It's not a scientific article, but just a cross-country runner at the top of his game giving us a quick update without breaking his stride, and I'm all here for that.
As Basil Exposition said "I suggest you don’t worry about this sort of thing and just enjoy yourself".
Sure. The critique is with his incoherently labeled images, not his prose or passion project.
If LLMs can't produce readable technical diagrams ("Figure n"), avoid them for diagramming.
Don't insult the reader with strings like Oarity, Ed3e Csess, Ouclstnoc, relinemen, ressore Critieal, Foll Throughput, Witnese/Refin, ecstate ta Mayor, and Hsoide Fulures Stafey Fen.
https://miro.medium.com/v2/resize:fit:1400/format:webp/1*X3z...
https://miro.medium.com/v2/resize:fit:1400/format:webp/1*7Cr...
https://miro.medium.com/v2/resize:fit:1400/format:webp/1*blw...
I like how ‘refine 1’ is correctness, followed by ‘refine 2’, which is of course ‘oarity’.
I’m convinced that AI has totally broken the brains of its users; how on earth could you look at this thing and think “yes, that is a reasonable thing to post in public”? For whatever reason, for many users, the lesson of the AI revolution seems to be “just produce arbitrarily shoddy nonsense, it’s fine, nobody cares anyway”… The worrying thing is that this may be _true_.
Thrilled to see someone else using a triple-em dash in the wild⸻keep holding the line.