Hacker News Reader

Comment by WhrRTheBaboons | Hacker News Reader

i hope this was vibe coded, because I'd hate to think that tiny dark-gray text on a black background was a conscious choice

View on HN

Hi, author here. I didn't vibe code this and have put considerable amount of effort in explaining the ideas. I have used Claude for review and help with visuals. My aim was just to help others understand these concepts easily. Thanks for your feedback, i'll improve the design and reading experience.

My thoughts on reading it (and looking around your site) was that you were building these explanations to help cement your own understanding of the topics and that you were just sharing the results because you thought it might be helpful to others as well. Knowing that you actually built it with the intent to teach just makes me think that it's nice of you to try to be a teacher even if there's a few folks who want to be jerks about your efforts. The world needs more teachers (and learners, too). Gaining, improving, and sharing knowledge is one of humanity's true super-powers. Without it, we'd all still be living like animals in the wild. Thank you for your efforts to contribute in a positive way. At least you're actually making stuff, unlike many of them what gotta always just complain about the things other people make and never actually make anything themselves.

Hi, thanks for your kind words. My intention is just to share my understanding and explain things in the way i wished someone would have explained me if i was not familiar with that topic. And it's a rewarding experience for me too, i do find something new about the topic or find some gaps in my thinking too. And since I am only starting, i am okay with all the feedback and work on my rough edges. I want the readers to walk away with new clarity and understanding.

This is a great article. The current knee-jerk reaction to AI is nothing but human negativity bias. There is a colossal amount of information here akin to what you might find in an intro to computers class. I appreciate seeing things like this.

Thanks, glad you liked it.

for the record, i only meant the visuals/CSS. i just do not understand the LLMs' recent insistence on this inaccessible design language and its ubiquity. vibe-coded frontends have always looked poor but at least they were readable until this recent trend took over.

there were no comments when i was posting mine, so i wanted to clarify i didn't mean to pile on the ai-written accusations. didn't seem like that was the case to me from the parts i've read. i'm just in the corner shouting at the claudes for being bad at frontend.

The text is also pretty clearly AI-generated. I guess there's now a market for "I asked an LLM so that you don't have to", but the funny thing is that it's such a wall of text that no one upvoting it will have actually read it. So it's vibe-writing, vibe-coding, and vibe-reading. Full end-to-end synergy.

Hi, author here. Sorry, you felt that way. I did put lot of effort in communicating my understanding, and used Claude only for review and visualizations.

I'd suggest splitting the content out into clearly defined sections, lead by the visuals.

Put the visualization and a short explainer, then have additional content show up if the reader drills in.

I don't think there is anything wrong with the content being run through an LLM. Networking is crazy complicated.

Thanks for your feedback. I'll try to do that.

Oh come on. Just seven days before this 60-page essay, you apparently cranked out another 40-page one:

https://fazamhd.com/mental-models/software/

That's 100+ pages of in-depth technical writing in a matter of days. Amazing, really amazing. Also, not something that a human can do.

And FWIW, Pangram classifies vast swathes of "your" article as entirely AI-generated. Yes, I know the tool is not perfect, but between your superhuman productivity, and all the subjective tells, and the output from that tool, it goes onto my mental "AI slop" pile.

What I can't understand is why people can't just own it that they used a chatbot and that it was more than just "for review".

Case closed.

Short publication intervals. That's the heuristic you're going with? I have dozens of such unpublished long form articles on fundamental topics, just waiting for me to carve some time to polish. I expect that on a sabbatical I could finish and post them in quick succession. It's an unreliable indicator to determine AI-writing.

Author here, when I share something on my blog, i take full responsibility on the writing and the concepts being explained. This couldn't have been generated with just LLMs. If it was that easy, i wouldn't have bothered to work on it. I am sorry you felt that this was just slop, but i really thought people would enjoy reading it.

I have my rough drafts ready for all the topics which i want to explain. And I actually spent majority of my week working on the networking article, it's a lot.

How were you able to pick this up? Not challenging your assertion. Just really curious. Can you point to some clues? I read it (with my own eyes). I can't see actual evidence the text itself is artificial or at least, that it is not human-curated.

I wonder whether mathematicians got the same reaction when calculators appeared: "You didn't compute every step by hand, so the result doesn't count."

It seems that author didn’t check vibe coded result. Mobile version has tiny text even even smaller… :)

At HN, there should be some tag explaining the project is vibe coded.

Hi, author here. Sorry for the bad experience. I'll put more effort on the design and reading experience

Issue is LLM itself. For design and typography it needs precise instructions otherwise you get output that it is information-heavy but churn rate is high as ppl can’t use (read) it.

I have to control output everytime but after 25 years in magazine layout I know what has to be controlled.

Tailwindcss has typography plugin named prose. And shadcn/ui recently published typeset package as well.

Emilkowalski/skills is good too.

Just add something more specific into prompt and the result will be much better. Cheers.

Hi, thanks a lot for your suggestions. Design is one area where i would like to improve.

Reply on news.ycombinator.com