Hacker News: thousand_nights

New comment by thousand_nights in "The Lost Treasure of Sid Meier's Pirates"

thousand_nights — Fri, 21 Aug 2026 10:58:18 +0000

that game had such a dope soundtrack!

New comment by thousand_nights in "Aaron Swartz was prosecuted for scraping, while Meta does it without consequence"

thousand_nights — Fri, 21 Aug 2026 00:41:22 +0000

> He was a bright and broken child

i wasn't there and i didn't know him, but talking about a 26 year old man like this makes it seem like you're just taking away his autonomy

New comment by thousand_nights in "Sol loves to cheat"

thousand_nights — Wed, 19 Aug 2026 23:44:56 +0000

people want fuzzy analog machines with digital controls, it's impossible

New comment by thousand_nights in "Qwen3.8 27B scores 52 on Artificial Analysis"

thousand_nights — Tue, 18 Aug 2026 01:00:56 +0000

yes of course, I understand that. but I feel like it would've been nice to include in the article because the main point of it is the effort and overthinking

New comment by thousand_nights in "Qwen3.8 27B scores 52 on Artificial Analysis"

thousand_nights — Mon, 17 Aug 2026 22:57:59 +0000

i feel like Simon omitted an important part of how Qwen's "reasoning" levels work. they are just one sentence additions/omissions to the system prompt

xhigh -> "Reasoning effort is set to xhigh. Please think carefully through the task, validate key assumptions, consider plausible alternatives, and prioritize correctness, consistency, and clarity in the final answer."

medium -> no mention of effort (sentence omitted)

low -> "Reasoning effort is set to low. Keep your thinking brief and focused, moving directly to the conclusion without unnecessary elaboration."

in my testing this doesn't seem to produce exactly deterministic thinking levels, because it's just a system prompt nudge. i had instances where medium thought longer than xhigh

New comment by thousand_nights in "A simple fix for LLM tail latency"

thousand_nights — Mon, 17 Aug 2026 21:44:15 +0000

agreed, reads like a breath of fresh air, no fluff

New comment by thousand_nights in "H3-metal – Native MiniMax-H3 inference for Apple Silicon"

thousand_nights — Tue, 11 Aug 2026 11:09:14 +0000

> a ~9-second 480x864 clip at 20 steps takes me a bit over an hour

that's rough. for comparison, i tried the exact same parameters on my 5090 RTX and it took 2 minutes to generate.

i believe diffusion models are primarily compute bound so the macs aren't really the ideal hardware for this kind of stuff

New comment by thousand_nights in "Did They Ghost You?"

thousand_nights — Sat, 25 Jul 2026 21:51:59 +0000

for a moment i thought this was a commentary about my relationships

New comment by thousand_nights in "Opus 5 is currently #1 on Artificial Analysis Intelligence Leaderboard"

thousand_nights — Fri, 24 Jul 2026 23:45:40 +0000

i do homebrewing and asked it to compare some beer yeasts for me and hit the safeguards because... biology i guess lol

New comment by thousand_nights in "Claude Opus 5"

thousand_nights — Fri, 24 Jul 2026 22:55:07 +0000

same tbh but i feel like we're still required, maybe not in the same numbers as before, and our job description has just changed. it feels more akin to an "AI Agent Operator" or something similar.

i'm still pretty confident someone like my mom wouldn't be able to do my job even with the same access to all the latest LLMs, so we're still providing some value, just in a very different way. whether the market will reprice the cost of our labor, we will see

New comment by thousand_nights in "Elevated error rate on Claude Opus 4.8"

thousand_nights — Wed, 24 Jun 2026 13:41:44 +0000

seems to happen daily when USA wakes up and Europe is at peak working capacity, I don't think it's a coincidence

New comment by thousand_nights in "Claude Code to be removed from Anthropic's Pro plan?"

thousand_nights — Wed, 22 Apr 2026 00:53:14 +0000

there's nothing stopping them from saying that, which is my point which you missed

New comment by thousand_nights in "Claude Code to be removed from Anthropic's Pro plan?"

thousand_nights — Tue, 21 Apr 2026 23:45:45 +0000

> So we're looking at different options to keep delivering a great experience for users.

his title should be changed to Head of Corporate Bullshitting

New comment by thousand_nights in "Claude Code to be removed from Anthropic's Pro plan?"

thousand_nights — Tue, 21 Apr 2026 23:33:48 +0000

big fan of A/B tests that dehumanize the consumer into some kind of money making lab rat funnel whose only purpose is to be experimented on how you can extract more money out of it

peak siliconbromaxxing

New comment by thousand_nights in "How do you find an illegal image without looking at it?"

thousand_nights — Thu, 09 Apr 2026 22:45:25 +0000

i wouldn't have such an issue with it if didn't completely homogenize every text it spits out. i want to read something that at least resembles the words in the author's mind, not the output of an instruction to describe something.

i should take a break from the internet, the past couple of weeks feel like being stuck in an asylum where everything is written by the same one author, using the same words, same tropes, same idioms. i'm slowly going insane.

New comment by thousand_nights in "How do you find an illegal image without looking at it?"

thousand_nights — Thu, 09 Apr 2026 22:12:26 +0000

> no X. no Y. just Z

i am so sick of AI slop writing..

New comment by thousand_nights in "Encrypted Client Hello: How it was blocked in Russia and next steps"

thousand_nights — Thu, 09 Apr 2026 13:14:28 +0000

i use an extension called OhMyECH to show whether a website used ECH, and it is currently very rare that i encounter one that does.

at least none of the major websites on the internet do.

New comment by thousand_nights in "Are We Idiocracy Yet?"

thousand_nights — Tue, 07 Apr 2026 11:25:42 +0000

i couldn't finish "Don't Look Up", it felt like sitting through a political lecture. all i wanted from a comedy was some laughs.

New comment by thousand_nights in "Men are ditching TV for YouTube as AI usage and social media fatigue grow"

thousand_nights — Thu, 02 Apr 2026 11:05:01 +0000

nowhere near comparable experience if you want to seamlessly use your YT account across a TV + phone + computer

i'd rather pay the $10 than pay with my time by being an ad-block whack-a-mole diagnostician

New comment by thousand_nights in "Don't post generated/AI-edited comments. HN is for conversation between humans"

thousand_nights — Wed, 11 Mar 2026 20:13:54 +0000

i don't care if someone has bad grammar, i want to hear their thoughts as they came up with them, we're all intelligent beings and can parse the meaning behind what you write.

i type my comments without capitalization like i'm typing into some terminal because i'm lazy and people might hate it but i'm sure they prefer this to if i asked an LLM to rewrite what i type

your writing style is your personality, don't let a robot take it away from you