cubefox

25 ms·

631.
by cubefox 6mo ago
Wow, nice find, thanks! Here is the next picture, "Lesson 3 E. Morning Light": https://ibb.co/8DkdrG9q
632.
by cubefox 6mo ago
See also: https://ndlsearch.ndl.go.jp/en/imagebank/theme/hokusaimoyo Hokusai Moyo Gafu: an album of dyeing patterns (ndl.go.jp) 170 points by fanf2 10 months ago https://news.ycombinator.com/i
633.
by cubefox 6mo ago
More 1-bit pixel art: > MacPaint Art From The Mid-80s Still Looks Great Today - https://blog.decryption.net.au/posts/macpaint.html Previously discussed here: https://news.ycombinator.com/item?id=4454
634.
by cubefox 6mo ago
Yeah, I also just tried it, it doesn't taste bitter at all. It tastes like clear soup broth, which usually contains a lot of MSG.
635.
by cubefox 6mo ago
> So a little MSG to get your taste buds extra sensitive to other flavors is a net good. Salt and MSG are sometimes said to strengthen existing flavors, but I'm pretty sure they mainly just contribute their own unique taste: salty a
636.
by cubefox 6mo ago
Yeah. In the past I assumed that some people just sometimes randomly behave aggressively towards me for no good reason. But usually the reason is probably that I was unintentionally rude or strange with some sort of nonverbal communication
637.
by cubefox 6mo ago
Well, he argued in this 19 years old essay that syntax highlighting for programming languages is bad. Which might have been somewhat defensible in 2007, but is obviously completely false in 2026 AD.
638.
by cubefox 6mo ago
Abstract of the technical report [1]: > We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models — DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSeek-V4-Flash with 2
639.
by cubefox 6mo ago
Apparently I was not the first to have this thought. See https://english.edward.io/ https://github.com/SichangHe/natural_syntax
640.
by cubefox 6mo ago
> Everything should try do some basic syntax highlighting IMO. Interesting idea. So even syntax–highlighting natural language. Grammar highlighting, as it were. Prepositions, verbs, question marks, etc. An LLM could do it. Would it actua
641.
by cubefox 6mo ago
> and to note that it is the editorial team's responsibility to do things like check quotes. Publishing things online for free (as Ars does) is difficult business. I doubt they can realistically afford an "editorial team"
642.
by cubefox 6mo ago
It is completely unclear from this article whether this means Apple does no longer cache dismissed notifications somewhere.
643.
by cubefox 6mo ago
I checked all the images on the blog post and I'm quite sure that the one you talk about isn't there.
644.
by cubefox 6mo ago
To ask for verification whether a photo is real or fake?
645.
by cubefox 6mo ago
Not sure why this is voted down, it's true.
646.
by cubefox 6mo ago
That's the original pronunciation
647.
by cubefox 6mo ago
Dungeons contain mainly puzzles, but otherwise it's a lot of combat and story and side quests, similar to most RPGs.
648.
by cubefox 6mo ago
You forgot to consider whether all this is worth $60B.
649.
by cubefox 6mo ago
You will likely witness strongly superhuman AI, which dwarfs any changes your grandmother saw.
650.
by cubefox 6mo ago
Demon’s Souls??? I guess Zelda, Metroid and Half-Life also count as puzzle games then. :)
651.
by cubefox 6mo ago
It also had a large capacitive touchscreen rather than one with a stick and a keyboard underneath, impossibly smooth scroll inertia, a MEMS gyroscope for automatic landscape mode, etc. The GUI was also optimized for the capacitive touchscre
652.
by cubefox 6mo ago
This is the only paper which really does this: https://proceedings.neurips.cc/paper_files/paper/2024/hash/7... They train directly in the 1 bit domain, without any floating point weights. They don't
653.
by cubefox 6mo ago
The problem is that neither Sundar nor Demis are remotely as focused and competitive as Sam and Dario.
654.
by cubefox 6mo ago
Siri was GOFAI (handwritten software) rather than a model written by a machine learning algorithm.
655.
by cubefox 6mo ago
Current models don't yet use RLVR with self-play though, at least as far as we know. They use RLVR with large numbers of manually created RL environments. But they will probably use self-play soon. See https://www.amplifypar
656.
by cubefox 6mo ago
The NYT was right: https://www.nytimes.com/2026/01/08/technology/apple-ceo-tim-...
657.
by cubefox 6mo ago
They do use diffusion models, but I don't think they would make a detour via images. They can just generate audio directly with audio diffusion rather than image diffusion.
658.
by cubefox 6mo ago
Quite fitting then that it is called "unnecessary censorship".
659.
by cubefox 6mo ago
If you synchronize the seed at game start for speedruns, the seed is the same for everyone, and players can again manipulate their luck, so nothing was gained.
660.
Claude Code sometimes hallucinates user messages(lesswrong.com)
2 points by cubefox 6mo ago|2 comments

More ›