yorwba

21 ms·

571.
by yorwba 9mo ago
It just means that there are two indistinguishable coordinate views a + b i and a - b i , and you can pick whichever you prefer.
572.
by yorwba 9mo ago
Mao tried picking industrial locations for survivability in case of war: https://en.wikipedia.org/wiki/Third_Front_(China) Turns out that places that are hard to reach for the enemy are also hard to reach in peacetime,
573.
by yorwba 9mo ago
Yes, in those movies, the hot white guys (and sometimes girls) usually ride on top of the muscular horses. So when you want to show a horse riding a man as a visual gag, why not make the man a hot white guy with a gruff beard? You act as th
574.
by yorwba 9mo ago
Gary Marcus successfully wrote an article about getting image generation models to show a horse riding an astronaut, which is all I needed him to do. (Actually he wrote two, but this one felt more concise.) Take it as an existence proof, no
575.
by yorwba 9mo ago
The message is "We watched Lord of the Rings and Game of Thrones and liked the medieval aesthetic enough to emulate it."
576.
by yorwba 9mo ago
There's definitely some hostility: https://mp.weixin.qq.com/s/A5shO-6nZIXZvJUEzrx03Q
577.
by yorwba 9mo ago
Where do you see a press release countdown? Alibaba consistently doesn't release weights for their biggest models, but they also don't pretend that they do.
578.
by yorwba 9mo ago
There's also the "horse riding astronaut" challenge in image generation: https://garymarcus.substack.com/p/horse-rides-astronaut-redu...
579.
by yorwba 9mo ago
Seedance 2.0 isn't even listed among ByteDance's models: https://seed.bytedance.com/en/models Pretty sure it's a scam to capitalize on excitement about a model that is expected to be released soon.
580.
by yorwba 9mo ago
If this were a competition, some people would try hard to win it. But the goal here is exploration, not exploitation. Once the answers are revealed, it's unlikely a winner will be identified, but a bunch of mathematicians who tried pro
581.
by yorwba 9mo ago
A quick overpass-turbo search for "brand:wikidata=Q24933714 in Moscow" https://overpass-turbo.eu/s/2kaO (Q24933714 being Wildberries https://www.wikidata.org/wiki/Q24933714 ) reveals tha
582.
by yorwba 9mo ago
You're not the first to have wondered: https://news.ycombinator.com/item?id=20730027
583.
by yorwba 9mo ago
This is not a benchmark. They just want to give people the opportunity to try their hand at solving novel questions with AI and see what happens. If an AI company pulls a solution out of their hat that cannot be replicated with the products
584.
by yorwba 9mo ago
Sometimes it just takes a motivated teacher to do things differently: https://fivetwelvethirteen.substack.com/p/tech-free-january
585.
by yorwba 9mo ago
Don't forget about the examination system grouping students by ability and filtering out the worst performers altogether. Of ≈17 million students who took the zhongkao and graduated junior high school after grade 9 in 2024, ≈10 million
586.
by yorwba 9mo ago
> But the reason that attention is quadratic is that each token gets evaluated with respect to each other token. They haven't changed this at all. Section 2.5 seems like it's deferring this to an appendix. They defer it to the
587.
Tech-Free January(fivetwelvethirteen.substack.com)
1 points by yorwba 9mo ago|0 comments
588.
by yorwba 9mo ago
When Alibaba succeeds at producing a GPT-5.2-equivalent model, they won't be releasing the weights. They'll only offer API access, like for the previous models in the Qwen Max series. Don't forget that they want to make money
589.
by yorwba 9mo ago
SWE-Bench Pro consists of 1865 tasks. https://arxiv.org/abs/2509.16941 Qwen3-Coder-Next solved 44.3% (826 or 827) of these tasks. To solve a single task, it took between ≈50 and ≈280 agent turns, ≈150 on average. In ot
590.
by yorwba 9mo ago
I think for the median HN user, the biggest takeaway from this article would be that well-known Chinese open-weight models like Minimax, Moonshot's Kimi and Zhipu's GLM are ultimately minor players in China when it comes to market
591.
A Three-Way Race for China's AI Super-App(chinai.substack.com)
1 points by yorwba 9mo ago|1 comments
592.
by yorwba 9mo ago
It's a bug with the way this disambiguation page is interpreted: https://en.wikipedia.org/wiki/Kudu_(disambiguation)
593.
by yorwba 9mo ago
The flatworm genus Obama is not named after Barack Obama. https://en.wikipedia.org/wiki/Obama_(flatworm)
594.
by yorwba 9mo ago
Previously: https://news.ycombinator.com/item?id=46618544
595.
by yorwba 9mo ago
"Each oat meal comprised 100 × g of rolled oat flakes (Demeterhof Schwab GmbH & Co. KG, Windsbach, Germany) boiled in water."
596.
by yorwba 9mo ago
The Reuters article focuses on exports because that's what their target audience is likely to be interested in, but the People's Daily article they reference in passing https://paper.people.com.cn/rmrb/pc&#x2F
597.
by yorwba 9mo ago
The model is explicitly trained to produce as uniform a distribution as possible, because it's designed for batched inference with a batch size much larger than the expert count, so that all experts are constantly activated and latency
598.
by yorwba 9mo ago
BYD is selling a lot of cars, but they're also making a lot more cars than they can sell at sticker price in China, as does every other Chinese car company. This oversupply leads to all kinds of distortions, like dealerships registerin
599.
by yorwba 9mo ago
GPUs are more efficient than CPUs for LLM inference, using less energy per token and being cheaper overall. Yes, a single data center GPU draws a lot of power and costs a fortune, but it can also serve a lot more people in the time your C
600.
The Interventions We Need(fivetwelvethirteen.substack.com)
1 points by yorwba 9mo ago|0 comments

More ›