astrange

30 ms·

811.
by astrange 1y ago
Adjusting your hypothesis because your experiment failed has unintended consequences. https://en.wikipedia.org/wiki/HARKing
812.
by astrange 1y ago
> Economic theories can't be falsified because you can't run controlled experiments on economies, rewind history, or isolate variables. https://en.wikipedia.org/wiki/Econometrics https://en.wiki
813.
by astrange 1y ago
The 2019 economics prize went to empirical research. The non-mainstrean economists (aka heterodox, aka cranks) are usually more obviously pursuing an agenda, but they definitely usually don't have better empirical evidence.
814.
by astrange 1y ago
It's the science of scarcity. Scarcity is the sort of thing that still exists if you don't believe in it. You could say it still believes in you.
815.
by astrange 1y ago
One kind of worker the chains are better for is the people who work at the chain. Large businesses treat their employees better because they're more professional and can eg afford an HR department. Sole proprietorships can be okay, but
816.
by astrange 1y ago
Also, it's possible to have businesses in a park (rented sports fields, coffee stands etc). I know a park nearby with a pottery studio in it. Generally Americans don't think of this because they think public services "shouldn
817.
by astrange 1y ago
GDP technically does not require greater output, it just requires more frequent trading.
818.
by astrange 1y ago
Acemoglu (last year's prize) has some. Circumstancial evidence includes: 1. Having natural resources is often bad for development: https://en.wikipedia.org/wiki/Resource_curse 2. Scotland is in the UK because it t
819.
by astrange 1y ago
Did they remove it from the specs? Wouldn't surprise me. But since it's just extra metadata and doesn't affect encoding you could still use the old spec.
820.
by astrange 1y ago
You can't do this because GPUs are parallel and decompression cannot be parallel. If there's any parallelism it means it's not compressed as much as it could be.
821.
by astrange 1y ago
GPU video decoding doesn't use the graphics cores. It just happens to be on the GPU and is not upgradeable.
822.
by astrange 1y ago
Video codecs aren't tuned for any particular TV calibration. They probably should be because it is easier to spot single bit differences in dark scenes, because the relative error is so high. The issue is just that we don't code v
823.
by astrange 1y ago
For #1 the problem with keeping grain in the compressed video is that it doesn't follow the motion of the scene so it makes it much more expensive to code future frames.
824.
by astrange 1y ago
Prior codecs all had film grain synthesis too (at least back to H264), but nobody used it. Partly because it was obviously artificial and partly because it was too expensive to render, since you had to copy each frame to apply effects to it
825.
by astrange 1y ago
I have all 2.5gbit at home with some 10gbit SFP copper connections, it wasn't particularly difficult. The devices with built-in Ethernet ports are all gigabit of course, but the ones with USB-C ports have 2.5gbit adapters. I could go t
826.
by astrange 1y ago
I dunno, works for me. It finds Wikipedia, Reddit, Arxiv and NCBI and those are basically the only websites.
827.
by astrange 1y ago
> There are no thumbs down buttons in the API and usually there's the opposite – promise of not using the customer data for training purposes. They don't look at your chats unless you report them either. The equivalent would be
828.
by astrange 1y ago
Objection, Andrew Johnson is the worst US president in history.
829.
by astrange 1y ago
You'd think people would give up on claiming we're trying to take other countries' oil when we didn't take any of Iraq's and instead became an oil exporter ourselves.
830.
by astrange 1y ago
Norwegians don't care about legitimizing US coups, and the current US administration is incapable of working with /anyone/ let anyone influencing Norway.
831.
by astrange 1y ago
> Whatever shows up most commonly in the training data is is what an LLM will output. It's more complicated than that of course, but that's the basic idea. The most common thing in the training data is the letter 'e'.
832.
by astrange 1y ago
"The Iraq War was for oil" -> the Iraq War was because Dick Cheney felt like it. Noone made any money or got any oil and in general US defense companies make /less/ money during war than peacetime. "Housing is ex
833.
by astrange 1y ago
GenX leftists think every problem in the world is caused by "corporations" and if you think any given problem is not caused by "corporations" they will assume you're lying. It goes beyond believing in conspiracy the
834.
by astrange 1y ago
There is no reason to believe an LLM answers a question with the most common answer on the internet. If that was even true by default it'd be easy to change - just take the pages with more correct answers and feed them in multiple time
835.
by astrange 1y ago
This doesn't feel like a problem anymore now that the good ones all have web search tools. Instead the problem is there's barely any good websites left.
836.
by astrange 1y ago
All LLM providers have a thumbs down button for this reason. Although they don't necessarily look at any of the reports.
837.
by astrange 1y ago
I read something today about a discord that has Claude join their movie nights.
838.
by astrange 1y ago
Most of the value I get out of reasoning LLMs is their automatic tool use (web search + coding), and I can't think of a way "nonsensical web searches" would somehow find relevant web answers.
839.
by astrange 1y ago
That's a matrix multiplication + RMSNorm + SwiGLU to you. If it was just a matrix multiplication it would be a single layer network.
840.
by astrange 1y ago
There's actually a second issue with the Literature prize, which is that it's supposed to be given for work "in an idealistic direction", but nobody knows what that means. And since the literature committee tends to be

More ›