RestartKernel

How much data until we can intelligently route hard and soft human expert mixtures?

As a Dutch person, I'm very doubtful that's the case, but I'm willing to bet a good ESL speaker is more aware of common grammatical errors than some native speakers. For example, the your/you're mixup makes no sense if you've had to explicitly learn about English contractions in the first place.

Is there any benefit to individual states joining the WHO in lieu of the federal government? As silver lining, I mean.

Is there a term for "LLM psychology" like this? If so, it seems closer to a soft science than anything definitive.

> To post, submit a story whose title begins with "Show HN".

I always enjoy how low-tech this website is. Not necessarily great UX, but it's quaint.

I can see how that would work. For a small website, you're probably only looking at a few comments per day at absolute most, and you'll be reviewing them anyway to filter/approve them. (And the implementation lends to easy automation, since you just need to programatically update the Markdown file instead.)

What are the costs looking like to run this? I wonder whether you would be able to use this approach within a mixture-of-experts model trained end-to-end in ensemble. That might take out some guesswork insofar the roles go.

Don't expect ethics from GPTZero. If you upload a large document, they'll give a fake 100% AI rating behind a blur until you pay up to get the actual analysis. This clearly serves to prey on paranoid authors who are worried about being perceived as using AI.

That seems reasonable compared to an actual developer (depending on your region), but I had hoped for these models to make simple tasks like this fast and cheap so those developers can focus on the difficult stuff.

> I think LGBTQ acceptance in Japan is significantly better than any Western country.

Sorry, but I don't see your reasoning support this at all. The relative lack of Abrahamic religion would make an impact for sure, but Japan is more socially conservative by most relevant metrics. How does this one factor overcome that?

This makes me want to write a post about the rabbit hole that is icon optimisation. It drives me insane when websites suffer from layout shift simply because they are not inlining their icons, for one.

> The repetition does indeed seem to because of the poor quality of the first telephone lines.

I'm pretty sure it's also commonly pronounced "moshi mosh~" as a side-effect of this repetition.

Does it (/ did it originally) actually carry such respect from a Japanese perspective? To me, it seems like a pragmatic solution to cope with bad telephone lines more than anything.

Their enterprise offering is more for fresh retrieval than training. For training, you can just download the free database dump — one you would inadvertently end up recreating if you were to use their enterprise APIs in a (pre-)training pipeline.

Why use a non-English source? Doesn't that encourage people to reply to the headline rather than the article?

A bog standard devcontainer works fine too.

Thanks for the reply. Are you therefore not planning to sell within the EU for now?

Just for my understanding: so full notification support would essentially only apply to anyone in the EU willing to import from the US (or directly from the country you're manufacturing these devices in).

Exactly. I've bought custom keyboards off individuals under such limited expectations, but everything makes this watch otherwise look like a proper product release.

Just to spare people reading this a few clicks (from their FAQ):

> Yes, we warrant against manufacturing defects for 30 days after you receive your order. Ship us the defective watch, and after we receive it back, we will ship (no charge) you a replacement.

Full warranty as of now: https://archive.is/HxXvL

Makes sense for a smaller company. Though proudly advertising a feature you may implement thanks to EU regulation, while avoiding compliance yourself, seems in somewhat bad taste.

Not that I am unaware of the immense difference in scale and influence between Apple and Pebble, and the different nature of these regulations in this light.

How big is the overlap between people who know what a CSV is, but who can't do this themselves with Python or their LLM of choice?

This comment made me check your profile to see what country you're from, since I don't know of any people this brazen other than my own. Maar natuurlijk, you're Dutch too.

The Nothing CMF Watch has a true (two?) week-long battery life due to its RTOS, but it's kind of mediocre otherwise. More like a souped up Casio.

Looks great, but a mere 30 day warranty for manufacturing defects is kind of insane. I'm glad they're planning to fully support notifications on iOS in the EU due to the DMA, but that warranty wouldn't even be legal if they sell directly within the EU.

During my Bachelor's, I wrote a small "immutable" algebraic machine learning library based on just NumPy. This made it easy to play around with combining weights by simply summing two networks by whatever operations are supported on normal NumPy arrays.

... turns out, this is only useful in some very specific scenarios, and it's probably not worth the extreme memory overhead.

Speed:

- 3942ms

- 4281ms

Guess it depends on your region. This is from East-Asia.

This is incredibly cool! How do you think these modern language features would have affected Dreamcast development back in the day? (I have no idea of how difficult the console was to develop.)

Thanks for the link, that's a very interesting statement piece. There must be some word though for the artistic illiteracy in those X/Twitter replies.

Very good work for a high schooler, but explaining backpropagation and cross entropy loss is typical of ML paper-mill, though the feature extractor is interesting. These results really call for more domain-specific analysis to reach any significance deserving of headlines.

This seems super well-done, albeit completely impossible for laymen. Great work, I'd love to see someone with the necessary knowledge play this.

The ranking is very interesting if substantiated, but learning Polish in response is somewhat mental.