Hacker Newsnew | past | comments | ask | show | jobs | submit | suchire's commentslogin

Reminds me of a quiz tournament that I participated in once, where the answer was “Apollo 11” (eleven), but the host thought the answer read “Apollo Two” and told us we were incorrect.


Is their WER graph just completely made up? It’s comically bad


Given how little the US invests in public education, probably never


>Given how little the US invests in public education, probably never

While I understand the sentiment, this isn't factually true: from figures I've seen, the US actually spends more per student in pre-college schooling than any other nation.

It's a lot like the US healthcare system: the US spends (much) more per-capita, but gets much worse results overall. The problem isn't the level of funding, but how it's used and who runs it.

Another factor is US education is that the funding is largely from local sources and locally controlled, so if you live in a wealthy county, you might have (relatively) very good schools, while kids in poor counties will have poorly-funded schools. This is mostly a problem unique to the US because it loves keeping power at local levels so much.


'Invests' doesn't necessarily have to mean about money, like you assume.


> The problem isn't the level of funding, but how it's used and who runs it

aka rent-seeking behavior from entrenched corporations and interests.


At one point, apparently it was fashionable amongst teens to type characters by using pinyin and always selecting the first character in the list of options, regardless of the intended actual character. That was essentially phonetic writing, but as a result, texts were incomprehensible to parents (the desired outcome).


If the texts were incomprehensible to parents, how were they comprehensible to their intended recipients?


It's the same logic as writing a sentence like this:

Y U gna be late

It's grammatically completely incorrect. But you can still understand it.

When it comes to chinese/japanese characters, many have the same phonetic reading. So you can do something similar, while selecting the wrong characters.


I think it's just easier for beginners (or teenagers) to go from phonetic to meaning. I guess advanced Chinese readers don't even read the words out loud in their head and go directly to meaning. I'm beginner/intermediate at Chinese and surprisingly, I noticed that my pinyin often seems better than many Chinese natives.


It sounds like the Chinese version of 1337speak.


The real answer to your question: the most commonly used Chinese input method allows you to type the first pinyin letters only, and the algorithm will figure out the most likely Chinese characters you want.

It's not "the parents" can't read it. It's that people who don't use electronics have a harder time reading it.


(I misread the top comment)


Read the nonsense text aloud and then listen. Presumably with practice, you don't actually to actually speak aloud, and your 'inner monologue' voice is sufficient.


I once rode a bus from Boston to NYC in the middle of a winter storm at night. The bus got lost and the trip ended up taking close to 7 hours. These buses had TVs for showing video tapes, and so partway through, the driver decided to show “The Ring”. Let me tell you, an involuntary horror movie screening in the middle of a blizzard is quite a memorable experience.


This is a paper describing the pharmacokinetics (the rates of various phases of entering the blood and clearing out of the body) for various formulations of the drug, and their hypotheses for why this might be the case:

https://pubs.acs.org/doi/10.1021/acs.molpharmaceut.3c00626

tl;dr: the molecule is poorly soluble in water, so by suspending a bunch of microparticles and injecting them subcutaneously, the drug very slowly dissolves over time, and it’s very potent, so only a little bit is necessary to do its job.


This workflow is exactly what Descript does. Transcript-based editing, filler word removal, noise reduction, volume normalization, Overdub spoken word correction using the speaker’s voice, eye gaze correction for video, etc.

Disclaimer: I work at Descript


San Francisco has Golden Gate Park, Mt Sutro, Lands End, and the Presidio as forests. I think park maintenance in those places is less to ensure trees thrive and more to keep up trails, remove hazards, etc.


Vermeer made only low dozens of paintings as opposed to, say, Monet, who made thousands, so likely they just mean something like “including one of Vermeer’s paintings, which are rare”


Jenkins was a fork of Hudson. No one talks about Hudson anymore


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: