Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

What Google Translate can tell us about vibecoding

ingrids.space

Post reply on HN
Print view View original

71–80 of 175 posts

Re: What Google Translate can tell us about vibecoding

#71

Post by djmips »

kasey_junk wrote: post #51
ohthatsnotright wrote: post #42

Earlier quoted context omitted.

What a strange thing to ask for a citation on when CEO pay, stock buy backs and corporate dividends are at all time highs while worker pay and honestly just affording to live continue to crater.

Productivity is up and labor wages are up. That’s why I asked. It wasn’t an attempt at a rebuttal it was a request for reading material as it’s a heterodox view. The normal conversation is that productivity growth has slowed and the divide has increased, not that more productivity creates lower outcomes in real terms. https://www.bls.gov/productivity/images/labor-compensation-l...

What is 'real hourly compensation'?

Re: What Google Translate can tell us about vibecoding

#72

Post by aidenn0 »

felipeerias wrote: post #43
tkgally wrote: post #5

> ... a translators’ and interpreters’ work is mostly about ensuring context, navigating ambiguity, and handling cultural sensitivity. This is what Google Translate cannot currently do. Google Translate can't, but LLMs given enough context can. I've been testing and experimenting with LLMs extensively for translation between Japanese and English for more than two years, and, when properly prompted, they are really go…

It is hard to convey just how important it is to be able to provide additional context, ask follow-up questions, and reason about the text. I live in Japan. Almost every day I find myself asking things like “what does X mean in this specific setting?” or “how do I tell Y to that specific person via this specific medium?”. Much of this can be further automated via custom instructions, so that e.g. the LMM knows that t…

I'm bad with names, so for any Japanese literature, I need to take notes; it's not unusual to see one character referred to by 3 names. Then you might have 3 characters that are all referred to as Tanaka-san at different points in time.

Re: What Google Translate can tell us about vibecoding

#73

Post by ethan_smith »

krackers wrote: post #4

This seems like a terrible comparison since Google Translate is completely beat by DeepL, let alone LLMs. (Google Translate almost surely doesn't use an LLM, or at least not a _large_ one given its speed)

Google Translate has actually been using neural machine translation since 2016 and integrated PaLM 2 (a large language model) in 2023 for over 100 languages, though DeepL does still outperform it in many benchmarks.

Re: What Google Translate can tell us about vibecoding

#74

Post by lodovic »

tartoran wrote: post #9
darvinyraghpath wrote: post #3

Fascinating thought piece. While I agree with the thrust of the piece: 'that llms can't really replace engineers', unfortunately the way the industry works is that the excuse of AI, however grounded in reality has been repurposed as a cudgel against actual software industry workers. Sure eventually everyone might figure out that AI can't really write code by itself - and software quality will degrade.. But unfortunat…

Maybe AI can't replace engineers but it surely can apply downward pressure on engineers' salaries.

I don't believe that. Software has become so expensive in the last decade, that only very large enterprises and venture capitalists were still building custom applications (ymmv). LLMs make it cheaper and faster to create software - you don't need these large teams and managers anymore, just a few developers. Smaller companies will be back in the game.

Re: What Google Translate can tell us about vibecoding

#75

Post by 3eb7988a1663 »

sudahtigabulan wrote: post #47
noobermin wrote: post #46

Earlier quoted context omitted.

Americans are extremely polite, and American English is replete with niceties in everyday speech. It's funny, having been a tourist abroad, it's only living in a significantly non American country for a few years that lead me to realise this. I even lived in another country albeit a former US colony for ten years and didn't even notice it there given American influence. I stopped saying, stuff "I would like a latte t…

> could I get a latte today actually To me (non-American) the above sounds like sarcasm, not politeness. Adding "today" and/or "actually" could mean you've had it with their delays. I like to joke that Americans always seem to find ways to get offended by innocuous things, but in this case the joke is on me.

American - I fully agree with your interpretation. Throwing on the time component gives up all pretense of being polite.

Re: What Google Translate can tell us about vibecoding

#76

Post by numpad0 »

eurleif wrote: post #61
jezzamon wrote: post #52

I was thinking about this comparison recently along a slightly different axis: One challenge when working with translations (human or machine translations) as that you can't vet whether it's correct or not yourself. So you just have to either trust the translation and hope it's the best. It's a lot easier to trust a person than a machine, though I have had someone message me once to say "this translation of your arti…

You can kinda validate machine translation by doing a circular translation, A->B->A. It's not a perfect test, but it's a reasonably strong signal.

Round tripping is definitely better than nothing, but it lets through a lot of "engrish" errors with multi-meaning words and with implied perspectives. Cape as in clothing and cape as in pointy seasides, general as in military rank vs general as in generic, etc.

e.g. "Set up database as: [read replica]" and "Database setup complete: [a replica was loaded.]" may translate into a same string in some language.

Re: What Google Translate can tell us about vibecoding

#77

Post by bluefirebrand »

sudahtigabulan wrote: post #47
noobermin wrote: post #46

Earlier quoted context omitted.

Americans are extremely polite, and American English is replete with niceties in everyday speech. It's funny, having been a tourist abroad, it's only living in a significantly non American country for a few years that lead me to realise this. I even lived in another country albeit a former US colony for ten years and didn't even notice it there given American influence. I stopped saying, stuff "I would like a latte t…

> could I get a latte today actually To me (non-American) the above sounds like sarcasm, not politeness. Adding "today" and/or "actually" could mean you've had it with their delays. I like to joke that Americans always seem to find ways to get offended by innocuous things, but in this case the joke is on me.

Tone of voice and mannerism matters here a lot

To me (Canadian, not American) "Could I get a latte today actually" sounded something like "Normally I get something other than a latter but actually today I would like a latte instead"

Not rude at all, but kind of assumes some context

Re: What Google Translate can tell us about vibecoding

#79

Post by bluefirebrand »

lodovic wrote: post #74
tartoran wrote: post #9

Earlier quoted context omitted.

Maybe AI can't replace engineers but it surely can apply downward pressure on engineers' salaries.

I don't believe that. Software has become so expensive in the last decade, that only very large enterprises and venture capitalists were still building custom applications (ymmv). LLMs make it cheaper and faster to create software - you don't need these large teams and managers anymore, just a few developers. Smaller companies will be back in the game.

> Smaller companies will be back in the game.

Smaller companies that can afford smaller salaries you mean?

I'm not sure that is at odds with what the post you were replying to said

Re: What Google Translate can tell us about vibecoding

#80

Post by vasco »

jezzamon wrote: post #52

I was thinking about this comparison recently along a slightly different axis: One challenge when working with translations (human or machine translations) as that you can't vet whether it's correct or not yourself. So you just have to either trust the translation and hope it's the best. It's a lot easier to trust a person than a machine, though I have had someone message me once to say "this translation of your arti…

> So you just have to either trust the translation and hope it's the best. It's a lot easier to trust a person than a machine

Only because you think the machine will be incorrect. If I have a large multiplication to do it's much easier to trust the calculator than a mathematician, and that's all due to my perception of the calculator accuracy.

Post reply on HN