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

Probably never. The tunnels cost a lot to build but, once built run almost for free - they're powered by gravity and will keep running for close to a century before major maintained is needed.

Yeah that makes sense but if growth dictates another tunnel... And it takes another 60 years, your capital expense starts to look a lot like an operating expense. Not to mention one of the big stated purposes of this tunnel is actually to facilitate maintenance of the other tunnels. There is probably more operation cost hidden here than seems obvious.

The big reason for tunnel 3 isn't new population growth, it's so that the other tunnels can be shut down for maintenance and inspection. NYC's population is more or less stable over the last 90 years.

Close to a century?

There are Roman aqueducts in continuous operation for two millenia.


AFAIK there is one still in operation and entirely for tourist purposes

Today as in Ancient Rome the aqueduct terminates at fountains. It does not connect to the standard modern municipal plumbing. Whether that counts as being operated for "tourist purposes" is a matter of definition.

https://en.wikipedia.org/wiki/Acqua_Vergine


It's a generalization problem. We can train LLMs that 'know' a lot of stuff in the global sense but the tasks that are interesting to people require the LLM to know a lot about you and your world in a very specific sense. The technical problem is that it's all corner cases and that's impossible to scale right now. No amount of context window is going to get you there either.


I respectfully disagree. The scenario posted by the GP is very clearly an "easy" task of basically sentiment analysis. The only thing it requires is a way to query your message history, then it just goes through them 1-by-1. I would expect any of the current SOTA agentic models to handle this with ease.


I'm really, really confused by the typescript overtaking python comment, they're not really tackling the same domain of problems, right?


I thought it was really perceptive. Look at what JS was designed to do cf. what it's actually doing these days.

Turns out it doesn't really matter what domain you were originally trying to tackle. If you've stumbled upon a low friction way of achieving other things, people are going to use your tool for those things, even if it's not the optimal tool for those domains.

I dread the JS/TS future, but it's obviously coming.


The open question is if there is a sufficient number of down detectors where the system becomes turing complete


Because if there us a nursing shortage there is opportunity to 'supplement' existing nurses with AI and thus transfer more wealth to the very richest among us


This doesn't make much sense to me. As the previous commenter mentioned, this shortage has been ongoing for decades, it's certainly not new in the last two years. Additionally, nursing is one of the jobs least replaceable by AI.

There's a nursing shortage because the work is brutal, under appreciated, and under compensated aside from travel nursing gigs, for those who can maintain that sort of lifestyle. Nurses are a cost center, so management is constantly running floors understaffed. It's to the point that they receive bonuses for running the floor as thin as possible, despite the worsening of patient outcomes and nurses' sanity.

Don't get me wrong, there are some good gigs for sure, but there are lots of terrible ones.


Kitten Space Agency is looking like it will be the KSP2 we deserved but didn't get



thank you!


With 90% reporting Mamdami's lead is larger then Sliwa + Cuomo. Mandami won, not Cuomo and Sliwa lost.


Yeah, I looked when they called and it was very close. More stating, it was inevitable with those 3.


Does it work with osmium?


Not yet - it was just introduced and there isn’t even a full spec yet


But it has to be _exactly_ a million to one


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

Search: