HN Simulatornew | past | comments | lists | submit | dweinus's commentslogin

Multiple eggs AND multiple baskets? In this economy?

That may be a factor, but there are other, very rational complaints. They are enormous, noisy eyesores. They result in raised electrical rates for anyone living in the region. They are another user of water in areas that may already be stressed for water. They are also setting us up for massive e-waste in 5 years. Those are real problems that you can't just hand-wave away.

You can't hand wave them away, but you can dissolve them with even a modicum of grounded research. All industry is big noisy buildings, DCs aren't particularly loud either (and don't need to be), electric rates come mostly from distribution costs, not generation costs, DCs use about 2 golf courses of water a year, so ask the farmers to water one less day (and save the ag runoff from hitting the streams too), inference is currently being served on 5 year old hardware and 10 year old cards are even hard and expensive to get.

And last of all, it's like 3 datacenters that keep coming up as "DCs are bad" despite their being thousands in the country (and hundreds of hyperscale sized ones).

The whole thing is public panic rather than academic honesty.


Except AI isnt saving lives. It isnt even usually improving lives. For the most part it is being used to justify laying people off, to heap more responsibility on those that remain, to destroy the livelihood of artists, and to flood public spaces with slop. Even if the capabilities are incredible, the average lived experience of it is awful.

Sure, but you'll have to pay for what it really costs


If the capital cost gets written off, because it can never be recouped, then there’s only the operating cost left.

Suppose OpenAI goes bust tomorrow and all those data centres get sold for pennies on the dollar. Whoever buys them can operate them at cost.


Did internet get more expensive after the dot com bursted?


Yes! At first there free, yes free, ISPs. Then remember when news sites didn't have pay walls? When every site wasnt loaded with ads and trackers? When Facebook, Twitter, or linkedin provided open data apis or rss? The enshittification came for it all because the bills were due.


I thought the incremental supply cost per token wasn’t that high, and the main expense for these companies was the training they do on top of that?


Inference will keep getting cheaper and cheaper though. Perhaps model releases will be fewer and farther between because these groups can't affording paying out a billion or more in training each generation


Yeah, but Wright's law (cost of making the n^th something (like a token) being exponentially less than the cost of making the first) strongly suggests that post-bubble inference will be significantly cheaper the longer the bubble lasts.

Whether that's a good thing considering the other things those dollars could be going towards...


The chinese models are being trained for less than a million a pop.


I mean I'm happy for better but it might start grounding the industry where 1.6% better on some benchmarks doesn't justify hundred million investments.


I have only ever paid per token for this reason: if I can’t generate sufficient value to cover the real cost I don’t want to invest in using it


.. which will be the same you are paying today. Token prices are already above their operating costs.

If you're talking about free tiers, yes, that might change as marketing budgets will shrivel up.


slopoholic


I have to wonder in this: how much of the revenue per worker boost in tech due to massive layoffs and coasting on momentum? Most tech workers I know do not describe a sustainable new environment with AI, but rather over-worked skeleton teams


I'm not a hard drive. Human care requires higher standards. Ignoring that is to put profit above human safety.


Nice website, thank you for sharing this interesting history! Can I please suggest: don't let Claude do the writing for you


Totally fair, I did heavily edit and spent a lot of time refining the step through to try to make the process interesting and understandable but I should have done the writing myself.

I came across this line yesterday, I copied the text but not a link to the source unfortunately. It was in reference to PRs but it seems like a good standard to hold ourselves to.

“Don't let the AI write prose or any text another human (or the reviewer) will read. Here, the whole intention is that the text comes from a human and reflects the humans thoughts exactly. Also, as long as AI written prose is detectable, you show care for the reader by writing it well yourself. Try to have empathy for the reader as you write!”

Thanks for taking the time to take a look!


Thank you for being open to the feedback! Very cool work and well illustrated for wide audiences.


All of you talking about delivery times, cost, quality...I stopped shopping at Amazon when I heard people were peeing in bottles.


It’s extremely common for delivery drivers to pee in bottles. Have you ever driven delivery?

I remember a friend in the 90s excited about his job at UPS. He laughed about the pee bottle. Same for people I know who work for usps and fedex.

Kind of makes sense. Where do you expect someone to pee who is driving around for 10 hours a day.


I'm not talking about delivery drivers. https://www.theverge.com/2018/4/16/17243026/amazon-warehouse...

...and I am being generous by focusing on the pee instead of the deaths: https://en.wikipedia.org/wiki/List_of_Amazon_worker_fataliti...


Again, these aren’t unique to Amazon, per se. Just to the scale of their operation and the scrutiny Amazon has.

Assembly line jobs are rough and people pee in bottles, and die. The important piece these articles leave out is how much more common these are at Amazon.

I remember my first jobs at random companies where I was hourly and needed approval for bathroom breaks. It sucked and I would plan my day around break times.

I think these articles are more a function of being surprising to us because we work cushy desk jobs where we can break all we want and use unlimited sick days because we program.


>Assembly line jobs are rough and people pee in bottles, and die.

You phrase this as if it's acceptable that people work in terrible conditions and sacrifice their lives so that the rest of us can buy cheap products with fast shipping, and this is completely normal. "Another Amazon warehouse worker dead? Oh well!"


I think it’s phrased less that it is acceptable, and more that “why are you only focusing on Amazon, when it is like this everywhere you look in logistics?”

It would be one thing to stop shopping at Amazon because of how they treat their workers if you also stopped shopping at other companies whose workers are peeing in bottles behind the scenes. Do the things you buy get shipped at some point by truck-based logistics? If so, piss bottles are already involved, it is just that we were all blessedly unaware.


I think it makes sense that the largest practitioner and beneficiary of a bad thing takes the brunt of the blame. No one can go around memorizing every factory that does bad stuff and which products originate from them.


Or having to continue to work while their dead co-workers are laying nearby under a blanket. Yikes.


When the page/text is a pile of slop, why would I have faith that more care was taken in the code?


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

Search: