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

We do have a chicken and egg problem. I think the idea here is that it's a systemic issue and the enforcement is focussed on individuals. This is analogous to the concept of getting everyday people to recycle when the companies creating the products have greater control over how much garbage is produced.

Employers need to stop taking advantage of undocumented workers at artificially suppressed wages. This has acted like a subsidy keeping these poor business models afloat. This has led us to the situation we are in now, where we've become dependent on undocumented migrants (food production etc), who we are being taking advantage of (lower wages, less rights), and also trying to villanize & deport them (the article above). All simultaneously.

It's possible with careful coordination of industry, legislation, and immagration, we wouldn't be here. But now that we are, we need to either find a way to improve the situation or reverse it.


I don't think it's chicken and egg at all. I think lots of employers employ immigrants illegally, and then the immigrants take all the political heat. Anyone pissed about "all these illegals" should be at least just as pissed about all the businesses illegally employing them.

We should stop letting employers do this, and then we all discover that we still really want to employ immigrants, we should enable that, legally.


Thanks for sharing. That was a great story.


And thank you for posting the thank you because it was that that prompted me to bother to follow the link. I too think it was a great story, well told.


I wonder if the, now abandoned, Intel Optane drives could help with this. They had very low latency, high IOPS, and decent throughput. They made RAM modules as well. A ram disk made of them might be faster.


Intel PMem really shines for things you need to be non-volatile (preserved when the power goes out) like fast changing rows in a database. As far as I understand it, "for when you need millions of TPS on a DB that can't fit in RAM" was/is the "killer app" of PMem.

Which suggests it wouldn't be quite the right fit here -- the precomputed constants in the model aren't changing, nor do they need to persist.

Still, interesting question, and I wonder if there's some other existing bit of tech that can be repurposed for this.

I wonder if/when this application (LLMs in general) will slow down and stabilize long enough for anything but general purpose components to make sense. Like, we could totally shove model parameters in some sort of ROM and have hardware offload for a transformer, IF it wasn't the case that 10 years from now we might be on to some other paradigm.


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

Search: