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

Remember how she wasn't locked up? That set the precedent.


> Remember how she wasn't locked up? That set the precedent.

Locking her up would have set a far worse precedent, and I think that the other norm-breaking behavior of the current administration does not support the idea that prior punishment of past administration members for insecure data management would have led this administration to more secure data practices.


No better time than the present to establish a better precedent. Though more realistically, nobody needs to be going to jail in either case - but if there are zero consequences for anyone involved well, that's telling for a leader who frequently criticizes his opponents for not firing people when they do poorly.


> I drive Lyft/Uber full-time

Stop driving Lyft/Uber. It's for those who don't have any skills beyond driving a car.

Sit down and determine what are the skills that you currently possess that are the most valuable on the market right now. Then sell those to the highest bidder.

My two cents, that isn't a job.

> I know I can earn 3-4x more doing freelance tech, but I can’t do that from a car.

Yes, you can. That's a limiting belief; it's just holding you back. All you need is a laptop. Need to make a meeting? Go to a local cafe. Nobody will even notice. More importantly, if you're doing remote contract work, perfectly acceptable to ask a percentage of the total upfront. That will go a long way to get you out of a car and into an apartment quickly.

> What would YOU do in my situation to break the cycle?

1. Stop driving Uber, stop renting motels, stop paying any bills, bring my costs down to near zero.

2. Start taking on as much contract work doing "freelance tech" that I can find, with as much paid upfront as possible.

3. Once I have 10k in the bank, rent an apartment and start cleaning up whatever mess is left.

4. Start making quality of life decisions. Maybe you don't want contract, but want a job? Maybe a nicer laptop? Closer to your kids? Live your life.

#1 is the hardest, because you're risking everything on your ability to deliver. You have to believe in yourself.


I only have a desktop. I thought about getting office space from the chamber of commerce for 250 per month... and I can use their fridge to save money on food (Greek yogurt, Jimmy Dean sandwiches, and string cheese is all I need)...


Don’t rent an office space. You dont have the runway to make that work given the current tech market. Focus on getting a solid enough job to get you actual housing. Temp agency, county jobs program, take anything that will pay you a living wage and provide stability.


> I only have a desktop.

Sell it. It's a paperweight for you right now. You can find a really nice open-box/used laptop for $300-400 that will allow you to work from anywhere. You can charge the laptop at a cafe or a library. $5 for coffee and you have a working space to take calls and work.


DC-AC Inverter will let you run the laptop from the car as a last resort..


Usually there exist also dedicated car chargers for the lighter socket.


will cost you energy though, making battery less reliable or wasting fuel. Best to charge where you get energy for free.


Chamber of Commerce is swimming with other hungry sharks. You'd be better off finding a coworking space (NOT a chain one like WeWork). Talk to the owner and explain your situation, ask for a month or two discount while you get your bearings and attend every meetup they have. Meet everyone, tell your story, share your skills. A small community will help take care of you in ways a CoC will not.


Kiln just opened in St George, I have a dedicated desk, but they have non dedicated desks for I think $150 or so, and has more amenities (and free snacks/coffee to save a little on food costs)


Is this a joke? Genuinely the worst advice I have ever read on this site.

> Start taking on as much contract work doing "freelance tech" that I can find

Why do you think he could find any?


He asked what would you do. That's what I would do. And that's what I did. I'm sorry to hear you can't find work, but that's not a problem everyone shares.


So an invoice attached to an email as a PDF is sent digitally ... those unfamiliar with PDF will think text and data extraction is trivial then, but this isn't true. You can have a fully digital, non-image PDF that is vector based and has what looks like text, but doesn't have a single piece of extractable text in it. It's all about how the PDF was generated. Tables can be formatted in a million ways, etc.

Your best bet is to always convert it to an image and OCR it to extract structured data.


This is simply not true. Maybe it’s easier and you do not need 100% precision. But it is actually possible to extract text and layout of digital pdfs. Else it would be impossible to display it. Of course some people still add image fragments to a pdf, but that practice is basically dying. I did not see a single pdf the last year we‘re it was impossible to extract the layout.


Another interesting statistic is that Europe has given Russia more money than it has given Ukraine in 2024. That's a fun one.

https://www.theguardian.com/world/2025/feb/24/eu-spends-more...


Europe has been dependent on Russian oil and gas for some time correct, but has dramatically reduced imports since 2022 - cutting Russian gas by 80% and oil by over 90%. Additionally more and more european and baltic countries are disconnecting from their grid see ("Baltic states unplug from Russia and join EU power grid"

- Source: https://www.bbc.com/news/articles/crlkwyljl8lo


Nitpick: Europe spent more on Russian oil and gas than it gave to Ukraine.

We're comparing spending and giving there.


https://www.amazon.com/dp/B09PJYMK77/zcF9kZXRhaWxfdGhlbWF0aW...

I'm sure there are power supplies for servers that go above 1600 watts too. If you really want to, you can ... but you really shouldn't.


We also have 20 amp wiring, 20 amp breakers, 20 amp sockets, and plugs too. A lot easier than going 240 volt. That will give you 2400 watts max.


Most residential wiring is 15A except for bathrooms.


And kitchens.

Otherwise, running your microwave, toaster, and coffee maker at the same time would likely trip the breaker.

And obviously, the stove/oven is on its own circuit unless it's gas.


My kitchen has at least 2 120VAC circuits, which seems to avoid this.


NEC code requires 2 separate 20 amp circuits in a kitchen. Usually they split the circuits across a countertop or window. It's been code since ... as far as I can tell, pretty much forever. Unless you're living in a house that is either non-compliant or built before NEC was required, then it should have this kitchen arrangement.


Most kitchens don't run on only one circuit for the obvious reasons you've described.


Don’t forget garages, too.


Ha! Find me where I can buy it for $2000.


Yeah, how do people have these in their hands already? Everywhere I look is sold out.


It's sold out to the people who have these in their hands already.


> No, there would just be less 4080s for people to buy.

Sure, and more 4090s at a lower price.


Nvidia wouldn't leave money on the table dropping the price of the 4090. There would just be more supply of 4090s at the same price. A card manufacturer selling them below MSRP would get immediately sanctioned by Nvidia.


Even a monopolist is constrained by supply and demand. If they could sell everything they make as a 4090 without trashing their margins, they would. The fact that 4080, 4070, 4060 lines exist means they can't.


Right. And the fact that a BMW 2-series exists means that BMW can't sell all the 7-series' that they want.

They're cutting down 4090s into 4080s to fulfill a demand for a cheaper chip, while still supplying their premium option. Your fanciful world concept of things being sold for no/minimum margins is just that: fanciful.


It's really slow. Like, unusably slow. For those interested in self-hosting, this is a really good resource: https://github.com/XiongjieDai/GPU-Benchmarks-on-LLM-Inferen...


You know, there's nothing wrong with running a slow LLM.

For some people, they lack the resources to run an LLM on a GPU. For others, they want to try certain models without buying thousands of dollars of equipment just to try things out.

Either way, I see too many people putting the proverbial horse before the cart: they buy a video card, then try to fit LLMs in to the limited VRAM they have, instead of playing around, even if at 1/10th the speed, and figuring out which models they want to run before deciding where they want to invest their money.

One token a second is worlds better than running nothing at all because someone told you that you shouldn't or can't because you don't have a fancy, expensive GPU.


> For some people, they lack the resources to run an LLM on a GPU.

Most people have a usable iGPU, that's going to run most models significantly slower (because less available memory throughput, and/or more of it being wasted on padding, compared to CPU) but a lot cooler than the CPU. NPU's will likely be a similar story.

It would be nice if there was an easy way to only run the initial prompt+context processing (which is generally compute bound) on iGPU+NPU, but move to CPU for the token generation stage.


Change your interface to the LLM to email. Then you're just sending emails and get your answer back in 15 min. For many cases that'd be useful.


> One token a second is worlds better than running nothing at all because someone told you that you shouldn't or can't because you don't have a fancy, expensive GPU.

It's a minuscule pittance, on hardware that costs as much as an AmpereOne.


Great point


It's not "really slow" at all, 1 tok/sec is absolutely par for the course given the overall model size. The 405B model was never actually intended for production use, so the fact that it can even kinda run at speeds that are almost usable is itself noteworthy.


It's a little under 1 token/sec using ollama, but that was with stock llama.cpp — apparently Ampere has their own optimized version that runs a little better on the AmpereOne. I haven't tested it yet with 405b.


This resource looks very bad to me as they don't check batched inference at all. This might make sense now when most people a just running single query at once, but pretty soon almost everything will be running queries in parallel to take advantage of the compute.


How do you run multiple queries from multiple clients simultaneously on the same HW without affecting each other context?


It depends on the framework. Here's a LlamaSharp example: https://github.com/SciSharp/LLamaSharp/blob/master/LLama.Exa...


My question wasn't about how to run multiple queries against the LLM but rather how is it even possible from transformer architecture PoV to have a single LLM hosting multiple and different end clients. I'm probably missing something but can't figure that out yet.


If you have a branchless program, you can execute the same step of the program on multiple different inputs. https://en.wikipedia.org/wiki/SIMD


> this is a really good resource: https://github.com/XiongjieDai/GPU-Benchmarks-on-LLM-Inferen...

Yes, it is.

But it has not been updated for seven months. Do things change so slowly?


I would never chatgpt my code because I don't want to send it to Microsoft. Slow is better than nothing.


Bummer that they have no stats for AMD, Intel, Qualcomm, etc (C|G|N|X)PUs.


> That means about 50% of the US can not afford a place to live.

So 50% of the US are homeless?


They are either overspending or under non-standard living arrangements: roommates, living with relatives, living in cars, etc.


They are a couple of paychecks away from being broke, and a health condition away from being homeless.


Or they inherited a home and can barely afford insurance, taxes, or upkeep on it. So they could afford to live there but couldn’t afford to buy it.


Barely is not can't. You either can or you can't. If you can't, you get kicked out of your home and they auction it off to pay for the taxes.

It happens. It's disgusting when it does. It IS a problem.

But it's not happening to half the country. That's nonsense. And when you say nonsense, it keeps others from taking the problem seriously.


In some form or fashion yes. Theres a reason why there are so many homeless tent camps everywhere. These people are not choosing to be homeless. They just can't afford a place. But remember a lot of them still have jobs. At Walmart, Amazon Warehouses, etc.


> In some form or fashion yes

Can you clarify? Because I'd find it hard to believe that half of Americans are literally homeless. It might make more sense if you're referring to home-owners.


“Half of Americans are facing housing insecurity” is probably closer to the intent but I don’t know if that’s actually true.


Pretty sure half the US isn't living in tents. I feel like I'd notice that, especially living in a city with a larger-than-average homeless population.

I agree that homelessness is a problem, but you appear to be arguing in bad faith. If 50% of the US actually "can't afford housing", we'd have $170M people living in tents, and that is demonstrably not the case.

I would believe a claim that states that a large portion (maybe 50%, maybe more, maybe less) of the country are facing financial insecurity that makes them feel like their housing situation is precarious. But that wasn't the claim put forth upthread.


What we can't see is the amount of people that are living with roommates and in toxic situations they'd love to exit but are unable to due to a lack of housing options and a lack of money.


170 million people living in tents? Wow.


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

Search: