Hacker News new | past | comments | ask | show | jobs | submit | LPisGood's comments login

Yes, a big part of the size is because the military is a massive and horrendously inefficient jobs, education, housing, and healthcare program.

Don't forget all the beak-wetting that happens along the way when signing contracts etc. That's where an actual difference could be made.

I think people VASTLY overestimate the amount of graft in military procurement.

Lockheed only has a $100b market cap. Raytheon has $200b. General Dynamics $74b

The reality is that US defense spending pays American designers and American laborers high prices for their American effort. We pay basically the same prices for ammo and supplies and services as other countries.

When we pay $13 billion for an aircraft carrier, that's just what it costs to build a gigantic boat with nuclear reactors. The French paid $4 billion for their aircraft carrier, and a $12 billion Gerald R. Ford class is over twice as large as the Charles de Gaulle (40k tons vs 100k tons), and much much much more advanced.

Americans love to misunderstand the cost of military things. They will scream about the F35's $1.5 trillion "price tag", ignoring that the estimate is for 50 years of operations and maintenance as well as initial purchase. Actual purchase price is about $90 million a plane, which is reasonable. Which makes sense, since being not stupidly overpriced was a key point of the program. The operational cost is about $40k a flight hour, which is roughly the same as the F-14, another high tech superplane program.


40k per flight hour is actually extremely reasonable for an advanced aircraft - good luck trying to charter a large cabin private jet for cheaper.

Do you know about the president banning the federal government from working with people represented by specific law firms he doesn’t like? Are you aware he has been revoking clearances of all lawyers working at law firms that have brought suit against him and/or his government?

These blatantly corrupt abuses of power against officers of the court are not propaganda.


In the last weeks I have used it for things like very specific drug jokes and a baseball bat.

I just got a new MacBook and I felt reasonably inundated with requests to opt in to things.

American conservatives are increasingly not grounded in facts and reality. This isn’t partisan, it’s just an observation of reality. I used to identify as a conservative, but they have become less and less grounded as a party.

I feel like this is a very dramatic view of things. Have you ever been in a management position?

I have and this poster is spot on, he needs to go higher up the chain though. Investors hate employees, even founders. Founders are out to get rich. Executives are out to get rich but don't have what it takes to be founders. All of these people detest labor. They are the enemy you must work with to buy food. Treat them as such.

I feel like this is a very cynical way of looking at things.

I know some excellent people in leadership that have been promoted from lower level management jobs. I’m not sure the career change made them no longer care about people.


This is definitely how the capital class views labor. Don’t be fooled, and ignore at your own peril.

I wouldn’t really consider someone that moved from middle management to upper management/executive leadership to be a part of the “capital class.”

Investors, board members, maybe even some CEOs, sure.


If they’re upper management, they are tasked with doing the dirty work for capital. That’s their job.

That doesn’t mean they don’t care though. Now this could come from spending time at a company where the executives were only ~3 levels of indirection from devs.

They would lay you off in the blink of an eye or be instantly fired. “Caring” doesn’t really come into play.

It's not that the career change makes them not care about people. It's that it's practically impossible to get into upper management without eating others. People who don't embrace their sociopathic tendencies don't make it - they get out-competed by those who do. The very occasional exception just proves the rule, and usually doesn't last in any case because once they get to that point, they still have to compete to remain there.

I think his statement could have benefited from and/or implicitly in the list of titles.

I certainly care very deeply about my people, and letting someone go is a last resort after trying to work things out. My boss cares that I care.. their boss.. we're numbers.


Is that company in data storage?

No. It's a small company ~400 employees in total, including contractors. The CEO is a very technical guy. The average tenure of the employees let go was ~8 years.

Well to be fair 10+ years ago it’s hard to find too many dollar for dollar investments that beat MAAMA.

True! But it's crazy that the risk, successful startup didn't even beat established tech for returns.

It isn’t crazy, it’s that startups are just that bad of an investment when averaged out. You have better odds playing the lottery.

I was wondering that myself.

It led me down the path of imagining two data centers: one in your basement, and one in some office building with a reception area, some cubicles, etc. It seems like the former would see the government breaking down doors and yanking servers off the rack, ripping ethernet cables. The latter would probably see a phone call or perhaps a discussion with reception, followed by a security officer or something.

The suspicious workload could be the exact same in both cases, but this is one of those neat little spots where being a Real Business has massive advantages.


Selling compute that is on my home network using storage that resides at my house and networking that leads back to where I live to random internet people sounds like a nightmare that cannot possibly be worth the few hundred bucks a month it may earn you.

To wit, no serious business would use your services, so your market is pretty much limited to:

1.) Amateur hobbyists (very little money there)

2.) People with Bad Intentions (horrendous from a legal and security perspective, and ethics if that’s your thing)

3.) People in your social circle


Not even Amateur hobbyists; they will tend to break down into:

1. enthusiastic enough to have their own homelab and host their own kit.

2. have money and willing to spend on 'proper' hosting - either AWS/Azure/Hetzner/OVH style or cheap VPS style. You say "few hundred bucks a month" but a low quality virtual server starts from from $1/month - https://lowendbox.com/blog/1-vps-1-usd-vps-per-month/ - and will be in a datacenter. You'd need a lot of customers or a very competitive offering to make even one hundred bucks profit every month.

3. have no money or don't want to spend money, using Oracle/Amazon free tier, SDF free shell, a free account from a friend with a homelab, etc. They don't make good customers.

That leaves people who have money and are willing to spend it on a low quality product instead of a higher quality product, which is basically your third option - friends who are giving you money and putting up with a worse deal out of friendship - charity, basically.


I have some friends with Nextcloud accounts that run on my home server, and a couple have accounts on my Matrix home server.

They understand that if anything goes wrong I will offer a complete refund (which is $0), because they’re my friends. I couldn’t imagine doing it for other people though.

It also helps that they’re IT or at least interested in IT.


My friends have offered to pay for some of the hard drives, but I refuse because once I accept money, there's an implicit owing them to keep it up (whether that's real or in my head). Now it's just a best effort, and I prefer that over a couple hundred bucks.

Just getting people to pay their domain fees on time (or at all) was a nightmare back then - so nowadays there's only one deal on infrastructure I own:

It's free and comes with no guarantees. If you need a domain or subdomain, you own it, pay for it, and just put a DNS record to the IP I tell you.


It's less about coming back to a "home", and more about self-hosting.

Self-hosting a service, website, or SaaS happens alot more than most people realize especially with fibre to the home, and even things like Starlink.

Sharing your home services doesn't seem mission critical, unless it is. It's easy to just pop in a QNAP or Synology for any individual for themselves.

Businesses who have data residency and processing requirements are live and well, and pay a premium.

The pendulum of cloud vs local is starting to swing the more AI models become available in datacentres to keep close to your data in 2025. Google just announced privately hosting to Gemini a few days ago iirc.

For self-hosting, one has to stabilize power, internet, and decide on equipment. This is orders of magnitude more doable, easier and cheaper every 5 years going back 15.

Just have to be clear on how experienced you are with this in the past and most recently. The cloud is ridiculously profitable and overpriced because in part of how much this has come down market just not used.

Of course there's knowledge involved, and again, there's plenty of people who quietly have this knowledge, and it's far easier to obtain now starting with something as simple as Proxmox and/or Docker.


I know of a few "serious" businesses doing this to scrape sites (mostly pricing data) via consumer IP blocks.

You can get residential proxies as an alternative, but it's pretty expensive.

It’s not really that expensive, I guess it depends on your use case but it’s only like $4/GB (edit: misleading, see replies): https://oxylabs.io/products/residential-proxy-pool or https://brightdata.com/proxy-types/residential-proxies

Usually you only need some subset of the data per page load if you invest some time looking at dev tools you can probably find the API call you need and save yourself a few MB.

They all offer scraping APIs now too that can be cheaper for certain use cases where you only need a subset of the data that is actually loaded. Like $1.3 per 1k requests: https://oxylabs.io/products/scraper-api/web/pricing or $1 per 1k: https://brightdata.com/pricing/web-scraper


Providers like Oxylabs can be quite restrictive, preventing access to many of the common sites that scrapers choose to target.

https://faq.oxylabs.info/en/articles/8826164-restricted-targ...

Additionally I believe that $4/GB is an introductory price. When I went into the Oxylab dashboard, it showed me $8.


>Providers like Oxylabs can be quite restrictive, preventing access to many of the common sites that scrapers choose to target.

Most of them seem pretty reasonable?

"Entertainment & streaming" - who's trying to scrape netflix's library?

"Banking and other financial institutions" / "Government websites" / "Mailing" - seems far more likely it'll be used for credential stuffing than for "scraping".

"Ticketing" - seems far more likely that it'll get used by scalpers than for scraping

The main targets of scraping - e-commerce sites (for price comparisons) and social media networks (for user generated content) are fine to scrape. Is there some use case I'm missing here? Is there a huge contingent of people wanting to scrape ticketmaster or bank of america?


I used the term "scrapers" pretty loosely, but yes, in many cases they are more bad actors than actual scrapers. However as they say the list may include other sites, I suspect Oxylab adds sites to the list at the site owners' requests (Amazon, Target, etc are likely to be on those lists)

Hmm that’s unfortunate. I’m actually scaling up a data journalism project this month which is why I’ve been looking at these.

I’m curious if you can suggest a happy medium between curl-impersonate on VPS (dirt cheap) vs residential proxy ($8/gb)?

Personally I’m not trying to hit any of those common sites.


US court says Brightdata's web scraping service can be used to scrape Facebook:

https://www.courthousenews.com/wp-content/uploads/2024/01/me...


Damnit I didn't even knew this existed with such insane pricing.

"Residential proxies is based on traffic and purchase model. Pay as you go model starts at $7.35 per GB, and can be discounted as low as $1.84 per GB when purchased in bulk."


Yeah but if you’re just scraping it’s only a kb or 2 per request. Once or twice a day to check the price of an item would let you track thousands of items for years for just 8$

Depends on the site. Some sites are sending down several megabytes of Javascript or images per request. Some sites even send down massive JSON payloads to page through instead of doing it iteratively.

Exactly. Or if you decide just to scrape whole thing with headless browsers. It would be ridicilously expensive.

But I would guess this type of proxies is mostly use to send data rather than receive. You can access geo-locked sites through standard vpn.


I don't even see the market for 1, because amateur hobbyists will either (1) host their own server, because that's kinda the point of the hobby, or (2) go for a large commercial host like Hetzner, Digital Ocean, Linode or any of the other dozens of business-grade options that are available for <$5-10/month.

Renting parts of a homelab sounds just as strange as renting out the extra space in your family’s refrigerator, at least at first. But thinking about it more.. internet speed/cost in much of America is insane compared to Europe, and depending on what you’re trying to do with the bandwidth it’s not like you can actually address that with cloud, where the free tier is almost useless but above the free tier it’s hard to do much without quickly getting into significant expense.

There’s collocation for mid tier usage patterns that’s cost effective maybe but I imagine it’s on the decline in general these days, and it never seemed that cost effective for hobby stuff unless you had a group of people who were splitting it.


> internet speed/cost in much of America is insane compared to Europe

Can you expand on this? I'm in the US and have 1 Gbps symmetrical fiber for $70/month. The US is a big country and some people will have it very slow, but usually not the same type of people who self-host.


I have 8/2.5 Gbps for slightly less than half of this price.

This is a typical offer in fiber now in France.

Not that it matters because my equipment is 1 Gbps but at least I can flex :)


> a few hundred bucks a month

This is more than I expected a hobbyist to be able to pull off. How/where would you market to achieve this cash flow? If I could make $$$/mo with 1 computer at home, could I scale to $$$$/mo by adding compute/storage?


You can easily pull that with a single server if you put some GPUs in there. See vast.ai

reading reddit about that sounds like that's not feasible

We make a service like that: https://borg.games/setup

The target is $100/m per RTX 3090.


How does that compare to the typical cost of electricity to power a 3090 for a month? I honestly have no idea, this isn't a gotcha question.

At $0.14/kWh and 250w power limit (that I'd recommend to avoid noise) it would be ~$25/m

Hosting a few higher CPU requirement video game servers would put you there. Something like 2 dedicated cores, 8-16GB RAM, and 100-200GB Disk for $30/mo.

Join us for AI Startup School this June 16-17 in San Francisco!

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

Search: