I started self-hosting after noticing that my AWS bill increased from like $300 per month to $600 per month within a couple of years. When looking at my bill, 3/4 of the cost was 'AWS Other'; mostly bandwidth. I couldn't understand why I was paying so much for bandwidth given that all my database instances ran on the same host as the app servers and I didn't have any regular communication between instances.

I suspect it may have been related to the Network File System (NFS)? Like whenever I read a file on the host machine, it goes across the data-center network and charges me? Is this correct?

Anyway, I just decided to take control of those costs. Took me 2 weeks of part-time work to migrate all my stuff to a self-hosted machine. I put everything behind Cloudflare with a load balancer. Was a bit tricky to configure as I'm hosting multiple domains from the same machine. It's a small form factor PC tower with 20 CPU cores; easily runs all my stuff though. In 2 months, I already recouped the full cost of the machine through savings in my AWS bill. Now I pay like $10 a month to Cloudflare and even that's basically an optional cost. I strongly recommend.

Anyway it's impressive how AWS costs had been creeping slowly and imperceptibly over time. With my own machine, I now have way more compute than I need. I did a calculation and figured out that to get the same CPU capacity (no throttling, no bandwidth limitations) on AWS, I would have to pay like $1400 per month... But amortized over 4 years my machine's cost is like $20 per month plus $5 per month to get a static IP address. I didn't need to change my internet plan other than that. So AWS EC2 represented a 56x cost factor. It's mind-boggling.

I think it's one of these costs that I kind of brushed under the carpet as "It's an investment." But eventually, this cost became a topic of conversation with my wife and she started making jokes about our contribution to Jeff Bezos' wife's diamond ring. Then it came to our attention that his megayacht is so large that it comes with a second yacht beside it. Then I understood where he got it all from. Though to be fair to him, he is a truly great businessman; he didn't get it from institutional money or complex hidden political scheme; he got it fair and square through a very clever business plan.

Over 5 years or so that I've been using AWS, the costs had been flat. Meanwhile the costs of the underlying hardware had dropped to like 1/56th... and I didn't even notice. Is anything more profitable than apathy and neglect?

The most likely culprit was talking to other nodes via their public IP instead of their local ones. That gets billed as interent traffic (most expensive). The second culprit is your database or other nodes are in different AZs and you get a x-zone bandwidth charge.

Bandwidth inside the same zone is free.

That floor plan is typical Albert Heijn. I remember from my time living in The Hague. Grocery stores in most other countries are usually a lot less complex.

The big system integrators are often pretty terrible at their jobs but it isn't the only cause.

Extremely expensive software projects in government have a common thread in every case I have first-hand experience with. The government has no consistent vision of what they want or who is the final arbiter of these decisions, and no person in the government is accountable for the outcomes. Both the requirements and responsibility are spread across so many people that for all practical purposes there are no clear requirements and no accountability.

The government software programs that run well in my experience have the organizational equivalent of a BDFL. A BDFL doesn't really exist in government; even when someone acts in that role they are often reassigned to other projects at random.

> IMO, this is fraud, of a criminal nature. People should go to jail.

I have been in companies getting money from government programs. It's not fraud from the government side, at least not for what I've seen.

The problem is that companies see government programs as a way to make easy money. If the government pays a company for X, that's because that company has expertise in X. So it's easy for the company to bullshit the government employees and sell them crap.

Companies abuse the government funding as much as they possible and legally can, and then we blame the government. And on top of that we would want fewer regulations? If you want to be able to punish abuse from companies, you need regulations, and you need to apply them.

>Companies abuse the government funding as much as they possible and legally can, and then we blame the government.

So, in well-known conditions of ineffective spending without competition government chose to waste money bypassing market? How that's not government's fault?

Not sure if you have a specific example in mind or speaking generally.

What I am saying is that generally, private companies abuse the government money when they can. Just like private companies enshittify their products to make more money. It's all about making money.

Why are governments getting bad software from companies? For the same reason users are getting bad software as well. The software industry produces money, not good software.

My point is that "private companies abuse the government money when they can" is well-known fact. Thus, government as a customer should make everything possible to buy software on competitive market basis. If govt chose easy way to blindly gave shitton of money to some company with vague acceptance criteria and got bad result in return - this is 100% govt fault.

My point is that they are genuinely trying. I have been in contact with such people, and they are not idiots.

They have government money that they want to spend wisely. And experts from private companies convince them that they can solve their problems.

If the government employee was the expert, they would not have to contact the private companies in the first place. The private companies know that, and they abuse their dominant position by convincing (sometimes downright lying) to the government employees.

It is a very difficult position to be in. It's not about buying a car and being able to just test it. Many times the government funding goes for some kind of R&D. Which makes it easy for the experts to bullshit them and never produce anything useful.

Those who say that it's 100% the government fault should maybe try to go work there. They could truly help their country, if they could actually do better. But chances are that they can't.

I am sorry for good people genuinely trying. I understand that it is a hard problem to solve - but private companies chose their contractors and suppliers and as well facing similar issues. Somehow private companies are managing to evaluate results and found accountable persons while governments are known for being very undemanding customers.

"We have tried hard, but failed" is still failure despite any good will. Good will with courage and having skills with competence are very different things and often govt employees have neither because otherwise they will be more successful using this skillset in private companies.

> but private companies

Two things:

* The government is often in the situation where they want to fund some effort, like R&D. Say China is way ahead of the US in drone technology, and the US wants to catch up because that's a risk (we see how drones are used in the military now). So someone will have money to spend on US companies who will do everything they can to get as much money as possible. How can the government employee know which company is abusing? First they are all abusing, and second they are all failing already (otherwise the government wouldn't be in this situation where they need to fund them). This is a very hard problem.

* When it is about buying e.g. an IT solution for the company, I have seen private companies fail just as much as governments. McKinsey and their friends do the same bullshit to everybody, be it government or private companies. Don't think that private companies don't waste money. It's just that when it is the government, it is transparent and we like to complain about the government.

> The problem is bureaucracy; too many regulations and massive hidden corruption

I keep hearing the "too many regulations" argument, and I totally disagree. Too few regulations (or rather no enforcement of them) resulted in the TooBigTech monopolies we have today. Of course, they got so successful because of the lack of regulations, but now it's simply impossible to compete with them. Removing regulations (assuming that there are regulations that impact them today, which I doubt) would help them, not the competition.

And we have precedents:

* Whenever the EU tries to do some antitrust, it impacts TooBigTech (which is almost exclusively US), and as a result the US bullies the EU to stop it. If regulations were weakening the EU, why would the US government fight them?

* Let's continue with the US as the example of fewer bureaucracy in this case (the complaint is that the EU cannot compete with the US because of the EU's bureaucracy): look at examples where a non-US company takes over a market (or threatens to take it over). Huawei smartphones (not the infrastructure like antennas, this is different), TikTok, DJI. What do US companies do to win against them? They lobby like crazy to add regulations that will stop the competition.

The US hasn't managed to compete with TikTok: they made it illegal instead.

When Huawei was becoming very big in smartphones in the US, they got banned.

The US hasn't managed to compete with DJI, and the biggest US drone companies are spending a ton of resources trying to get DJI banned. DJI is so superior that even banning them is tricky: it has to be done slowly because banning them right away would disrupt entire industries for lack of viable alternatives. That's how far US drones are from DJI drones.

"Too many regulations" is wrong. The successful players get protection from their government (be it the US or China), and it's high time the EU protected its own players, too. With regulations, just like the US and China does (when they don't abuse their dominant position to bully the EU).

That seems like a lot of words to basically just admit that Europe’s regulations are anti-competitive.

I especially enjoyed reading the logical fallacy of drone companies that are so small/non-existent that DJI cannot be banned quickly, but those same companies mysteriously have enough money to bribe politicians for a ban (and the much bigger DJI can’t outbid them).

Also: You wrote in a previous comment that nobody can compete with Apple due to lack of antitrust regulation:

> > Is there anything that stops today anyone from starting a new Google or a new Microsoft or a new Apple in Europe?

> The same thing that stops anyone from starting a new Google or Microsoft or Apple in the US, probably.

But now you are saying that Huawei was about to unseat them and therefore needed to be banned. So… which is it?

Wow, that's a lot of bad faith. I'll try once, and once only:

There has been a ton of money thrown at US drone companies in the last 10 years. A TON. From the government and from VCs. It's not that those companies are so small or non-existent: it's just that consumers do not buy their products. Which is why most drone companies have conveniently pivoted to the military now. And with the military funding, they have money.

And they have been lobbying A LOT to ban DJI, and they are winning that fight. But that does not mean that the consumers want to buy their US drones. DJI drones are still vastly cheaper and better, and professional users (including US government entities) rely on DJI. So much that it is unreasonable to just ban all existing DJI hardware. It has to come progressively so that those consumers can get used to paying a lot more to get worse drones.

There is no question here: without regulations, NOBODY can REMOTELY compete against DJI, period. If you don't understand that, then you don't understand the drone industry.

> But now you are saying that Huawei was about to unseat them and therefore needed to be banned. So… which is it?

Huawei and TikTok and DJI are Chinese companies. US people never forget to mention that they don't exactly play by the same rules (e.g. they can grow in their protected market until they reach a size where they can try in the rest of the world). US companies cannot do that (try to build a US smartphone manufacturer and compete with Apple in the US, just for fun).

But in those very rare situations where US companies get competition (and that will happen more and more from China), and those US companies suddenly find themselves in the weaker position, the VERY FIRST thing they want is more regulations.

"When I win, that's because I am the best. When I lose, that's because the rules are against me". You think Europe sucks because their tech companies lose against the US? Have fun comparing the US to China then :-).

I was kind of retired, earning passive crypto income for several years after 2019 throughout COVID. Best time of my life. I was living on a Mediterranean island, splitting my time between snorkeling and open source work.

Then I got cheated out of my position in the crypto project. Literally scammed by the project founder with the full support of government regulators who are supposed to be preventing this stuff and lost all my income overnight. The regulators literally facilitated fraud instead of preventing it... And I had the pleasure of being gaslit about it while also being gaslit about COVID by a different set of regulators. I became a conspiracy theorist during this time! Now I'm forced to work again...

It's especially infuriating in this age of perma-bailouts where the system is basically bailing out everyone with assets.

I figured out that the system is a scam. I can prove it to anyone in excruciating detail, with citations. If anyone should be bailed out, it should be me. I shouldn't be forced back in the hamster wheel. It's hard to compete against others who think the system works a certain way and don't realize how the hamster wheel works. I shouldn't have to compete with delusional fools who think that their effort spent on the hamster wheel is going to yield any rewards.

Anyway it drives me nuts how the only people who can afford to retire, choose not to... And those who are desperate to retire, can't! This is so pervasive, it feels like a psyop.

Yes LLMs aren't very good at architecture. I suspect because the average project online has pretty bad architecture. The training set is poisoned.

It's kind of bittersweet for me because I was dreaming of becoming a software architect when I graduated university and the role started disappearing so I never actually became one!

But the upside of this is that now LLMs suck at software architecture... Maybe companies will bring back the software architect role?

The training set has been totally poisoned from the architecture PoV. I don't think LLMs (as they are) will be able to learn software architecture now because the more time passes, the more poorly architected slop gets added online and finds its way into the training set.

Good software architecture tends to be additive, as opposed to subtractive. You start with a clean slate then build up from there.

It's almost impossible to start with a complete mess of spaghetti code and end up with a clean architecture... Spaghetti code abstractions tend to mislead you and lead you astray... It's like; understanding spaghetti code tends to soil your understanding of the problem domain. You start to think of everything in terms of terrible leaky abstraction and can't think of the problem clearly.

It's hard even for humans to look at a problem through fresh eyes; it's likely even harder for LLMs to do it. For example, if you use a word in a prompt, the LLM tends to try to incorporate that word into the solution... So if the AI sees a bunch of leaky abstractions in the code; it will tend to try to work with them as opposed to removing them and finding better abstractions. I see this all the time with hacks; if the code is full of hacks, then an LLM tends to produce hacks all the time and it's almost impossible to make it address root causes... Also hacks tend to beget more hacks.

Refactoring is a very mechanistic way of turning bad code into good. I don’t see a world in which our tools (LLMs or otherwise) don’t learn this.

> I don’t see a world in which our tools (LLMs or otherwise) don’t learn this.

I agree, but maybe for different reasons. Refactoring well is a form of intelligence, and I don't see any upper limit to machine intelligence other than the laws of physics.

> Refactoring is a very mechanistic way of turning bad code into good.

There are some refactoring rules of thumb that can seem mechanistic (by which I mean deterministic based on pretty simple rules), but not all. Neither is refactoring guaranteed to be sufficient to lead to all reasonable definitions of "good software". Sometimes the bar requires breaking compatibility with the previous API / UX. This is why I agree with the sibling comment which draws a distinction between refactoring (changing internal details without changing the outward behavior, typically at a local/granular scale) and reworking (fixing structural problems that go beyond local/incremental improvements).

Claude phrased it this way – "Refactoring operates within a fixed contract. Reworking may change the contract." – which I find to be nice and succinct.

Refactorings can be useful in certain cases if the core architecture of the system is sound, but for some very complex systems, the problems can run deeper and a refactoring can be a waste of time. Sometimes you're better off reworking the whole thing because the problem might be in the foundation itself; something about the architecture forces developer's hand in terms of thinking about the problem incorrectly and writing bad code on top.

I think overall, the idea of money is messed up on many levels. What we call 'money' today doesn't even have an identity. It's the most important thing in the world, it's also the most heavily utilized thing in the world but almost nobody knows what it means.

- It's backed by nothing.

- It's not a fair medium of exchange because it physically cannot circulate very far from 'money printers' (not many hops) before it's taxed down to nothing. This means that it's unevenly scarce based on social proximity; unfair by design. Cantillon effects on steroids.

- It doesn't even exist as a single cohesive concept; the US dollar in your bank account is not the same as the US dollar in your friend's bank account and it's not the same as the US dollar which European traders use to buy derivatives (e.g. Eurodollars)... There are literally thousands of different ledgers (banks, institutions, in different countries), each presenting its own interface supposedly showing their holdings of this mythical unit called 'The US dollar' which is actually thousands of different currencies, which happen to share the same name, scattered around the world and held together only by regulators whose only shared interest is to print more units for themselves than the next guy does. Slow and fallible human regulators represent the only layer of 'consensus' which exists for the entire fiat monetary system; they move at snails' pace in a world of high frequency trading.

> - It's backed by nothing.

Money is never backed by nothing, or it's worthless. It may not be backed by anything physical, but it's always backed by some form of trust. National currencies are backed by trust in the corresponding government and institutions.

But that trust is often backed by nothing. Especially if you don't own assets; then from that perspective money is really working against you and is backed by pure coercion... But coercion is not an asset and it doesn't have net positive value; at least not to the victim.

It has value from the perspective of the oppressor I guess... I think this is where it derives its value.

I keep having to repeat this.

The problem is systemic. The legal concepts of 'Corporate personhood' and 'Limited liability' don't make sense. Just think about what those terms mean.

Of course you cannot expect accountability in the long run.

Is a corporation really a legal person? Can it go to jail as a person might? Does it need a visa to operate in a country as a person might? Will a corporation die as a person will? Sounds like it's getting all the benefits of personhood and none of the drawbacks... Our legal system literally gives more rights to non-sentient entities than it does to us humans! No wonder things are getting out of hand!

What does 'Limited liability' mean? Who has to deal with the repercussions for the excess liability which may exceed beyond the limit?

It's deep corruption, codified into law. Of course these corporations will get worse. They will get satanically worse. Just watch what happens internally; a decade from now, the current CEOs will look morally responsible by comparison. It's a systemic issue. Total violation of the social contract at a deep human level.

Why don't we say weapons are legal persons and provide limited liability protection to the person wielding it? Then criminals could kill people and the court could pass judgement that the gun must serve 20 years in its holster while the criminal walks free... That's about as fair as what we have now. If you conceive of a corporation as a weapon. There is nothing in the law to explicitly prevent this exact scenario. The corporation could theoretically use up CEOs as a gun might use up bullets... The investors would bear no liability.

There are many people in this depraved world of ours who would be willing to be a corporate bullet. People will go to jail to provide for their family. With corporations sucking the wealth out of society, it will create new levels of desperation, this will surely happen.

Limiting human liability through coercion has always been the problem. I signed no contract with any LLC founder to limit their personally liability. Yet somehow some bureaucrat hundreds of miles away signed for me and now none of us can hold people accountable for their actions without armed men preventing it. Just look at what companies like Dupont, Monsanto, Bayer have done to entire towns. Limited liability is the problem. Uncle Sam is the worst uncle.

It's scary because OpenAI could do this if they wanted. They could show people different things. They could learn about people and manipulate them individually.

What's wrong this postgres?

Which alternatives to PostgreSQL would you like to see get more attention?

All of them. Nothing wrong with Postgres, I like Postgres. But the more alternatives the better. My favorite database is RethinkDB but officially, it's a dead project. Unofficially it's still pretty great.

Outsourcing manufacturing was very short-sighted in light of the automation which was taking place and accelerating.

No doubt this short-sightedness was the result of our debt-based monetary system. The disconnection of money from long term value-creation created a cycle of speculative booms and busts which made short term bets the most viable strategy to ensure that execs would get their bonuses.

Also, the perverse legal concepts of 'corporate personhood' and 'limited liability' sealed our fate, ensuring that companies could pollute our land and water with chemicals... China was all too happy to send children's toys full of phthalates and other endocrine disruptors our way, ensuring that the next generation would be pacified and struggling with hormone-related issues (I leave you to infer cultural implications...)

Seems like China got their revenge for the Opium wars!