> Sounds like standard terms from lawyers – not very friendly to customers, very friendly to company – but is it particularly bad here?
Compilers don't come with terms that prevent you from building competing compilers. IDEs don't prevent you from writing competing IDEs. If coding agents are supposed to be how we do software engineering from now on, yeah, it's pretty bad.
"Zero heap allocations: Parser results are stack-allocated using OxCaml unboxed records and local lists" - honest question, why?
On almost any platform on which you want to run a HTTP server - including bare metal - it usually doesn't matter if you keep state near the stack pointer or not. What matters is that you use it well, making it play well with CPU caches, etc. Or is there something specifically horrible about OxCaml's heap allocator?
In a conventional GCed language, you need to minimise heap allocations to avoid putting too much pressure on the garbage collector. The OxCaml extensions allows values to be passed 'locally' (that is, on the callstack) as an alternative to heap allocation. When the function returns, the values are automatically deallocated (and the type system guarantees safety).
This means that I can pass in a buffer, parse it, do my business logic, and then return, without ever allocating anything into the global heap. However, if I do need to allocate into it (for example, a complex structure), then it's still available.
It's kind of Rust in reverse: OxCaml has a GC by default, but you can write very high-performance code that effectively never uses a GC. There's also emerging support for data-race-free parallelisation as well.
The webserver I'm putting together also uses io_uring, which provides zero-copy buffers from kernel to userspace. The support for one-shot effect handlers in OCaml allows me to directly resume a blocked fiber straight from the io_uring loop, and then this httpz parser operates directly on that buffer. Shared memory all the way with almost no syscalls!
I think there are several advantages of stack allocation:
* freeing stack allocated memory is O(1) with a small constant factor: simply set the stack pointer to a new location. In a generational garbage collector, like OCaml, minor garbage collection is O(amount of retained memory) with a larger constant factor.
* judiciously stack allocating memory can improve data locality.
* unboxed data takes up less space, again improving locality.
Overall, I think this about improving constant factors---which makes a big difference in practice!
Unboxed records are fine, but stack-allocated lists make me nervous. What happens when someone gives you 8 megs of headers, and you run out of stack?
This code seems to put a 32k limit on it, but it's a manual check and error return. What about code that forgets to manually add that limit, or sets it too high? How do you decide when to bump that limit, since 32k is an artificial constraint?
By default in oxcaml, "stack" / local allocations happen in a separate stack on the heap (which the runtime allocates for you). If you allocate enough to exceed that capacity, it will resize it dynamically for you.
To me the torque is more important than the horsepower: one nice thing about Waymo is that they're quite good at gentle, constant acceleration up San Francisco's steep hills. Human Uber drivers have a 50/50 chance of making me carsick when they try this, out of some combination of inexperience and driving a car without enough low-end torque.
"Corporations" is meaningless in this context. If you want to own a rental home in the US, you should set up an LLC, both for liability reasons and because it makes it easier to deal with taxes and expenses.
In the parts of the country I lived in, I've never seen big corporations own single-family rentals en masse. They usually go for apartment complexes, which are far more profitable if you have the capital to buy / build one. Commercial real estate too.
If you click around your neighborhood, a lot of single-family homes are owned by living trusts and "Bob & Kate" LLCs, but that doesn't mean there's any hedge fund money involved.
Just to reinforce this, the last place I lived in was owned by a corporation...of one man, who lived a town away in a modest house and worked as a paper pusher by day.
There are well over a million homes in the metro Atlanta area. If the top three owners have less than 4% of a market, need if they act in a completely coordinated manner they have approximately zero market power.
How so, could you explain that a bit? I could see them causing a price dip or spike in purchase prices if they bought or sold all at the same time, but that would affect their own prices that they pay or receive, right? What is the market manipulation with 4% of housing stock?
Houses are unique and have irreducible transaction costs which makes the market for them very inefficient and slow relative to a commodity. For one example, if you are in the market for a 3-bedroom house with a garage, the market is already segmented much more narrowly than can be the case in an efficient market like that for a commodity. If you have to move into one as soon as possible for a new job, and you know closing will take a minimum of 3 months, the market for your prospective houses is going to be extremely small without even factoring in other distinctive characteristics like driving distance and schools. 4% of the aggregate market may represent 25-30% of your “market” nonetheless.
Thanks for that! I guess I don't see how there could be market manipulation without also damaging the manipulator, especially in a market that is as transparent as housing, with nearly every sale being at a public price.
Rental manipulation is much much easier, and probably more prevalent. But unfortunately the price-gouging lawsuits from using software to share pricing information have been settled with the landlords paying peanuts.
I'm not sure deliberate manipulation per se, but the market would be warped by a single participant that owned 4% of the aggregate market. This is especially true if that participant didn't adhere to the normal holding periods and purchasing rationale as the remainder of the market participants. Consider that market manipulation concerns are (some of) the reasons significant holders (>= 5%) of even extremely liquid public companies are required to publicly report ownership and ownership changes.
This assumes that the entire market is for sale at a given time, which is not true. If you have 3 kids and two parents who need to drive to work, there may be only a single digit number of viable houses for sale at a given time in your school district.
> e never seen big corporations own single-family rentals en masse.
I just sold a house to a big corporation that owns about 12,000 homes. There's a whole industry for enabling these buys, opendoor, offerpad, etc... It's usually a wash selling your home as is to a wholesale deal vs. prepping your home and selling it, the difference being done about 60-90 days faster than via retail.
The company I sold to already owned four houses on my street. It's crazy.
I could have a principle and live in a falling apart money pit in a declining neighborhood or move to a brand spanking new home in a neighborhood on the way up.
Part of the problem is how hard it is to sell a home in the first place. I'm not interested now, but for a while I was looking at needing to move states, and that was going to involve selling my home and buying (or renting) in the new location. And all the math was saying that anything other than getting really lucky with a sale just before moving was going to cost me a LOT of extra money and be a drain and a hassle on top of all of the stresses involved with moving to a completely new state.
I really hate the idea of selling to these "we'll buy your home fast" shops, but I have to be honest that had I needed to make that move, it would have been a very real possibility.
That's not really meaningless though. You just explained in more detail some attributes of corporations, many of which are precisely the things many people are criticizing when they criticize the fact that corporations own lots of housing. Of course we all know that there are still humans behind the corporations.
Depends on how unique your legal name is. Buying your own home as an individual creates a public record with your address and your name. This gets ingested by lots of people search websites like https://www.fastpeoplesearch.com/ (which I have used). But if your legal name is really common, it would be harder for anyone else to deduce which one is actually you.
For what it's worth, I just searched that site you linked with my name and the zip code I lived in for the first 18 years of my life, and it seems to have a pretty muddled view of who I am. It has my correct age, full legal name, and current address, as well as two of my past apartments and my address for those 18 years where I lived in the zip code I searched, but it also lists my parents current home that they bought over a decade after I had permanently left that state in a city I've never lived in. The landline number it lists is the number my parents had at the house I grew up in, but they don't even use it now, and it wouldn't have been an effective way of reaching me since I used to live at that house, and the mobile number it lists is one I've never seen in my life. For family members, it lists one of my brothers and parents, but not my other brother or either of my two living maternal grandparents, although it has a strangely long list of names I've never heard of, some of whom have my mother's maiden name but aren't my grandparents or any of my relatives with that name who I'm aware of, as well as a bunch of people whose names I don't recognize with last names I'm not aware of being in my family tree.
I'm not saying that site isn't potentially useful as a starting point to find out some stuff, but it hardly seems worth influencing a major decision like what legal entity to purchase a home with.
(edited to add): It also says there's no public record of me being married, which definitely is not the case. My wife literally co-owns and also lives in the house it lists me as living in (and has the correct purchase date for it), so you'd think that whatever algorithm is used to build the dataset would be smart enough to see if there's any other ways we're legally tied together. It also says it doesn't have any records of business associations I have when last year I registered a single-member LLC for contract work last year. The LLC literally has the same name as me followed by " LLC", because apparently no one else had registered that before in my state, which at least gives some evidence that my name isn't overwhelmingly common.
I'm not saying it's not worth trying to buy a house in a way that ensures privacy. I'm saying that there are probably better ways to consider the risks than looking at a site like this when it literally shows more people I've never met or even heard of than people actually related to me.
If someone is concerned about being stalked, they certainly should consider how to protect themselves if they're purchasing a home, but that would be equally true even if sites like this didn't exist. For someone who isn't otherwise already considering using an LLC to purchase a house for other reasons, I don't think a site like this is worth taking into account.
This suggests that a single-family home is generally a bad investment. Maybe not if you're living in the house yourself (no renter-landlord inefficiency), but still idk. Could make more sense in certain areas where all the value is in the land.
It's segmented. If you don't have a whole lot of capital but want to invest in real estate, you buy a rental home or two.
If you do have a lot of capital ($100M+), you don't, because you can spend that money to build or buy an apartment complex that gives you something like 5x as many tenants per dollar spent. And is cheaper to maintain in the long haul.
LLC may be a type of corporation but when people complain about corporations buying up homes they really mean C-corps, not LLCs owned by Uncle Bob who likes to flip houses.
Right, and when people share stats such as "60% of homes are owned by corporations", they're either clueless or are trying to deliberately muddy the waters.
About the most important tenet of communism is collectivism. When you attempt collectivization on a national level, there's always a significant portion of the population who doesn't want to play along and wants to keep doing their own thing. That's the end of the road for your political system unless you do a bit of mass murder, which is why every "successful" communist state resorted to that.
So yes, of course, no political ideology has "let's murder millions of people" as its founding principle. But some political systems require it.
or you let them be hermits that they are. No need to murder people. If they don't fit in with the collectivism then they are shunned from society. Much like social media.
It's important to remember the Anti-Comintern Pact started as an anti-communist agreement between Germany and Japan. Look what that did.
After making a secret deal with them to partition Europe. They didn't come around on principle, it's just that Hitler eventually decided to invade Russia too.
And after winning the war, Stalin proceeded to kill millions for good measure.
Stalin did the Molotov-Ribbentrop pact to (successfully!) buy time and breathing room for the inevitable Nazi assault after his numerous entreaties to Western/Allied powers were rebuffed.
That's one interpretation, another interpretation is that Stalin was expecting Hitler to struggle in France for years get worn down like in WW1 and then Stalin would attack and make whole Europe a satellite to USSR.
France buckled in months and Wehrmacht then attacked Red army which did not have setup defense positions because they themselves were preparing to attack...
That pact started the Eastern campaign of the Nazis. You could argue that it also bought time for the Nazis, as they could gain territory without worrying much about resistance from the Soviets.
I'm not really inclined to do too much Monday morning quarterbacking of the USSR's defeat of Nazi Germany. I'm especially not inclined to give it the time of day when it's coming from Westerners who never say a peep about their own countries having given Hitler territory and financing.
> And they are heavily moderated against negative discussion/ragebait.
So? You have to do that because it takes one toxic person to poison the well. HN is aggressively moderated to get rid of articles and opinions that don't belong too. Without it, it would be just a constant stream of self-promotion and politics.
The point is that in certain other places, someone (the moderators) worked to nourish a positive culture and it worked. HN didn't and it shows. I don't think that negativity is necessary to keep the forum interesting. Especially given that HN's negativity really isn't all that insightful. A lot of negative takes are bad, and many of them are written without reading the article, or by cherrypicking a single sentence and attacking that.
I'm saying it's aggressively moderated in some respects (off-topic content, politics, etc), but it's not moderated to root out a certain breed of snarky, I'm-smarter-than-you negativity. Many other forums police that second part and are doing just fine. This includes forums dedicated to technical hobbies.
In fact, computer science, electrical engineering, and mathematics are pretty uniquely toxic and we keep rationalizing it.
I remember working on a technical blog post for my company, trying to anticipate many of the possible HN rebukes and proactively address them as much as we could. And I remember having a conversation with a PR person who was genuinely taken aback by the hostility we've come to expect in our industry.
You don't get tech without negativity. And honestly HN is very tame compared to most forums when it comes to the deeply negative.
The problem with maintaining (only) positivity in tech is you turn into $large_companies marketing department. We have to step up and say security flaws exist. That companies outright lie. That some idea (when it comes to programming) are objectively bad.
Hence why the OP is here on the thread talking about what negativity means in this particular case, because it also counts criticism.
This is something we tell ourselves to rationalize bad behavior. How come that 3D printing forums or woodworking forums or car maintenance forums can exist without toxicity, but tech somehow can't? There are people pushing products everywhere. You can ban marketing content or set ground rules for it.
Further, performative cynicism really isn't that helpful. It's not insightful to hear that every company is evil and greedy, every personal project sucks, every scientific study is wrong, and every blogger is incompetent.
I think this is a common view, but it assumes that most of one's negative hot takes are good. And frankly, I've seen HNers being confidently wrong more times than I can count.
2) Two of the maze walls appear to be smudged, but nothing else in the image is.
3) The floral pattern on the plate is messed up - it repeats, but berries randomly change positions and the leaves change shape.
4) The milk carton has a splotchy texture that's really common in ChatGPT-generated illustrations.
I suspect it's not 100% gen AI - for example, the sharp outline around an out-of-focus xmas tree feels like a human-made artistic choice - but I'd bet it's at least partly composed from AI-generated elements.
I hadn't even looked at the floral pattern. It's awful.
I wonder how pissed Apple was when they realized they had paid him a few thousand (going by his posted rates) for "art" that used gen AI to a significant enough degree to have nonsensical components.
>for example, the sharp outline around an out-of-focus xmas tree feels like a human-made artistic choice
If I had to guess, it looks like he did the components of it separately and then just pasted them all in together. For example, you'd expect the light on the plate to brighten the milk carton near it. And the shadow on the plate doesn't line up with the bright light shining on the carton's side. These are just classic sloppy bad artwork mistake.
Notwithstanding your previous points, these sorts of AI detectors flag many false positives, they're not worth relying on. The only one that could actually work is specific watermarking in the images themselves, such as what Google does with SynthID in their generated images.
The structure of the carton is off. The spout. The opposite end of it.
Being a trained artist myself I don’t think many artists would deliberately make that kind of design choice considering given the overall style of the artwork.
Thing is, from a commercial standpoint, who cares? It's an image used to promote a TV show on a social media account. There will be a tiny, loud community of AI haters who make a stink on social media, but the vast majority of people will be oblivious, and the quality issues with the image will have zero bearing on whether anyone watches the show; whereas the fact that the post exists probably WILL drive a few to watch who wouldn't have otherwise.
The fact is we're in a world where algorithms consume more attention than ever. There's more content than ever. We're more sedentary and glued to useless shit on a screen than ever. We consume content like we're pigs, and what do you feed pigs? Slop. I don't like this, and I try to keep my diet as healthy as I can, but it's probably still worse than it was 20 years ago, and most people don't care at all. It's all consumption for the sake of mindless distraction. The slop exists because the demand exists. People will watch the show anyway.
> There will be a tiny, loud community of AI haters who make a stink on social media, but the vast majority of people will be oblivious
Not only oblivious but actively for it; I know many people who even watch fully AI generated content on TikTok or Instagram Reels. In fact, they know it's AI yet still like it, probably because some can look pretty cool or funny.
I actually started listening to AI music unknowingly, and was merely impressed when I later realised it was AI. However, I don't think AI videos have reached that point yet.
What's worth mentioning is that there has always been large quantities of "slop" even with human-generated art; the good stuff has always been a minority.
reply