HN user
The pure logical concept of a Turing machine is a tuple of states, symbols and a transition function. These all very much do exist, and they're very much not written on the tape. So no, there isn't "only the tape and its content".
I do not agree data is a derived term in computation theory. Code, maybe I can be convinced. But absolutely not data.
You are redefining both the term "code" and the term "data" in ways that I don't think most people would agree with - and more importantly, in ways that make it impossible to talk about von Neumann architecture.
That's less "code is data is a fact of computing" and more "if you ignore the part of code that's not data then all code is data".
I'm worried about your unusually low threshold for suicide.
LLM is only devaluing security work by competition. As in, it's making dependency scanning so easy and cheap that researchers don't have much left to do on that front. The end result is very easy and cheap dependency scanning available to everyone. It makes the next heartbleed less likely, not more (though actually the class of bugs it protects against is closer to log4shell).
I would argue that a reply to "you can't scale your house consumption" doesn't need to say how much the situation sux or what amount of gadgets will or will not unsux it.
You can get entry level Chevys, Kias, and Hyundais for < $25k, which is about $14k in 2001 dollars
For comparison, an entry level car in 2001 was <$9k. That's <$18k in today's dollars.
The extra safety features is part of why transportation has become less affordable, yes.
I actually don't have good data on this, but it seems like there's pipeline is still borked from COVID era irregularities.
That, plus Obama-era Cash for Cars, plus engines becoming so complicated and expensive to fix that 10+ year old broken cars are scrapped rather than repaired at much higher rates.
You used to be able to buy a car for literally less than $1000 that would last you a few good years with minimal maintenance.
Repair costs of old cars: Repairs can be expensive, but again, cars today are much more reliable than they were in the 90s or earlier.
Not 5x more reliable, as in requiring 5x fewer engine renovations over 15 years now that it costs 5x more to do so.
Which is an extremely important point to make when the claim is "you don't even have a choice" and "you can't scale".
There is a choice. We would just prefer nobody had to make that choice. We consider it a failure of the modern society that this choice is on the table. But housing consumption can be scaled, and that it can has important economic implications (such as, social housing can be made cheaper by making units smaller. It literally wouldn't be possible to pull it off if housing consumption was perfectly inelastic.)
You misunderstand. Keep your job. Move away. Drive to work 3 hours one way everyday. Have no life except work, commute, sleep and family.
And yes, it will drive your QoL to the ground, I'm fully aware of that. I'm just saying it's neither impossible nor infeasible nor unheard of. It's "merely" undesirable. Extremely undesirable. But possible.
I didn't say it's nice to downsize. I said it's possible to downsize.
That's why I said "worse come to worst". Where I'm from, it's not that uncommon to drive 3 hours to work, and as a result have no social life outside closest family. People do it because the alternative is homelessness.
The food also became qualitatively different over the same period. We're not eating the same meat, bread, fruits, cheese etc. that our grandparents did at our age.
Durable goods became much less durable, too. Furniture is the biggest one I'd say - they're just not built to last anymore. Whether it's planned obsolescence or just extreme cost cutting, we aren't comparing apples to apples here.
You can to a degree. You can buy/rent a way too small house/apartment for your large family (what many immigrants do). You can also get roommates. If worse comes to worst, you can move another hour or two away from the big city.
I don't know about the first half of those 50 years, but over the second half (2001-today), transportation - or at least cars - got significantly more expensive. Disappearance of cheap new cars + massive supply reduction of used cars + repair costs of old cars + gas prices.
Because despite geographical location, Australia is a western country, and all western countries are like that.
In the past, the vast majority used not to.
Try to actually use it on a site and chances are you'll get an ADA complaint.
I really doubt that. We've had 2 decades of godawful letter-based CAPTCHAs and I don't think anyone got into legal trouble because of it?
People didn't write tests not because it was boring, but because they didn't have time to do it because implementation is more important. That problem didn't go away - it's just instead of not having time to do tests, now we don't have time to review whether the AI wrote tests that make sense. And most of the time it doesn't. They're running the code but not asserting anything. Or they're asserting test inputs instead of implementation outputs. Or outright swallowing errors to make tests pass.
Having such tests is, in fact, worse than not having tests at all. Non-existent tests are just as useful as bad tests, and are much cheaper to maintain.
Not from what I've seen. Still lots of 1.0's, 0.1's etc. counted as 0.
I really wish the analysis didn't cut off at exactly 10 numbers. Rerunning the query with higher LIMIT, I noticed the following:
- The very next two numbers are 100 and 0. While 100 is consistent with the article's explanations, 0 still placing high despite fixing the query warrants further investigation. A quick glance at returned headlines shows the problem of phantom zeroes is not, in fact, fixed.
- The query doesn't group decimals and integers together. 2.0 is at #17 with 10k hits, while 1.0 is at #26 with 5k hits. So not only is the "version number" explanation for top numbers wrong - the claim that earlier version numbers are more common than later version numbers is wrong too.
It's marketing. Attach yourself to some wildly more popular brand for no reason other than to boost your own popularity. Like JavaScript did with Java.
I'd say checking out a malicious branch is in the same category as downloading a malicious attachment. By which I mean, it's kinda on you.
The biggest obstacle to huge refactoring has always been minimizing the risk of bugs, not losing any features, and ensuring compatibility with the existing ecosystem. The reason it's become easier in the age of AI is because we stopped caring about these things.
Using the allocator interface is only required if you allocate, which in this context you explicitly don't. If you want to use static preallocated memory in Rust, you'd have to use #[no_std] and only have access to the core part of the standard library. On the plus side, this is a very well supported and stable configuration, the core library is still pretty rich, and there's plenty of no_std libraries to choose from in the ecosystem.
I used "most" in the pedantic sense of "just a little bit over 50% if you count individual lines and not whole libraries". Most code doesn't operate on Box, Vec, String etc. directly, and is happy with & and &mut to underlying data. And even when it does, it's usually used for giving static lifetime to data, which is a non-issue here because all your preallocated memory has static lifetime already.
Is it nonsense to say C doesn't have any protections?
Why do you recommend the living definition of not having any protections whatsoever for protection?
The funny thing is that Rust works perfectly fine with statically preallocated memory. It's not even against its idioms - most code will work with zero changes. In a way, the borrow checker is the perfect tool for this situation, since nothing owns any memory and everything is borrowed from the static allocation.
I like my datacenters like I like my railway infrastructure - covered in enough protective layer of rust that it will never suffer structural damage even when running for years under extremely heavy load.
If you literally want one and only one allocator instead of the Rust's built-in one, that's been supported for years[1]. Almost 8 years, to be precise. It's very easy to implement and works with the entire standard library and all its abstractions, and almost every 3rd party library will support it out of the box too (except the ones that go out of their way not to support it for some reason). Borrow checking and all the other safety features are still fully supported.
Now, mixing different allocators is a different beast, and much less supported. But it sounds like you are very much not interested in this use case, right?
p<=0.05 is so easy to achieve for actual positive results (just increase sample size) that I don't think it's productive to treat any result with p>0.05 as positive evidence for any purpose other than deciding to rerun the study with a bigger sample size.