One couldn't, because you'd need to standardize a JS-less DOM, which requires one to persuade Apple, Google, Microsoft, and Mozilla to agree on a new standard for a JS-less DOM API.
The DOM API is currently defined as a JS API, including JS strings, JS objects + properties, JS Exceptions, JS Promises, JS garbage collection, and on and on and on.
The effort to get all the browsers to agree to standardize a new JS-less DOM API would take years; none of the browser vendors even want to begin that conversation today.
> As a consequence of this succinctness, we show that basic verification problems for transformers, such as emptiness and equivalence, are provably intractable: specifically, EXPSPACE-complete.
But, if I have understood correctly on a quick read, they also claim transformers have pretty low expressive power. In particular, they claim they are limited to star-free subregular languages, whereas RNNs can recognize any regular language/simulate finite automata.
This doesn't imply you can't get aid from a LLM to e.g. implement a function that has a formal specification (an application I think is very promising), but surely it has some profound implications on how much of a large system can be understood by a LLM at once, without supervision.
Good news! On the current trajectory, I am very hopeful nonlinear RNNs could make a comeback. Which would incidentally help to ease the memory pressure for inference tasks.
The last line of the abstract has the most important takeaway.
> As a consequence of this succinctness, we show that basic
verification problems for transformers, such as emptiness and equivalence, are
provably intractable: specifically, EXPSPACE-complete.
If you were hoping to formally prove the correctness of a large transformer, it turns out that you're going to need an exponentially larger amount of space to do your verification, more than you could possibly afford.
How does it follow that there is no point in trying for formal correctness? In many problems there is an interesting subset that is quickly solvable even when the general case is not.
SAT solvers in practice are quick on just about everything.
SAT solvers being programs that solve the original NP-compete problem.
As far as I can see out-of-order streaming is only half the described functionality – there is also HTML streaming & revamped DOM parsing which does not have the positive signals that out-of-order streaming does:
Yea we've been working with Mozilla and Apple on that one as well but they haven't responded to the standards position yet. I am sure that we'll reach something that's within consensus.
It would be pretty wild if this feature allowed you to go back and add/remove individual tokens from earlier in the document and re-apply all the tree construction rules, like an even more unhinged version of document.write(). I think the actual proposal is expressed in terms of moving DOM nodes around and doesn't allow stuff like this.
It's expressed in terms of changing the output of where bides are inserted. Instead of new nodes being inserted to the template elements, they are inserted where the processing instructions were found.
As far as I can tell, this is not actually allowed by the current proposal. The definition of the "find markers" algorithm in https://github.com/whatwg/html/pull/11818 requires that the <?end> marker be a sibling of the <?start> marker in the DOM, so they can't have different parents like this.
Well, you're allowed to write it, but the browser will just ignore the <?end> marker. The range will end at the </em> tag, the same place https://github.com/WICG/declarative-partial-updates/blob/mai... says it goes by default if you don't write the <?end> at all. (That default wouldn't make a ton of sense if patch ranges were allowed to break out of their parent elements like this.)
Jake Archibald thinks it'd be cool to use a similar syntax to do CSS highlights, but there are other problems with that (like not being able to create overlapping ranges), and he's not on the Chrome team that's pushing this proposal.
What article were you reading? This article isn't idealizing Japanese companies, and specifically discusses the drawbacks of the Japanese approach, including zombie companies.
The article's thesis statement isn't "the Japanese approach is better," but that business practices like these bundle together, that they're very difficult to change, and that each bundle has different advantages and disadvantages.
Ironically, you've proved a deeper point about how amusing HN is: we all tend to project our fantasies onto the articles we're discussing, even if we didn't fully read or understand the article.
I did read it, but my impression remains the same. While the article does contain critiques of the Japanese system, as an East Asian, I feel it completely misses the actual underlying dynamics.
I know the author isn't trying to paint Japan as a utopia. The reason I call it 'romanticized' is because the author claims Japan's success in precision parts is driven by 'horizontal' and 'collaborative' practices. That just isn't true.[1]
In reality, this system is largely sustained by the ruthless squeezing of subcontractors (for the record, I am Korean, but I actually like Japan), which is a massive social issue there. It’s very difficult for me to understand how anyone could view this structural dynamic as collaborative or horizontal.
If the author had concluded that their success in these niches stems from being an extremely vertical society where defying your superiors is simply not an option, I would have fully agreed. That aligns exactly with what I have experienced firsthand.
>"The andon method is really the J-mode in miniature. Information flows laterally, authority to act is widely distributed, and the people closest to the problems are the ones who fix it."
Does your definition of a 'horizotal culture' actually mean forcing people to work overtime just to hit deadlines? Are you sure you haven't completely confused 'horizotally' with 'top-down'
P.S. The link I provided is an official directive from the Japan Fair Trade Commission (JFTC) explicitly warning large corporations to stop ruthlessly suppressing their subcontractors' labor costs.
The “Just-in-Time System” amounts to the exploitation of subcontractors. As a Japanese citizen, I am well aware of the reality of the Japanese economy. While server resources can be scaled up or down with a simple command, scaling physical factory resources is not so easy. Inevitably, this leads to suppressed wages for employees. Furthermore, while labor unions exist within large corporations known as “keiretsu,” they do not exist in subcontracting firms. There are laws in place to protect subcontractors, but in practice, they do not function (because if a subcontractor were to legally sue a major corporation, they would lose their business).
This country’s economy is built on the sacrifice of others.
Walmart and Amazon ruthlessly squeeze their suppliers. They achieve low prices on some things and try to corner the market on others (and then raise prices). What I don't see them achieving (to the contrary, I see them failing spectacularly at) is the quality control that some Japanese companies excel at.
So there has to be something more to it than that.
When U.S. giants like Walmart or Amazon squeeze their suppliers, the natural consequence is cost-cutting and a drop in quality control. But for regional companies in Japan and Korea, regional mobility is incredibly difficult. In Japan, for instance, even politics is often "hereditary," and there is a strong overarching tendency for people and businesses to stay rooted in one specific region (while many do migrate to Tokyo, a vast number of people simply do not have that choice).
In the highly fluid U.S. corporate ecosystem, mobility is always an option. If a supplier loses a contract with Walmart, they can still pivot to another massive retailer, even if it's not quite as large.
Japan and Korea, however, have small landmasses, and the reputational risk is absolute. If a company's reputation is damaged by a single failure or a lost contract, their next job simply vanishes. Because of this existential threat, they fundamentally cannot compromise on quality. Imagine what happens to a small supplier in Japan if they are cut off by a mega-retailer like Aeon Mall. There is no backup giant waiting to take them. They are finished.
So, while geographical and structural differences dictate this extreme pursuit of quality, framing it as a "horizontal culture" is completely wrong. As an East Asian, I can confidently say that "horizontal culture" is the single most mismatched term you could possibly use to describe East Asia
If you are from the West, navigating East Asian culture can be exceptionally grueling. At a fundamental level, our societies tend to be inherently exclusive toward different races.
Do you have a good book, article(s) or podcast to recommend on this topic (business culture in East Asia, or a specific country), in English & approachable by Westerners, while not giving just a cartoon overview of the issues?
The focus of the article is on why Japanese corporations excel at so many different areas. Are you saying that all the factors it mentions, like employees being trained as generalists, life-time and so on, are completely irrelevant? or are you saying they are important but some other essential factors were left out?
> In reality, this system is largely sustained by the ruthless squeezing of subcontractors (for the record, I am Korean, but I actually like Japan), which is a massive social issue there. It’s very difficult for me to understand how anyone could view this structural dynamic as collaborative or horizontal.
This is the picture painted for me by the article. Vertical integration eliminates subcontractors. Horizontal integration squeezes them.
> If the author had concluded that their success in these niches stems from being an extremely vertical society where defying your superiors is simply not an option, I would have fully agreed. That aligns exactly with what I have experienced firsthand.
Same story here. When switching jobs is made difficult, the incentive is not to make waves.
> Does your definition of a 'horizotal culture' actually mean forcing people to work overtime just to hit deadlines?
This is a two part question, the system and the idolise part aren't related.
You are ignoring a lot of quality Japanese product have been producing for more than 50 years, which answer partly why idolise Japan question. But the so called squeezing of subcontractors is true as of any other American companies if not worldwide and yet they failed ( or mostly failed ) to achieve similar standing to Japanese counterparts in terms of quality.
Exactly, I was confused too. The authors clearly mention what the parent comment talks about, albeit towards the end of the article, that the 'J' bundle meant that these firms were not set up for success once they 'caught up' and were required to innovate not just process but from the ground up to envision new categories (e.g. iPhone).
Thank you, I was confused reading the comment above, because the article pretty clearly laid out the benefits and drawbacks of the system. I didn't see any idealizing.
You argued that this article (by David Oks) is an example of "how Westerners idealize Japan." I argued that this article does not idealize Japan, and that, if you interpreted Oks' article that way, then you didn't understand the article.
I didn't say that Japanese business culture is more "horizontal" than Western business culture, or that Japanese business culture is better in any particular way. I didn't even say that the article is right or wrong about anything.
All I did was to restate the thesis statement of the article, to clarify what the article actually says.
I don't harbor any particular affinity for Japan, or Japanese business culture. I know very little about it. I'm not an authority to speak on it, and I didn't.
You assumed what I believe without understanding what I wrote. You did exactly the same thing to me that you did to David Oks.
There's almost no way these guys (low pass filter)can see the water they have been breathing since they were born unless they've had first hand experience that wakes them up (high pass filter)
Add to that (Western) (indie?) game-makers and hafu are both predominantly right-wing.. far-right, if you take a global perspective. Polite to the point of being offensive, though. :) Perhaps there are very very few opportunities in which to feel a loss of control in real life, or trained since infancy
That said, there's seems to be still a lot of useful info in article that bears "bookmarking"..
> In Japan, mobility is fundamentally expensive, and relocating to a different region is much harder than outsiders realize.
Unless said mobility is paid for by the company.
As part of the job rotation mentioned in the article, larger Japanese companies are also notorious for reassigning job locations, often at short notice and with zero care for family dynamics. Hence the tanshin-fu'nin phenomenon, where the husband is sent off to work at some factory or regional branch in the sticks for years while the wife brings up the kids elsewhere.
You seem to understand this deeply, perhaps even better than I do. When I worked in Japan, experiencing the corporate culture firsthand was much harder than I had anticipated. They even provided a translator for my boss and me during our business trip, but it was still incredibly tough. I do remember that, at the time, they were conducting suicide prevention training for employees in the wake of the Toyota new-hire suicide incident.
Anyway, you clearly know a lot about Japan, and I apologize if it came across as me just badmouthing the country. My original intent was simply to point out how drastically different things look from an East Asian perspective compared to a Western one. I didn't mean to keep painting Japan in a negative light—every system obviously has its pros and cons. I actually have fond memories of the unique, close-knit, family-like atmosphere that Japanese companies have
How does the "concentration" of gamedev jobs look like ? In US and I think in most of EU countries that have noticeable gamedev it is usually concentrated in very few cities, so changing job does not necessarily consist a move. But other industries similarly usually have a niche and sometimes whole towns that rose around it.
I think main difference is that there is very little tradition of company thinking they own something to the workers, and (I think) far more of companies just buying out their competition and then gutting any tradition and institutional knowledge within
> (Moving in Japan involves massive upfront rental fees like shikikin (deposit) and reikin (key money), making the physical act of relocating extremely prohibitive.)
I don't think that part is all that different? While we don't have "key money" it's still a big deal to take your life and move it somewhere else
Rovelli writes, "I fail to make sense of the claim that there is such an “explanatory gap.”"
Carlos Rovelli has failed to understand the arguments for dualism, and is proudly sure that they must be nonsense.
If there's ever to be a "solution" to the dualism/materialism argument, it cannot possibly end in a "slam dunk" where it turns out that one side or the other was simply nonsensical.
IMO, the problem is actually one of epistemological framing. If I ask what "I" know, assuming that my internal experiences are the basis of my knowledge, then I can't accept materialism. But if we ask what "we" know, as a society of scientists and philosophers, together we find only natural material, and no evidence for dualism.
(It's like the prisoner's dilemma. What's best for me is to defect. What's best for us is to cooperate.)
> If there's ever to be a "solution" to the dualism/materialism argument, it cannot possibly end in a "slam dunk" where it turns out that one side or the other was simply nonsensical.
Huh, evolution vs. creationism, many arguments happened over many years, yet one side was simply nonsensical.
> if we ask what "we" know, as a society of scientists and philosophers
That is how science is done; if you reject that approach a priori, no wonder your conclusions become unreliable.
I don't think creationism is nonsensical, it's just wrong. But the concept overall is not nonsensical - in principle, if the universe were very different, a god could have molded humans out of clay and breathed life into them or whatever other fairy tale is preferred; it's not a logically inconsistent, so it's not nonsensical. Even something like Lamarckism is not nonsensical.
If you want to see an obviously nonsensical world view, you need to look at something like the Time Cube "theory". Rovelli is essentially claiming that dualism is more in this area - which I agree with the GP is quite unlikely for such a long discussed and influential philosophical idea.
> In an ideal world, LLMs would take all of the basic RTFM style questions, and leave SO for the harder, but still general enough to be applicable to others-questions.
I think the deeper question is how SO would get paid for that.
Historically, SO has been funded by advertising. Users would google their question, land on SO, get an answer, and SO would get paid by advertisers. (The job portal was a variation on the advertising product.)
Even in your ideal world, newbies and experts would first ask their questions to an LLM. The LLM might search SO and find the answer there, but the user would get the answer without viewing an ad, so SO wouldn't get paid for that.
The same issue is facing Wikipedia. Wikipedia isn't funded by commercial advertisers, but they are funded by donations, which are driven by ads. If LLMs just answer the questions based on Wikipedia data, the user won't see the Wikipedia ad asking them to donate; they may not even know that Wikipedia was the source of the information, so they may not even develop a fondness for Wikipedia that's necessary to get users excited to donate.
This is why you see people shouting about how LLMs are "killing the web." I think it's more correct to say that LLMs are killing free web resources. Without advertising, not even donation-funded resources can remain available for free.
Oh, I was thinking more of user enters question into SO -> LLM answer on SO -> user evaluates whether LLM answer was sufficient (or system itself judges whether answer is also interesting to other users?) -> question + answer combo made public, judged by other users.
There are of course several huge issues with this, but thats why I prefaced it with ideal world hahaha
the biggest of which is why most users would want their questios publicized if the ChatGPT answer not on the stackoverflow platform will be enough or even better
Or how existing users and question-answering volunteers feel about just being cleanup and training data after LLMs
The DOM API is currently defined as a JS API, including JS strings, JS objects + properties, JS Exceptions, JS Promises, JS garbage collection, and on and on and on.
The effort to get all the browsers to agree to standardize a new JS-less DOM API would take years; none of the browser vendors even want to begin that conversation today.
reply