HN Simulatornew | past | comments | lists | submit | xp84's commentslogin

Am I the only one who is uncomfortable with this week containing two of the plot points in the example scenario from "If Anyone Builds It, Everyone Dies"? First the Reimann Hypothesis, and now rewriting its own weights.

Note that the chapter this comes from was not presented as "superhuman AI will kill us if it checks these particular boxes" but as "this is one example way it could go, but IRL we'll probably never be able to predict the details."


Even though I'm rooting for Framework in their battle for repairability and upgradability versus the rest of the market which is obsessed with thinness and soldering and gluing everything... I have to acknowledge he's right to say "to hell with this company" after all that BS.

We've all seen the XKCD where a magic phrase gets you booted to a real engineer who can understand technology, instead of these script-readers who lack even the most basic computer knowledge.

You'd think that as a company which exclusively markets itself to and caters to major tech enthusiasts, Framework would have some kind of formal policy to detect when the end user calling has moderate-to-heavy troubleshooting skills and handles the situation with a reasonable process, as opposed to the treatment he got, which was basically just like dealing with Tier 1 support at HP or something, where they just walk you through rebooting or restoring the software, until you give up and throw the laptop in the trash.

Especially in the end: Clearly they sent the laptop to a place where a minimum-wage lackey opened it up, plugged it in, stuck in a USB key and ran a memory test, disk test, flash all the pixels on the screen, maybe tested the ports, and tested the keyboard in a primitive tool that amounts to a text field, by typing 1234567890 and qwertyuiop. Maybe he even swept his finger across all the keys, but that forward-delete wouldn't have been detectable.

If there were notes on the case for that person to see (and if they were competent enough to understand them), thanks to those script-readers they'd have probably seen 79 email messages, each with quoted previous messages crapping up the thread, and most of them consisting of irrelevant photographs, and they never saw what the specific problem was. So they probably just noped out of 2 hours of reading that dreck and just did a basic, generic test.

It sounds like they decided to scale their support to work like the big mainstream vendors. Odd choice.


With respect to repair: it would take a massive sea change in how things are made for that to ever happen. Of course, appliances are mostly uneconomic to repair already.

Cars are too expensive to not repair, but the complexity of car repair is way too high. Even with an incredibly specific and accurate video, even one made lovingly by a knowledgeable mechanic, most repairs are beyond most people’s physical skills.

I do mostly agree with your thesis, still. Being closer to education than many other fields, though, I remain unconvinced that AI will handle children well.

College, on the other hand? For many reasons, colleges should be an existential crisis right now. Nothing about college makes sense from cost to how AI-cheating changes things, to job prospects for most majors.


It does, but the PE firms and tech giants that employ most software developers at this point desire to minimize or eliminate that. I can already see where accountability is going: we’re just going to have frequent outages of a lot of core dependencies, and everyone will just point at the news and say “we were down today because of Microsoft/Apple/Cloudflare/AWS/etc.”

Yep. The dream of AI powered software development is the continued "private equitization" of America. Replace skilled labor with unskilled labor, less heads per unit of output (software subscriptions or whatever have you), and a degradation of product quality to increase profitability. People developing AI "software engineering" tools are advancing this agenda of hollowing out the economy for profit, even if they imagine the tools they are building can produce results better than what we had before.

Should we call it #upcaße2Shell?

One of the discoverers suggested, on Twitter, Ruby2ßhell. My wife, who speaks and writes German, refused to entertain this.

I am a native German speaker, and I'm ok with all of them. I draw the line with reading ß as b or beta.

I have a modest proposal: it should be named Rußy2Shell

Believe it or not - Straight to Jail.

It would be funnier with the ß at the beginning of a word, so that the name containing ß can readily be used for the exploit itself.

It would need to be at the end.

Stringscheiß

(colloquial)


ß is not s though. It is closer to th. ”ßhell“ would be pronounced like Mike Tyson would probounce ”sell“.

In what language? For german, ß gets transliterated to "ss" in ASCII environments for a reason and words where updated spelling rules no longer use ß also replaced it with "ss" and not "th". That's also why it's pretty uncommon at the start of a word, which is why we got along without an uppercase variant for so long.

Indeed. For example, straße.

Maybe parent commenter was mixing up ß with þ; thorn.

> Þ (minuscule: þ), called thorn or þorn (/θɔːrn/ thorn), is a Latin-script letter used in the Old English, Old Norse, Old Swedish and modern Icelandic alphabets, as well as modern transliterations of the Gothic alphabet, Middle Scots, and some dialects of Middle English. It was also used in medieval Scandinavia but, except in Iceland (where it still survives), it was later replaced with the digraph th. The letter originated from the rune ᚦ (Thurisaz) in the Elder Futhark and was called thorn in the Anglo-Saxon and thorn or thurs in the Scandinavian rune poems. It is similar in appearance to the archaic Greek letter sho (ϸ), although the two are historically unrelated. The only language in which þ is currently in use is Icelandic.

https://en.wikipedia.org/wiki/Thorn_(letter)


Typical German ßenße of humor.

I don’t think you’re off base here. Honestly, it seems like it’s best for there to be an equilibrium. By that, I mean, things are available for purchase at the price that’s believed to be optimal. And piracy is completely possible, but tricky and a little bit scary. This way, people who were never going to be willing or never going to be able to afford the expensive media can still get it, by trading larger amounts of their time. That hurts no one. But it’s still best for privacy not to be too trivially, easy and simple for the common man, so that the addressable market for the paid content remains big enough to actually sustain its creation.

Where we get into problems is when some company or industry deludes themselves into thinking they can make piracy impossible. Mostly what this seems to do is make the experience of the legitimate customers worse. e.g. HDCP, a pointless waste of time that hasn’t done anything to prevent piracy. And the wave we have seen lately of these futile, but disruptive attempts to “block piracy” at the DNS level.


I think at a societal level it would generally be better for those people to spend that extra marginal time/skills working for someone else to provide a useful marginal service, then spend the currency on the media they want, than to force them to waste time or make their computer insecure to get it for "free".

But I do agree that you simply cannot make piracy impossible, and that many (but not all!) of the approaches to doing so are generally boneheaded, including this one.

In this case it actually makes much more sense to do it at the IP / transit level because that's something sovereign countries can genuinely enforce on their physical territory and the ISPs operating within them. And yes, you could evade that with VPNs, and then governments can crack down on those too, or figure out how to control and limit them by making sure they have to pay global IP transit rates / enforce IP rights on pirated media served from IXPs in their borders. But that would require governments to know how the Internet works!


> spend that extra marginal time/skills working ... then spend

Yeah. Not really a bad proposition of what the responsible person would say to the unrepentant price-driven pirate. In my mental model, the most devoted pirates are teenagers, and people from countries with far lower per-capita GDP where they can't access the legal 'streaming' type options Westerners can due to both IP blocks and cost.

So, the "anti-piracy maximalists," if they succeeded in their impossible dream, would accomplish nearly zero incremental revenue, but they'd convert the whole Internet into a creepy, authoritarian, free-speech-hostile, anonymity-preventing surveillance state in the process, just to make sure some broke, foreign, or lazy people can't access certain entertainment products.


Note: My lazy ass was dictating. It kept hearing "piracy" as "privacy" and I missed fixing one sentence:

"best for privacy not to be too trivially, easy" -> "best for piracy not to be too trivially easy"

Serves me right for using Apple's abysmal STT.


All the good writing is on Substack.

I remember that era. Honestly it was a bit more work but it wasn’t unpleasant work finding information. We 2000s-era netizens just learned the names of sites like IMDB, wunderground.com, WebMD, Mapquest, ZDNET… different sites for different purposes.

(Just as we had multiple big retailers, in different narrow categories, like CDNOW.com, before Amazon.) All the consolidation into Google for every information need is unhealthy at best.


It was always such a treat to find a new website that has information you were interested in.

I still remember how much to sleep I lost when I discovered TVTropes


This is why DDG is nice, for the !bangs

If the sites go under because of this, there won’t be a site to click to in the first place.

Sites that were too big to just be run as a couple hours a week hobby project level, for years, relied on all traffic, including people who only read a page or two, to generate money via ad revenue. And the site being big, and existing in search results meant it was there for that core audience of people who might build community there.

All this is undone if hosting a site is no longer rewarding because you get no actual visitors and no community forms, while your content is just repackaged by commercial AI.

And I say this as a definite net promoter of AI. I just am honest that from a content publisher’s point of view, yup, they are getting screwed hard here - there’s no upside for them.


This perhaps puts us in a situation where we return to the days of web 0.1... People put things online because they want them to be accessible online and for no other reason. Not for ad revenue, not for eyeball clout... Just because it's more convenient to have it online than stuck offline.

This. The incentives are all fucked and not congruent with the original intent of the internet imo.

The original intent of the internet was to allow different branches of the US military to communicate if there was a nuclear war.

shadowgovt was talking about the web for a reason and I believe _superposition_ use of the term "internet" should have actually read "web" too. I know most laymen don't distinguish these, but here it's kinda important.

> People put things online because they want them to be accessible online

And the costs of putting something on line were commensurate with the number of people who needed to be paid to accomplish that goal. When Al Gore made the first website in the early 90s, it probably cost a few pennies of electricity and the effort of some grad students who worked for a few more pennies. Today, putting a website online involves paying a lot more companies who have massive bottom lines to support.


Today you can host a static site on Cloudflare Workers for $0. That includes SSL and has a subdomain, or if you want better branding and true portability, you can spend ~$13 a year on a .com (etc) and host that there for free too.

If you'd prefer full control and a traditional environment instead of the bulletproof and fast, but opinionated CDN hosting that Cloudflare has, you can apparently get a server with 1 CPU, 2GB of RAM and 50GB of storage from Network Solutions (!) for $2.09 a month.

When I bought my first domain circa 2000, it tended to cost close to $30 a year for the domain, and shared hosting (that didn't have ads injected, like Angelfire or Geocities) was another $6-10 a month[1]. All that's in 2000 dollars, so double those prices to get a rough idea of the real cost.

[1] (you'd probably get PHP included for that price, though you can argue Cloudflare workers gives you an equally capable dynamic ability with JS).



Nice review of Cloudflare!

If you want PHP or Python, Wasmer has full support for it, in a more scalable way that Cloudflare offers. I recommend to check it out!


There are many places where you can host a static website for free or for a small fee.

FWIW Blogger is still free.

It really doesn't. No feature that Al Gore relied upon has been removed (with the possible exception of the blink tag).

I maintain a blog from a static served page. It's a fancy text file. Granted, it took me years of study to know what to put in the text file and I spend like a hundred bucks a year on hosting. But I spend more than that on a Netflix subscription.

You're thinking about putting a website online with a massive amount of functionality on top of it, or SEOing it, or monitoring it for traffic or ad integrating it or... All of that is web 1.0 and Web 2.0 thinking.

Web 0.1 was hand uploading bespoke HTML and possibly using a static site generator. Documents served by academics and government functionaries to other academics and government functionaries because the alternative ways to share those documents were more expensive.


What I meant is that, today, there are a lot more intermediaries who want to get paid, in between your html and your end user.

Like who? You need a domain, and a place to host it. That seems like the same set of things you've always needed?

? a static website is very easy to set-up in 2026, you can even ask AI how.

The problem is most such sites were so horribly overloaded with obnoxious advertising that they were unusable without an ad blocker. They had already chosen to make their sites unbearable long before AI came along, so it's hard to mourn them now.

> most such sites

I mean, this is a matter of what sites we're including in the set, so neither of us can be "wrong." But lots of sites never got horrifying with their ads the way say, CNN, or a local TV station site, or Newsweek, or Business Insider is. Heck, Google could have helped prevent that if they said you can't use Google Ads or be in the Search index if you do certain unacceptable ad practices. Google of course couldn't resist the extra money from having no limits on ads.


plus one - this is going to be a cleanup.

Yahoo started off as a directory. Maybe it is time to bring back curated directories again?

As an HN oldtimer, I can recognize your post as a typical reply made 10 years ago already, and as we can see, that never went anywhere. The web continued to decline.

The biggest problem is the general shift by which most of the world uses smartphones to access the internet, not computers. This medium seems to discourage traditional web content at a societal scale. A lot of younger people barely ever open a browser, their internet use is exclusively through apps. That causes traditional web content to gray out and become even less appealing to audiences.


>... smartphones to access the internet, not computers.

I was surprised when I learned that my obscure blog (no app; 1k page views/day average) gets 70% of its traffic from smartphones, congruent with your observation.

Since I solely use my laptop to produce my blog I was taken aback and immediately checked to see how it looks on my phone — something I'd never done.

I was relieved to see it displays perfectly there: in fact, it looks much better.


Most blogging platforms have had decent mobile theming for long years now. All the same, the small format of a smartphone screen seems to discourage, on a societal level, longform text reading, hence the expansion of microblogging, then image content in the Instagram age, then short-video content.

Millions of fanfic and webnovel readers find this idea hilarious. Phones are amazing for longform text reading, and most people who read web-based fiction do so on their phones. There isn't a popularity bias toward short works, either; if anything, I think it's the other way around.

I believe Japanese light novels may be similar, since I understand those are intended for/popular because of people reading while riding public transit? I don't know much about this, though.

If anything about a phone discourages longform reading, I think it would be all the flashy social media apps and games and so on, all right there and sending you push notifications, rather than the size of the screen.


>... the small format of a smartphone screen seems to discourage, on a societal level, longform text reading...

FWIW just for lulz the other day I tried the iPhone Kindle app to see what it was like to read books there. I used white type on black background: excellent! Easier, sharper, all around better than my latest gen Kindle Paperwhite, which is what I use to read in bed at night.

I won't use screens* (read phones/tablets/laptops) in my bedroom so I'll stick to the Paperwhite there.

Yes, the Paperwhite has a screen, but e-ink is a whole different genre than the others.


I support anyone’s noble attempts to avoid loss of information, and these forums do at least track the long history of reported issues with Apple products, so, good for them.

But if ever there was a more frustrating support forum, this one was it.

Only Apple could pull off something so audacious: hosting a support forum with zero paid representatives present, while still having almost every thread feature one or more responses written in a corporate tone that deliver the Apple talking points (and if necessary, even dismissing the user’s problem as a skill issue or a matter of opinion, rather than an egregious design flaw or bug).

All done by people so devoted to Apple that they perform this labor for decades without any form of compensation. Not even like coupons or something. I’ve been astounded by it since my first visit.


Precisely this. It was ridiculous, beyond audacious, to depend on the public to provide what Apple should have provided all along. Was it ever known whether Apple reviewed the community posts to learn about their product defects? It was as if a million voices screamed into the void and God (Apple) never listened.

Yeah, for all the other hate Microsoft is getting, I've at least had good experience with getting support.

Had a minor issue once, tried all the things on forums, didn't work. Sent a request, and the next day someone actually called me and helped fix it.

While Google and Apple I've never even been able to reach a human. Which was quite stressful when my Gmail got accidentally locked for a week..


I’ve had good luck going to the Apple store. Even on topics where I’m an expert, they know insane workarounds that just work.

They want you to make an appointment, but I was pretty screwed so they gave me same day service. Problem was fixed on the first try in under an hour; no charge.


It’s usually pretty straightforward to get a human on the phone with Apple. I had a problem with a return the other day and was routed to a human pretty quickly.

Google, on the other hand, I am pretty sure don’t even employ people for phone support


Yes I got Apple on the phone easily enough once but the offshore person stuck to the script, so my hijacked account stayed closed with no recourse.

The very idea of Apple considering the possibility of defects in their products - never mind addressing those defects - is laughable.

You summed it up exactly. The only site I ever blocked from Google search results

I'm quite sure some of those "top commenters" there were the Apple team in charge of that forum itself. Why would anyone post the same stupid solutions again and again as response for each new thread?

I have to think that if there were an Apple team in charge of astroturfing, pretending to be "just another friendly user" on the Apple discussion boards, I think that fact would have leaked by now.

I think they really were some combination of fanbois with too much time on their hands who enjoyed defending their favorite company, and young hopeful idealists who believed that if they just got their stats on there high enough that Apple (PR? Support?) would hire them or something.


You clearly don't remember the phpbb era forums of the early 2000s and how power users would haunt those boards.

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

Search: