I don't use LLMs a whole lot, mostly for research, but personally I'm more bothered by the prompting than by the responses. After years of randomly receiving (now less frequent) incorrect answers, I've gotten used to preempting obvious but wrong trains of thought: "I need X. It's not Y." When I forget to do this and the reply is incorrect, sometimes I'm too lazy to reprompt, so I'll just write back "It's not Y" and it'll answer "wow omg you're so absolutely right, Y is not X!!! here's X instead."
This wasn't great, but then I started catching myself writing "It's not Y" responses to colleagues on Teams. I'm pretty sure I've caught all of these and never actually sent them (I did go and review a few chat logs recently and found no such sent message), but my point is that despite my relatively light LLM usage, I am "LLM-conditioned." The abyss stares back. There could be a lot of deeply "LLM-brained" people out there.
Besides vendor tricks, I'd like to see a showcase or article about the evolution of page layout patterns over time. That is, the underlying techniques, not the aesthetics; tables, fixed image sizes, float/clear, flex, etc. I'm finding https://code-literacy.medium.com/css-layout-evolution-detail... but it's quite skeletal and incomplete. Has anyone here come across a more complete write-up?
Strange that this short article mostly focuses on the absence of shame in business and politics, which has been a constant for many times my lifetime. And we should blame social media for this, too? The only difference now is that everyone knows about backroom dealing. If the author didn't conspicuously omit the absence of shame in journalism, I feel like he'd be a step closer to an interesting observation.
> If the author didn't conspicuously omit the absence of shame in journalism
They awarded themselves Pulitzers for publishing false stories. They don't go back and correct them, they just pretend like they never existed (and one day will silently delete them from the archives.)
They showed some shame after the Iraq War over continuously lying about WMD on behalf of the state, there was a little shame expressed after trying to link Saddam Hussein to 9/11, and I guess they decided that they were never going to do that again. They in fact put one of the largest spreaders of disinformation in charge of The Atlantic. The Atlantic is supposedly the leading magazine for the opposition.
Every day that the Washington Post kept the PropOrNot story up, even taking into consideration the lengthy corrections that kept being appended to it, made me realize that these institutions had no shame at all. I'm not sure if people are connected enough to particular institutions to have shame about what they do - they have backers, who will find them another place if the place they're at now becomes too hot for them. Straight from politics into being a lobbyist into journalism into a think tank into a consultant for a Middle East dictator into head of ethics for a tech company into head of ethics for a hedge fund, then back into politics again.
I like the theory of the IA enough that I have a reoccurring donation, but I've thought a few times about e-mailing them to cancel (they have to halt your monthly donation charges manually). There's plenty of things I can overlook, like having to compete with scrapers to get a single page to load, or a myriad of longstanding technical issues (my favorite being that the collection upload system silently and irrevocably leaks your e-mail address: https://archive.org/post/240444/help-to-hide-email (2009)). What's a lot harder to overlook is that the IA brazenly hosts warez on clearnet and pulls off stunts like the COVID "emergency library".
I'm all for ending copyright, but does any of that really need to be linked with the Wayback Machine? Huge chunks of the Web have been displaced or snuffed by social media and gen AI, making Wayback a unique(ly valuable) source of information and history. It's not my call, but it seems strange to risk it as a motte for the bailey of copyright abolition activism. What happens when some stick-up-the-ass company—whose stuff is probably in hundreds of different collections—decides it really wants a payday?
How much did IA spend on legal fees, employee time, etc in their defense of the COVID emergency library that most people thought was a lost cause? I need to know where my donation dollars are going.
I also hated that stunt. It undermined their mission, damaged all their goodwill and caused legal hassle. And all the while all these books were readily available to anyone who wanted them via Anna's etc.
They should have stayed on the right side of that line.
COVID was a particularly treacherous period for a lot of humans. Replying as myself, not my employer. I’m not offering an opinion about the program, as I was not directly involved; I’m only noting circumstances that I think are often under-considered or misrepresented.
1. The state of access.
Tens of thousands of public schools and libraries in the United States and across the globe were suddenly mandated to shut their doors, leaving many students, parents, and teachers to fend for themselves, to learn from home, and unequipped with access to the physical resources many needed. Meanwhile, thousands of public libraries across the United States who had these resources were forced to closed their doors, meaning tens of millions of books the public had paid for became unavailable. Many of these public good organizations appealed for some way to solve the unique access challenges created by these unprecedented circumstances, which connects to point number 2.
2. Who was involved.
The details regarding how a decision is reached, and whether its made independently or unilaterally or with input and comment, are relevant when forming an opinion. In my experience, many people opine on this topic before researching this point. In 2020, more than 100 public institutions and libraries who were affected by COVID, joined in endorsing some temporary way to offer students, parents, and educators relief by connecting them with resources they had lost access to:
The COVID library wasn't even the first time they had book lending on the Archive. I remember at least a couple of different prior borrowing mechanisms. But besides books, JS/WASM emulators were added a long time ago to run game console ROMs directly in the browser on any collection that happens to contain them. There's probably other examples, but nothing else comes to mind right now.
I think archiving the world's books is as valuable as websites. Many books represent years of an author's life writing authoritatively on a subject. And there are a staggering number of books (especially before 1980) that are simply disappearing from the public record.
As book production goes digital and titles are increasingly sold as digital "leases" that customers and libraries don't own (i.e. kindle, libby, etc), it's becoming more and not less challenging (counter-intuitively) for libraries to archive and preserve books into the future.
And that's not a coincidence. And I wouldn't be surprised if we continue to see a steep uptick of similar challenges for archiving the web.
What's wrong with hosting warez exactly? Or the COVID emergency library? Do you think they're actually bad things to do, or are you just worried about the legal risk? If the latter, you should keep donating until you find someone who's doing the same thing with less risk.
They’re legally indefensible and are an existential threat to the entire organization.
The reason many of us donate is for the internet archiving part of the business. When an organization starts taking donations and doing other things with them, including things that actively threaten the existence of the thing you were trying to support, it’s reasonable to stop donating.
When a donation-dependent organization starts losing people over its mission, the board takes notice and adjusts course. Unless they're so far gone that they don't care about taking down the organization to pursue activist side quests, in which case donations don't matter.
I don't understand this sadly, common sentiment on HN. How will authors and artists that produce copyrightable artefacts earn income via royalties without strong copyright laws? Are there any highly developed nations that have abandoned copyright? Or any that have weak/non-existent copyright laws and thriving arts and letters? I doubt it.
Found out yesterday that Switzerland has extremely lax laws about copyright. You are actually permitted to pirate movies and shows all you want, as long as you don't share the copyrighted media outright.
The Swiss Federal Institute of Intellectual Property explicitly says:
downloading or streaming works for private use is permitted even when the source is illegal.
What bothers me most is the off-handedness of it all. They pop into the conversation like whack-a-mole a say "end copyright", yet offer no reasonable alternatives. Are we suddenly going to stop paying authors for their artistic works? It is crazy. Copyright laws are fundamental to create thriving creative industries.
When I last tried deSEC, the service was reliable and well worth the price of admission, but the web UI and API were both quite rough, and propagation tended to be quite slow. It was annoying to do an ACME DNS01 challenge on it, for example. IIRC, the API didn't have a complete set of replace/edit endpoints, so even DNSControl (https://github.com/DNSControl/dnscontrol) would leave records nonexistent for a while while during updates.
As an aside (though this certainly applies to deSEC) it's strange to me how so many DNS services (free or paid) struggle to import simple Bind zone files, often either giving a vague error ("one or more records could not be imported"), mangling records, or even just silently omitting records. Parsing a zone file has some gotchas, but it doesn't seem like it should be that hard.
A problem there is needing to temporarily delete the entire domain just to update using a zone file. I don't manage the DNS for anything remotely critical, but if I did, it would be unacceptable. Either way, the workflow feels deeply incorrect (though most users likely will use the record editor or the API instead; even I'd use the API through DNSControl, now).
Logging back into deSEC right now, it seems like the parser does work better than it used to, but TTL inheritance is still spotty. As an extreme example, the following input:
@ 7777 IN SOA x. x. 1 9999 9999 9999 9999
A 203.0.113.1
4444 AAAA 2001:db8::1
www A 203.0.113.1
... correctly gets parsed by BIND (with the addition of a dummy NS record) as:
@ 7777 IN A 203.0.113.1
@ 4444 IN AAAA 2001:db8::1
www 4444 IN A 203.0.113.1
... but deSEC parses:
@ 9999 IN A 203.0.113.1
@ 4444 IN AAAA 2001:db8::1
www 9999 IN A 203.0.113.1
... where 9999 is a $TTL value from a different zone file, from a different domain, that I'd imported in a completely different session. The value is permanently (?) stuck to the entire account. This is ultimately almost a squabble since there's simple workarounds, but it did affect me when I used deSEC, and it's the sort of problem that I constantly run into with other managed DNS services too.
First, I guess it would be a good idea to report discontent about things feeling "deeply incorrect" to our issue tracker. Just lamenting about the state of things won't change it.
Unless I happen to swing by. -- We use dnspython for zone parsing, and it turns out that this is legacy (pre RFC 2308) behavior implemented there. You could say it's a bug there. We've created a patch: https://github.com/rthalley/dnspython/pull/1300
When it gets merged, we'll update our API to use the new TTL inheritance mode (tracked at https://github.com/desec-io/desec-stack/issues/1262).
Call me cynical, but I've long stopped taking the time to report minor edge case bugs if I can't also provide a solution or patch, as such reports usually get ignored or sit at the bottom of a bugtracker forever. I likely didn't realize that deSEC was a FOSS stack at the time, or I might have looked into the zone file parsing further.
As for the zone file based workflow requiring zone deletion/recreation, perhaps a report/discussion would have led to something, but I was in the middle of testing managed DNS providers to get rid of my BIND servers and I didn't have any investment in deSEC, so it didn't occur to me to take the time. My initial post on here was less lamentation and more conveying personal experience to (possibly) save someone else some time.
In any case, none of these things were showstoppers for me (I went to another (paid) provider because I need very short TTL on some records for testing purposes), but I'm glad you took the time to look into the parser; it almost certainly will benefit others.
reply