PeterisP

AFAIK the transponder kind of turns itself on when powering on the plane, you'd have to explicitly disable it but then you'd have weird discussions with the airport tower guiding you to a free timeslot on the runway which would just delay your takeoff, since ignoring the airport tower is a good way to not get off the ground at all because you'll accidentally be hit by some other plane.

I mean, it's an obvious difference in the primary use cases - if you explicitly want an isolated answer, you might clear context and start from scratch, and if you explicitly want a discussion with a persistent companion (as these people did before any of that psychosis started) then you won't do that.

The key missing step which breaks the loop is that while indeed a larger and larger portion of the web is written by language models, that data isn't being used to train new models - at the beginning of LLMs people did indeed want to use "all the web" to train models, but that's not being done now anymore, you either take only old pre-LLM data, or you pay for new 'clean' data, or take extensive filtering steps to avoid accidentally ingesting synthetic data.

The main phrase of the title "model collapse is happening" is untrue and not substantiated in the article - all the true statements in the article are about the hypothetical problem, warning of the bad consequences that would likely happen if makers of major models did something they aren't doing, but they aren't doing that because that is a known issue that they're avoiding. It's like writing an article "Foot shooting epidemic is happening" with a long, solid (and true!) proof that if you'll shoot yourself in the foot, it will indeed cause serious injury...

The Iran war was one of the biggest economic boons Russia could hope for; the disruption of oil and gas exports from Middle East with the associated spike in global prices brought Russian economy back from the dead, as now their exports are so much more valuable.

The root cause of identity theft in USA and some other places is the lack of "proper" national identity and the associated use of various personal "secrets" (not that secret) for identity verification because there are no good easy other ways.

Businesses in Scandinavia and many other countries would not treat someone knowing your personal information as any evidence of identity (because it's not); having all that information is not sufficient to impersonate you there - identity theft does happen but it would require stealing or forging physical documents or actual credentials to things like bank accounts; knowing all of what your mother or spouse would know is not enough to e.g. get credit or get valuable goods in your name.

The current bills (e.g. NY one at https://www.nysenate.gov/legislation/bills/2025/S8102/amendm... ) require age assurance that goes beyond mere assertions, so when creating your (adult) user account it would be required to give away your privacy to prove your age - if you can't implement a way for anonymous/pseudonymous people to verify that they indeed are adults (and not kids claiming to be so), these bills prohibit you to manufacture internet-connected systems that can be used by anonymous/pseudonymous users.

That would be a violation of the copyright law or the GPL licence - you aren't permitted to take GPL code and redistribute it with some extra restrictions added on to it.

If it's not (fully) your code, you aren't free to set the licence conditions; Linus can't do that without getting approval from 100% (not 99% or so) of authors who contributed code.

What one can do is add an informative disclaimer saying "To the best of our knowledge, installing or running this thing in California is prohibited - we permit to do whatever you want with it, but how you'll comply with that law is your business".

What I'm confused about is how the proposed bills would apply to servers.

Like, in general, a software change to add an "age class" attribute to user accounts and a syscall "what's this attribute for the current user account" would satisfy the California bill and that's a relatively minor change (the bad part is the NY bill that allegedly requires technical verification of whatever the user claimed).

The weird issue is how should that attribute be filled for the 'root' or 'www-data' user of a linux machine I have on the cloud. Or, to put aside open source for that matter, the Administrator account on a Windows Active Directory system.

Because "user accounts" don't necessarily have any mapping (much less a 1-to-1 mapping) to a person; many user accounts are personal but many are not.

The key parameter for swap size is "how memory-hungry things you want to run", which isn't easy to measure, but paying for installed RAM is a somewhat usable proxy metric for that. If you were happy with 8gb, it's some evidence that your apps don't need much memory (and swap), but if you needed to pay for an upgrade to 32gb, that's some evidence that you're the kind of user who needs much more swap than those with 8gb of RAM.

where do you see this goal post moving? From my perspective, it never was "The AIs will never do this." but rather even before day 1 all the experts were explicitly saying that AIs will absolutely do this, that alignment isn't solved or anything close to being solved, so any "ethical guidelines" that we can implement are just a bandaid that will hide some problematic behavior but won't really prevent this even if done to the best of our current ability.

Abstractions are inherently a tradeoff, and too much abstraction hurts you when the assumptions break.

For a major example, treating a network resource like a file is neat and elegant and simple while the network works well, however, once you have unreliable or slow or intermittent connectivity, the abstraction breaks and you have to handle the fact that it's not really like a local file, and your elegant abstraction has to be mangled with all kinds of things so that your apps are able to do that.

If you discard all major video streaming sites (including adult entertainment) then you probably can get most of the way there; you're probably mostly interested in text communication and actual user data, not the video content which is so much larger than that.

The "Keep America Beautiful" ad campaign with a Native American character (played by an Italian actor?) who's sad about the polluted environment.

This shouldn't be a major issue because of Forward Secrecy (https://en.wikipedia.org/wiki/Forward_secrecy) principles built into modern TLS protocols, which ensure that even if the public/private key scheme is vulnerable to (for example) quantum attacks, the attacks have to be done now, as a MITM for the handshake, or otherwise the full traffic capture is useless for future decryption without getting some secrets from one of the endpoints.

That being said, it's not 100% used everywhere yet (Wikipedia mentions 92.6% of websites), and various means of tricking devices into downgrading to an older protocol would result in traffic that might be decrypted later.

It's just that out of these two claims only one is interesting and worth talking about (and that's the one mentioned in the title).

Thank you for your testing! That's a bunch of effort which I didn't do - but checking the other claim is much more difficult; a refusal is clearly visible, but saying whether out of two different codebases one is systematically slightly less secure is quite tricky - so that's why people are complaining about the lack of any description of the methodology of how they measure that, without which the claims actually are not testable.

One report (https://www.theregister.com/2024/03/27/america_ad_blocker/) indicated that more than half of Americans use an ad blocker.

It's interesting that I just read an inteview with YouTube CEO (https://stratechery.com/2025/an-interview-with-youtube-ceo-n...) who mentioned that YouTube fully intends to start getting a cut out of that sponsorship money ("to align interests better").

I'd probably be OK if all the content which doesn't get made without sponsorship wouldn't get made at all, and the people who work as content creators stopped doing so. There is an overabundance of new content, having 10x less content would be perfectly fine, and in pretty much every niche there are amateur enthusiasts who clearly (based on their amount of viewers) are giving their time away, and their content is in many ways preferable and "more real" than the professionals - so I'd be OK if all the professionals stop and these awkward amateur enthusiasts are all that remain.

The same applies to web and blogs; the ability to monetize them by ads (and I do remember the "old web" before it was the case) increased the content but drowned out viewership for the true enthusiasts running things in their spare time, which IMHO were more valuable and I think that regime was better; again, losing 90% or 99% of the content wouldn't be bad in my mind, there still would be more than enough for anyone to ever "consume".

Well in your example it didn't write less secure code (wich is the core claim of the article, and something new), it refused to provide an answer about Falun Gong, which the article also claims, but that's not the interesting part of the article as censorship of certain keywords is well known DeepSeek behavior since it was released.

Technically (and legally) the USSR also had the same three branches of government; just all controlled by the same party.

All the US megacorps tend send me emails saying "We want to change TOS, here's the new TOS that's be valid from date X, and be informed that you have the right to refuse it" (in which case they'll probably terminate the service, but I'm quite sure that if it's a paid service with some subscription, they would have to refund the remaining portion) - so they can change the TOS, but not without at least some form of agreement, even if it's an implicit one 'by continuing to use the service'.

At least historically it used to be the case that you don't ever want to use the browser API directly for compatibility reasons but always through some library that will be a do-nothing-wrapper in some cases but do a bunch of weird stuff for older browsers. And traditions are sticky.

FHE solves privacy-from-compute-provider and doesn't affect any other privacy risks of the services. The trivial way to get privacy from the compute provider is to run that compute yourself - we delegate compute to cloud services for various reasonable efficiency and convenience reasons, but a 1000-fold less efficient cloud service usually isn't competitive with just getting a local device that can do that.

The parent point isn't that shitty code doesn't have defects but rather that there's usually a big gap between the code (and any defects in that code) and the actual service or product that's being provided.

Most companies have no relation between their code and their products at all - a major food conglomerate will have hundreds or thousands of IT personnel and no direct link between defects in their business process automation code (which is the #1 employment of developers) and the quality of their products.

For companies where the product does have some tech component (e.g. refrigerators mentioned above) again, I'd bet that most of that companies developers don't work on any code that's intended to be in the product, in such a company there simply is far more programming work outside of that product than inside of one. The companies making a software-first product (like startups on hackernews) where a software defect implies a product defect are an exception, not the mainstream.

If a LLM (or any other tool) makes so that team of 8 can get the same results in the same time as it used to take a team of 10 to do, then I would count that as "replaced 2 programmers" - even if there's no particular person for which the whole job has been replaced, that's not a meaningful practical difference, replacing a significant fraction of every programmer's job has the same outcomes and impacts as replacing a significant fraction of programmers.

Intent makes all the difference in the eyes of law.

If a jury of your peers decides that you configured that firmware to protect your data from thieves, then it is not a problem.

But if a jury of your peers doesn't buy your assertion of that, and instead becomes convinced that you applied that (otherwise reasonable and legitimate) feature with the intent to destroy evidence of a crime so that it doesn't reach the court, then setting it up was a felony.

Your IT system may need to handle entries linked to people who don't have names, for example, recording that some treatment was given to a baby who died soon after birth before being given a name.

I don't get why I would choose a dataclass in cases where I've already decided that an ordinary tuple would be a better fit than a normal class (i.e. "anywhere you're tempted to use a namedtuple")

To me, namedtuples are a convenience to give a nicer syntax than ordinary tuples in scenarios where I don't want the overhead of having to store a copy of all the keys with every object, like a dict would. Dataclass seems to be even more stuff on top of a class which is effectively even more stuff on top of a dict, but all the use cases of namedtuples are those where you want much less stuff than an ordinary class has. And I don't want to have to define a custom class just as I often don't define a custom namedtuple in my code but use the one the database driver generates based on the query, which is a very common use case for namedtuples as efficient temporary storage of data that then gets processed to something else.

Those farmers with second jobs would need protection from local megafarms who are outcompeting them, not other countries - it's not about location, it's about size of the operation.

The general principle is that "pipelines" impose a restriction where the errors of the first step get baked-in and can't effectively use the knowledge of the following step(s) to fix them.

So if the first step isn't near-perfect (which ASR isn't) and if there is some information or "world knowledge" in the later step(s) which is helpful in deciding that (which is true with respect to knowledge about named entities and ASR) then you can get better accuracy by having an end-to-end system where you don't attempt to pick just one best option at the system boundary. Also, joint training can be helpful, but that IMHO is less important.