drdaeman

Excuse me… what? You must've misunderstood what I argued about. Yes, they’re not the same!

I'm going to reiterate for the third time, to break it down even further in a hope it clicks.

We have: 1. Predatory/addictive algorithms (A_bad). We don’t like those. 2. Old good user preferences-based algorithms (A_good). Those are fine. 3. Probably some other algorithms but let’s go with just two alternatives for simplicity’s sake, as extra (A_more) won’t affect the argument here.

Definition from the law (A_law) covers both (A_bad)+(A_good). As I already said twice, it’s overbroad.

You said “if (A_law) wouldn’t be bad there wouldn’t be an issue”. That is obviously true, because (A_law) contains (A_bad). What you said is true, and it doesn’t contradict what I said - that (A_law) containing (A_good) in addition to (A_bad) isn’t great and lawmakers should’ve done better.

Just so we’re clear, I’m not arguing that A_law = A_good (that’s not true, they’re not the same! but neither A_law = A_bad either, it's the fact that it's A+bad+A_good[+A_more] is the issue), or that A_bad isn’t bad, or that A_bad isn’t covered by the law, or anything like that. Your knowledge that A_bad is bad, is my knowledge too.

As for the second point, I provided a counterexample of how endless scroll can be a non-issue. If endless scroll can be a problem (no argument against that from me - yes, there are clear examples we all have seen) and endless scroll can be non-problematic too (my counterexample design), that, says endless scroll is not what’s problematic per se. That’s just simple propositional logic.

Endless scrolling can be a proxy/substitute for the actual issue if nothing better could be named, but I think a better target can be identified and pointed out.

In other words, all I’m unhappy about is that laws are poorly worded and loosely defined so they can potentially cause unnecessary collateral damage. All I’m saying is that lawmakers are missing the point and just vibe-legislate in the vicinity. They’re banning the thing they (and you, and I) see, but the way they do it kinda like how Plato defined what a man is. And I’m just plucking a chicken here.

If you see a logical flaw - please do point it out for me. I don’t see any.

Recommender algorithms as defined aren’t just the recommender algorithms you’re thinking about. It’s merely this idea that precise definition matter, nothing more. I’m not arguing that it was a bad idea to restrict some crappy patterns, or anything like that. Just that lawmakers defined that poorly.

And endless scrolling is pagination with a page size of one and a gesture (scroll down or swipe) for the next page. And trust me, “next page” can be helluva addictive. Endless pagination is an aesthetic preference. It may make it more slicker, frictionless, more visually appealing to get addicted, sure, but it doesn’t change a thing about the core element of the addiction. Same endless feed can be easily made counter-adductive by degrading relevance past some threshold, which clearly tells me the issue is elsewhere.

> The thing which measures your preferences and usage patterns then decides what to show you

That (plus a clause that exempts algorithms with user controls with full informedness and ability to confidently and reliably predict the outcome = old good filters and such) is the thing that should be in the legislation’s definitions section.

But then in the actual law it’s worse than “algorithm”, they’re actually banning broad over generalizations and design patterns instead, stuff like this:

>> recommender features – content selected and displayed based on information associated with the user’s account;

(Obvious legit use cases are regional content and user preferences.)

>> endless-feed features – continuous feeds of end-user content;

(Pure design pattern, pagination can be addictive at any span size.)

There is such a thing, or, rather, used to be. Problem is that security is expensive (essentially one needs to examine all possible states of the system), and it inevitably failed to keep up with the crazy growth of complexity of modern computer systems. It became impossible to maintain a model of a system with myriad of moving parts, so it became impossible to make behavior guarantees.

Remove the complexity (all the way down to the hardware quirks), and security will be doable again.

Cloud GPUs have ToS about purposes you’re allowed to crank numbers for?

I thought this only applies to LLM inference providers, but not raw GPU rentals.

Those tokens aren’t guaranteed (esp. with RE and security tasks - rooted my own TV last week, Claude crapped out on “cyber safety” grounds; but also no guarantees about the model served - providers can pull a switcheroo on weights or quantization at any moment, and new options may not work for you), and you’re throwing money at entities that aren’t aligned with your interests instead of entities who are interested in actually empowering you.

Except that there’s no such thing as a “normie”.

Everyone - save for people with developmental difficulties or ahedonia - has a specialty, something they know that the majority of other folks don’t. Every such domain of knowledge (useful or otherwise) has its shorthands and jargonisms.

Some domains of knowledge lag behind their technological adoption. Which is natural and fine, except that society at large got infected with this silly idea of normalizing mindless consumption with a shallow (if any) underlying mental model. In other words, the nerds are actively told to shut up and keep their mumbo jumbo talk to themselves and spare people from knowing how things work under the hood.

So it’s not just that those silly nerds bad at communicating. That too, but it’s meeting with the unwillingness to learn is what makes it truly bad. It’s in the very idea of a “normie” - an anti-knowledge, ignorance-normalizing, curiosity-dampening attitude - that’s to blame. Not that nerds use nerdy words on nerdy websites and that usage spreads to less nerdy areas - that’s part is just life as usual.

BTW, those “loud idiots” are the actual normal people learning and making mistakes (including the mistake of overconfidence - especially easy to make when something feels important) and having opinions. Modern mass communications changed the world’s landscape, so instead of a few peers in their kitchen they’re broadcasting their knowledge (including the errors in it) to orders of magnitude more people and that sometimes gains traction - but that’s a separate issue, very far from nerds and their words.

> Rule Number One: You may not use a single word an LLM suggests to you.

I think there's a use case where this rule can be ignored. Sometimes I ask LLM to help me pick a single word or short phrase in some language, that has specific meaning, and connotations. Or doesn't carry some connotations or associations. Especially when I have an idiom from some language in mind, but I'm writing in some other language and my brain struggles to find a good replacement.

LLMs are super handy as fancy vibe-based dictionary search engines. They're all about words that go with other words, so when I'm struggling for a good term, I see no harm in asking and picking an option that I feel closest to what I want to express. My choice, my voice, just a very fancy thesaurus machine.

What's the point of complaining about use of jargonisms in a specialist-leaning spaces?

Anyone with a resemblance of a functioning brain (that is, any "normie" included) can understand "ability to install and run custom aftermarket operating system, or modify pre-installed one", assuming they know what an "operating system" is. And if they don't know what it is, that can - and should (!) - be explained too.

This ability matters to people who don't know any much about any computer programming, because it affects their ability to use the devices they own. They may not know how it all works, but they do express regrets when something doesn't work the way they expect or want it to work. And we all know and hate enshittification here, don't we? Personally, I'm tired of saying "no, because Apple/Google decided only they can control this, and neither you or me can change this on this device" to "can this work so-and-so" questions.

And "open bootloader" is a jargon shorthand for that. Crude, because it should be actually more like "bootloader that can be configured to validate software integrity with user-provided key material", but most people familiar with the space and lingo get the right idea anyway, and shorthand needs to be short. And this is Hacker News, not a Wendy's, so jargon is more than expected.

If someone (a "normie") reads something and don't understand it, we should all be encouraging them to ask questions and try to give them accessible answers. Not shield them from this knowledge, not fetishize mindless consumerism, not behave like snobs, and especially not dismiss something important only because it's not commonly understood.

For marketing materials that target "normie" audiences, it should be just phrased in "normie"-accessible way (look how Valve does that with their hardware), but when we're compiling a HN wishlist from HN comments, complaining about jargon is silly.

/rant

It’s an interesting book for sure, but I had (minor, but still) issues even with the introductory premises. Just a few notes I made from the very first chapter:

- Jaynes used the term “reminiscence” in a way very different from Ballard and claimed no animal is capable of reminiscence. That was a very odd claim overall.

- Then he claimed that if we ever create a language that can express everything, metaphors won’t be possible. Which is, I think, almost obviously false, because many metaphors are constructed artistically, despite having exact terms.

- And he talks about spatialization as a primitive aspect of consciousness, but he seems to talk about relations/connectedness at best.

And so on - I misplaced my notes somewhere and only recovered a few random bits. Still, interesting reading, fun (albeit really questionable) core idea, but those papercuts were kinda off-putting for me.

Yes, of course.

My point was that computers are as secure as human(s) who programmed them were careful and competent. Computer security is ultimately human knowledge and reasoning competence (plus time/money tradeoffs, if made willingly)

Human security. Computers are fine, they usually do exactly as they’re programmed.

Yes. That’s “is used to cause harm” situation, exactly.

Isn’t the whole hope is that Google (and other app vendors) doesn’t even get a say at what’s possible through Siri.

All it takes is to expose application’s UI to the model, and let the model send actions. If a human can do it tapping on the screen, the machine should be able to do the same thing.

Sometimes that lead ends up in some school kids instead of enemy combatants or other plausible threats to life.

> In the physical world, it seems like when an tool/device/instrument causes harm (or is used to cause harm), we assign blame to either the user of the tool or its creator.

Firearms are a notorious example where some people get, well, weird.

I like the overall direction, but how to tax it without screwing over hobbyist archivists aka data hoarders? That is, how to assess the value of some datastore, somehow distinguishing between personal preference profiles, random LLM software slop and, say, vintage anime collections?

We didn’t have laws but every TV at the dawn of the era came with full schematics and documentation.

That said, I’m not arguing that we don’t need laws today - we do, or at least I have no other idea how to do a course correction from where we are. Preferably wide-range ones that alter the whole consumer-producer relationships.

But the key thing is not the laws, it’s the attitude (individual and corporate) towards others. Can’t legislate that, I’m afraid.

Which ones?

Aren’t we all living with an illusion of intentionality of our actions, and the uncanniness of a machine is because it doesn’t pretend?

Meaning is listener’s own perception (author’s own intent is a suggestion at best - if I’d be paid a dollar every time I thought some artist meant something when they didn’t or meant something different, I’d sure have a good wad of cash), and when a listener learns about the purely mechanistic “authorship” they start to feel uncanny unless they find an outing in appreciating the mechanicity itself.

Two completely different stories. One is public data scraping, another is private conversation scraping (where they're a first-party to the conversation). The key difference is that in the former case, no one made any promises, in the latter an explicit promise was made that data is not used for training (assuming opt-out).

Isn't modulo-twelve (if we ignore microtonal) the core underlying structure, and then you pick scale (like 2 2 1 2 2 2 1) and a key as a mapping, and then use degree as an 1-based index into the scale?

I don't really know a thing about music, so I could be wrong here - but that mental model at least made some sense to me when I tried to map all the concepts I've heard about into some sort of a mental system.

Same exact duration for all authors: as long as they’re around and capable of caring. It’s not the same when duration is expressed numerically, sure. But consider: no life is the same in this way (until immortality is discovered), yet all should be valued equally (or so I hope and believe). Timing life makes them unavoidably unequal - and thus differences in lifetime shouldn’t be a criteria to use for comparing rights. That’d be actual age discrimination.

I suggest to rather look at it under the angle whether estates should be able to inherit author’s identity (or some of it), to continue the will.

> The framing of something as a disease

Is a well-known cognitive quirk, Sapolsky covered that in “Behave” and “Determined”. As I get it, our morality piggybacks on the sense of disgust from insula activation, which lights up the same way as with pathogen avoidance. So comparing social groups to rodents, insects, or disease is a popular propaganda trick. Same here with viruses.

> a computer can never be sad or horny or have taste

So far ;-)

What we need are refinement types, where there’s a base type and a predicate. F* has this:

     val (/) : int -> (divisor:int { divisor <> 0 }) -> int

I’m not sure what kind of peripherals the article talks about, but I’ve had a lot of weird crashes with ASUS MB onboard feature that provides RGB lightning to various fans and whatever. I may misremember the details; it was 6 or so years ago when I encountered that, but their AsIO driver was an abomination of engineering. Not only was it a god driver that allowed users to mess with arbitrary memory and MSRs from userspace, it managed to somehow mess up SMBus and cause random lockups.

Guess there’s plenty of other hardware companies that aren’t producing better software.

“How” as in “what tool” is a different “how” as in “with what intent”. LLM was used to produce a bunch of arguments supporting a pre-intended outcome. Doesn’t matter much whether it was a machine or a human (and whether it was “expert” himself or a ghostwriter), but it does matter - a lot - how the outcome was the pre-selected and the task was to cherry-pick the facts to support it.

Yea, my bad, that was a brain fart. I guess I was thinking about it from the thing’s perspective rather than owners’.

Yes. Although, to be pedantic - stealing relocates (it doesn’t leave one less, it moves the only thing into another person’s possession), while copying duplicates.

Copy vs move is IMHO accurate semantics.