d110af5ccf

the reason the magnitude doesn't matter is that those counts will be much higher in longer documents ...

To be a bit more explicit (of my intuition). The vector is encoding a ratio, isn't it? You want to treat 3:2, 6:4, 12:8, ... as equivalent in this case; normalization does exactly that.

Do you have any non-fictional examples

Probably not a good approach to take here because there are plenty of examples of everything he said. However there are simultaneously plenty of counterexamples. For example certain west coast cities where shop lifting has been de facto legalized due to non-enforcement rooted in the name of the social cause of the day.

Actually what he wrote appears to match a certain political stereotype in the US while being the opposite of the other one. So I guess it says more about his view of the US than anything else.

> at least when they want any at all. It's not consistent.

This is because you are treating Americans as a monolithic group, failing to differentiate between various major clusters.

However I will say that at least

> A lot of Americans are way too into punishment. It's seriously like a fetish for them.

matches what I see around me.

It’s difficult to burn things hot enough to reduce potentially dangerous compounds

And that's before accounting for people illegally tossing things that contain heavy metals (such as batteries).

It actively harmed the environment. Once the thing has been manufactured doing anything less than driving it into the ground is wasting the upfront environmental cost.

"rote" drilling of basic skills

I think it depends on the nature of the task for which the skill is used. But I agree. There are certain things that are just not realistic to master without rote drilling of some sort of supporting material.

At the same time I think rote drilling gets a bad reputation because poor teachers will tend to fall back on it in instances where it isn't particularly useful.

I personally had a low opinion of it until university, where I encountered two independent topics where it was clearly necessary.

Unless each spoken dialect corresponds to a matching written dialect. But that just sounds like an even messier system than modern english.

The upstream bug: https://sourceware.org/bugzilla/show_bug.cgi?id=24323

An example of a problem provided in a different bug: https://sourceware.org/bugzilla/show_bug.cgi?id=11754#c15

But I don't really understand what's going on in that example.

Aside, it only took 9 years for that patch to get conclusively rejected.

You might have something unusual going on there. At the beginning I, and to the best of my knowledge everyone I grew up around, was also taught to sound words out. At some point I stopped needing to do that. That happened quickly for the common words and more slowly for the uncommon ones. At the extreme end, for example extremely specific medical or chemical terminology, I still have to do that on account of never having encountered the word before.

At no point did I intentionally memorize word shapes. At no point in my life have I ever been conscious of recognizing words by shape. That might well be what's happening under the hood, but if so it's not something I have any awareness of. I just see a word and (in most cases) instantly know what it is without thinking about it.

Presumably your job is to get results. If catering to someone in some way is able to achieve that then it's a perfectly legitimate approach to consider.

Formatted like a formal academic publication. No way (that I can tell) to grab a pdf. Comes across as a blog masquerading as academic literature to me. Am I wrong? Did I miss something and there's an offline version available?

Pages served up over http are ephemeral. An absolutely essential part of formal academic literature is the archival aspect - self contained, immutable, and referenceable in an unambiguous manner.

There's also an immediate practical aspect for me. I will likely never get around to reading this because I will forget it exists because my "reading list" consists of a pile of pdf files.

audio/video get out of sync when seeking a lot

I have a very similar issue with subtitles in MPV sometimes. Some unknown (at least to me) combination of codec, encoding settings, subtitle format, switching subtitle tracks, and seeking will utterly break it to the point that I have to exit and reopen the file. I can't be bothered to debug the issue (so far).

In general I don't understand the negative comments I see VLC get. I've never encountered major problems with either piece of software. I've encountered minor bugs and annoyances with both.

It's not at all clear that "restarting with a small subset of cards" is better.

Sure it is. It could grow as quickly as allowed for given the time invested by the user. That could mean a return to the full subset within the span of a single day, or it could mean many months. Perhaps even never. It all depends on time invested by the user going forward. Starving regular review for the sake of verification is an example of an algorithm failing when faced with the real world.

At minimum it is clear from what was said that (better) prioritization between conflicting goals is needed. That somewhat matches my own experience with it from years ago. The algorithm was simply not flexible enough to fit my own usage patterns. In other words I was not part of the target audience, which I found frustrating because I very easily could have been.

Agreed. Explicit is only better than implicit when you have a mechanism that can reliably identify instances where you forgot to do something.

Actually for resource management I'd go so far as to say that implicit is better than explicit in the vast majority of cases. Now if only the rules that determine what happens during initialization in C++ weren't so horribly convoluted.

wouldn't it be silly to have the llm produce the weather?

Typically the LLM queries external tools as appropriate and then incorporates the result into the response.

The comment you replied to described an edge case and explained why it's broken in that particular case. You haven't actually responded to the example provided.

it's just not clear if you're going to recall them or not. Figuring that out becomes the priority

Presumably the priority ought to be (re)starting with a small subset of cards and gradually trickling the others back in. The algorithm needs to account for time spent by the given individual and adapt to changes in that over time.

I haven't used Anki for about a decade so I'm not familiar with the current state of things. At the time a major factor in my dropping it was that I found the algorithm to be more of a hindrance than a help.

Similarly, if you visit certain instances you will be greeted with the polar opposite of that. The inhabitants of some instances are more political than others. If you actually create an account you will be able to pick and choose whose posts you want to see.

The strictly chronological approach seems to work fairly well as long as you only follow a few people but it doesn't scale very well.

Information loss, or the result of useful computation? VAEs exist after all.

different comprehension of what rules mean

You're swinging right back to the context and meaning of the rules that were presented during the assigned task. What I wrote isn't really about that. It's about the assigned task itself and the self assessment of whether or not it was completed faithfully. That's where the cognitive bias becomes plainly observable.

There are the rules presented during the task. Separately there are the instructions given for the task itself. To me it feels a bit like a failure to reason with layers of abstraction. Almost an inability of most people to reason about and interpret the rules differently in different contexts. They're stuck in the "real world" context and can't seem to switch to the "hypothetical framework" context laid out in the instructions.

From the other perspective, hyper-rationality is a dysfunction ...

When obstinately adhered to in a general context, certainly. This was not a general context. It was an exercise with specific and reasonably unambiguous instructions. Openly deviating from them would be quite different than what can be observed in this comment section - deviating while claiming to have followed them.

On any other website I would be inclined to assume a certain lack of literacy or comprehension. Not so with this audience.

I think it's a fascinating practical example of how "baked in" cognitive bias is. The sort of people that use HN tend to be highly analytical. Yet nonetheless we see a massive public display of people rationalizing their failure to directly answer the question that was very clearly and unambiguously asked while pretending that they did.

The logical exercise is extremely close (by design) to one that commonly occurs in everyday life. In real life people want to bend the rules to achieve a certain outcome when applying them. They don't want to say "well a rule was violated but I'm exercising discretion". That's on full display here even though no meaningful outcome is actually being determined in this case.

No, it is not over (yet). The node is nonsense. https://ieeexplore.ieee.org/document/9150552 DOI: 10.1109/MSPEC.2020.9150552

Edit: I have no idea why anyone would downvote a link to this article. It directly answers the question with a decent level of technical detail. We are nowhere near single atom sized features yet, despite what node names might lead you to believe. There's still quite a ways to go.

There's also nothing stopping the service from remembering this hypothetical SSN equivalent (or a fingerprint of it), exchanging data with other service providers, and tracking you that way.

There are certainly cryptographic primitives that can be used to solve the privacy concerns but I have approximately zero confidence that they would even be considered for use by the current technically illiterate batch of legislators.

I'm not going to choose to "prove I'm an adult" to most of the Internet, just the sites I want to use.

Except for purchases and other business or government type interactions I flat out refuse to do this. I simply do not use web sites or services that require a real name. I only rarely use ones that require an email.

Yes, this means I don't use most commercial services and major websites. I honestly think I'm better off for it. But it's also a line in the sand and I feel much better for observing it.

HN doesn't require an email to sign up. If it did, I almost certainly would never have created an account here.

Lots of people on HN have a strong reaction against any attempt to regulate the internet, from generations of moral panics about new things.

No. That is absolutely not why and betrays precisely the lack of understanding about these things that causes such a response to begin with. Any time you find yourself dismissing widely held concerns you should take a step back and reevaluate if you haven't missed something. Perhaps your life experiences or priorities are different than these people who you encountered who had a "strong reaction"?

At least for myself, the primary reasons for such concerns are the frequency with which new regulations have historically resulted in dysfunction and abuse, proved entirely ineffective at solving the stated problem, or both.

There are major freedom of speech issues, and there are lots of other technical and policy complications

Right here you acknowledge something which should be cause for very serious concern.

Also I think I'd object to your framing of "a public health crisis comparable to opioids". Much of that appears to be attributable to the side effects of the "war on drugs". At absolute minimum the vast majority of overdoses are clearly due to unreliable product quality which is simply not an issue for things purchased legally.

So really I'd like to avoid a second "war on drugs" type scenario that seeks to solve a social ill via government intervention, ultimately fails and makes the problem far worse in the process, creates various bureaucratic machinations that are incentivized to ensure future funding for themselves rather than outcomes that are actually beneficial to the general public, and erodes civil liberties in the process.

As you ask, how many ills is it actually directly causing versus merely exposing.

As op asks, what even is social media? Does email count? An email based mailing list? A federated group chat protocol like Matrix?

Speaking of federated protocols. Email, Matrix, Activity Pub, many others are federated and self hostable. How is legislation supposed to account for that without causing more problems than it seeks to solve?

On that note, are the downstream negative effects of such legislation even being considered? If one of the outcomes is "provide copy of your passport to register a user account" would that not be a situation rife for abuse, both by governments and corporations alike? Is this a "cure" for which the combined side effects are worse than the original problem?

Counterpoint: the anti-crypto crowd frequently condescends about these topics yet fails to provide rigorous explanations of how various laws apply when pressed.

When it comes to staking, who is being lent to? It doesn't seem like a loan to me at all.

The cynics self-select themselves out quickly as there's no likes to be earned.

If these are invite-only communities do they actually self select out or do they get kicked out? My experience has generally been that small groups are often cliquish and won't tolerate significant dissent even when it's civil.

I've experienced surprisingly civil and insightful discussion in online spaces that were very nearly unmoderated. It's not for everyone and it's certainly a mixed bag but there's a lot more than just a public/private dichotomy going on in terms of social dynamics.

Even HN is a good example of a solidly public space where I've had countless constructive interactions over many years.

Note that there are both HTTPS and SVCB record types.

Regardless of the standard being finalized there's already some support in the wild. To name a few prominent ones iOS, macOS, Cloudflare, Akamai, and NS1 all support both types.

https://github.com/MikeBishop/dns-alt-svc/blob/main/svcb-imp...

Perhaps a solution such as dnscrypt-proxy and configuring it to forward requests that match certain patterns to a validating resolver on the local network would work? Requests that didn't match could be forwarded to some other resolver that wasn't configured to validate.

If you are saying that LLMs "understand" (you, or anyone else)

I was not saying that.

why is it skeptics that have to provide evidence that they don't?

Because if a claim is going to be made in either direction then evidence or other reasoning to support it should be provided. My position is that the "sensible" default position in this case is one of "we don't know". Of course defaults are always some degree of subjective in that such judgments ultimately arise from our personal worldview.

Before you object that this is an unreasonable default in this case, consider that I can't even prove that other people are sentient. I can't articulate in quantifiable terms what exactly it means for a person to understand a concept. So if I accept that the default is "not sentient until proven otherwise" then there would seem to be an issue. I would then simultaneously be saying that other _people_ aren't sentient (I can't prove they are) and that other _people_ don't understand things (I don't even know how to measure such a thing, after all).

If you can't, does that mean that atheists, like me, can't continue to say very confidently indeed that he doesn't?

I'm not quite sure I follow. Indeed I hold that strong claims that God either does or does not exist are unreasonable on the basis that both would appear to be fundamentally untestable claims. At least, untestable short of such an entity choosing to provide incontrovertible proof of its existence.