HNHacker News
TopNewBestAskShowJobs

mhink

1,169 karma · joined August 4, 2011

[ my public key: https://keybase.io/mhink; my proof: https://keybase.io/mhink/sigs/ZGyLxMDpR82SpEWmF_6OjVUuT4LhvftzJPuAnZpSPlQ ]
submissionscomments
mhink··on 418 I’m a teapot
This always reminds me of a great find I discovered a while back when I was reading the HTTP/2 RFC for whatever reason.

Back before "429 Too Many Requests" was standardized, Twitter's API used to return a nonstandard 420 status code (with the caption "Enhance Your Calm") when rate-limiting incoming requests. They stopped doing this for understandable reasons, but the caption, at least, got snuck into HTTP/2, see:

https://datatracker.ietf.org/doc/html/rfc7540#section-7

down at 0xb, the caption for disconnection due to excessive load is, indeed, ENHANCE_YOUR_CALM. Always makes me smile. :)

mhink··on Show HN: Dead man's switch without reliance on your infra
> Basically a technically competent will executor.

I don't think OP is saying that an attorney *can't* get these things done, but that it would make them feel more comfortable knowing that a technically competent person and/or service will be performing the actual actions.

I do think there's a place for an attorney here, in the sense that they could be the trusted individual responsible for notifying DeadManService, Inc. that a particular person has, indeed, passed on and wishes DeadManService to run their instructions.

mhink··on Helping wikis move away from Fandom
> One thing I find interesting about playing video games in modern day is that with the proliferation of Wikis, there is assumed to be some kind of third party guide for every game. Especially in smaller/newer games it seems like developers sometimes don't bother putting necessary information in the game at all because they don't have the person-hours for it.

While this may have become more of a norm in recent years, online communities with community-supported guides have definitely been around since before wikis were common in the gaming community: most notably at gamefaqs.com. To this day you can still find plaintext walkthroughs for thousands of games, written 25 years ago by pseudonymous authors.

Which isn't exactly to dispute your point, just waxing nostalgic about the good ol' days. The RPG Maker 2000 forum was basically my introduction to programming, waaay back in the day.

mhink··on Show HN: AI-generated images that look like real life
> The dog in a park, the collar is obviously off.

Interesting note: my wife works with dogs, and I just took that photo over and asked her what breed she thought it was. Her first take was "it really just looks like an old mutt, I can't tell the marking colors. Maybe some Rottweiler in there?"

When I pointed out the collar she took a closer look, and then pointed out "oh, and look at its legs! They're way too short for its body!" Which, in retrospect, is the most hilarious error in the picture, hah.

mhink··on iPhone 16 Pro and iPhone 16 Pro Max
For AirDrop, at least, I've found it to be very reliable, with one caveat that always tripped me up before I realized it: *the recipient's phone needs to be unlocked*. This completely changed the game for me, and now me and my wife use it all the time.
mhink··on Can kids master the video games their parents loved?
Honestly... I don't think there's much difference between games and other forms of media in terms of longevity in the public consciousness. "Singin' in the Rain" was released in 1952 and might very well still at least ring a bell today, but what about "Monkey Business" or "Viva Zapata!", released the same year? Those films even had some pretty big names (Marilyn Monroe and Marlon Brando). Even nowadays, there are still plenty of "shovelware" movies out there.

That being said, I think the real difference comes down to the fact that games are (by necessity) relatively tightly coupled to the hardware required to run them. Movie projector systems, while complex, are still much "dumber" than a computer. I think it's safe to say that if you have a film print from, say, the 60s, there would probably be a range of projector models from the preceding decade that could play it.

By contrast, if you have a copy of a PC game from, say, 1996, you're going to have to pay a lot more attention to its specific hardware requirements in order to build a system that's capable of running it. That raises the skill floor to run it in the first place. Honestly, I think this is one reason console retrogaming is so popular: to play a game, you only have to get a hold of a working console. The problem these days is that consoles are becoming more PC-like, so they're kinda backing into the same problems as PCs.

mhink··on Turing's topological proof that every written alphabet is finite (2010)
That link is absolutely fascinating, and would definitely make for a great sci-fi concept!
mhink··on Timezone-naive datetimes are one of the most dangerous objects in Python
The trouble is when you have to deal with timezone-aware business logic. `date-fns-tz` does exist, but the amount of brain contortions required to keep track of everything is frankly exhausting, not to mention error-prone.

Suffice it to say that I'm waiting with bated breath on the Temporal API (https://github.com/tc39/proposal-temporal).

mhink··on Pi calculation world record with over 202T digits
I got the impression that the comment was a bit tongue-in-cheek.

The joke lies in the fact that saying "100% of real numbers" isn't *technically* the same thing as saying "all real numbers", because there's not really a good way to define a meaning for "100%" that lets you exclude rational numbers (or any other countable subset of the reals) and get something other than 100%.

mhink··on Deactivating an API, one step at a time
I was thinking about this kind of thing. Another idea might be to introduce artificial latency and gradually increase it over time? Maybe dial up rate-limiting? I'm not really sure if this is a better or worse idea, though.
mhink··on The Naughty Words the FAA Removed from the Sky
> If I'm reading correctly the Florida waypoints referenced were renamed because a busybody NY Times reporter complained?

Do you suppose that if a so-called busybody complained about a waypoint the FAA didn’t care about, they’d bother to change it?

It was $11k. You could probably pay that with the change in your couch.

mhink··on What's the difference between a motor and an engine? (2013)
While I think I agree with your definition of "engine", I'm not sure I agree with the idea of "motor" being a subtype of "engine".

To me, a "motor" is a specific part of a system's powertrain: it's the part which converts some other type of energy into kinetic energy for the purposes of moving the system. On the other hand, an "engine" is a self-contained system that establishes some kind of stable feedback cycle to perform a task.

With that in mind, a car's internal combustion engine is the motor of the car, since that's the place where chemical energy becomes kinetic energy, which results in motion after being transferred through the drivetrain.

If you take the same exact engine out of the car, mount it in some fixed position, and rig it up to generate electricity, I wouldn't technically call it a "motor" anymore, because it's not acting to move the system it's a part of. (That being said, I probably wouldn't split hairs here, since the engine was originally designed to be a motor. But you see my point.)

Importantly, this lets us fit diesel trains into our definition: they use diesel engines to generate electricity, but electric motors to actually make the train move. In an old steam-driven train,

It also addresses the question of what to call the part of a model rocket that makes it go up. It's the part of the rocket's powertrain which converts chemical energy into kinetic energy that moves the system. It's a *motor*.

---

So, with all that said, an interesting point is that the terms "motor engine" and "engine motor" both make sense, but don't mean the same thing. "Motor engine" describes an engine designed to perform the role of motor (as opposed to a "generator engine"), whereas "engine motor" describes a specific kind of motor (as opposed to a "chemical motor").

mhink··on It’s an age of marvels
I would agree if we were talking about, say, a team in 1970, but it's worth pointing out that we're talking about a team working during the very infancy of computer engineering (the late 40s/early 50s). Just off the top of my head, you had solid-state hardware, stored-program machines, and the Von Neumann architecture all just popping on to the scene. This is stuff that would definitely have been on their radar, but I think they'd be fascinated to realize how directly and fundamentally those inventions would affect computing.

Perhaps more importantly, I think they would be blown away by how directly and fundamentally computing would change the world. Doug Englebart and his team are the ones who really developed the idea of using computers for something other than performing calculations for scientists. They started that research in the early 60s, and didn't drop the Mother Of All Demos until 1968. So if I were one of the "ENIAC guys", and someone asked me how my work would affect the world, I'd probably just shrug and suggest that the computers I was building would help perform computations that would let other scientists make discoveries more quickly.

mhink··on Apple apologizes for iPad 'Crush' ad that 'missed the mark'
Yeah, I got these vibes too, but especially from the cinematography. It was kinda _lingering_ on all the little ways each item strained and broke under the weight of the press, almost like it was something to savor. It was weirdly voyeuristic, in a way.
mhink··on 'Underwater bicycle' propels swimmers forward at superhuman speed
I don't think this argument holds water (pun intended), though, since this device is completely human-powered.
mhink··on Double-entry bookkeeping as a directed graph
> No. This transaction does not increase liability in any absolute sense. It increases liability only from the bank's perspective. From your perspective, it increases assets.

What you're missing here is that if I were to keep my own records, I would also list two accounts. When I look at my bank statement online, what I'm looking at is the bank's records. If I kept my own books, it would be the same thing, but in reverse, like this:

    MHINK'S RECORDS
    - CREDIT mhink's cash account $100 (decreasing mhink's assets)
    - DEBIT mhink's account for the bank $100 (increasing mhink's assets)

    BANK'S RECORDS
    - CREDIT the bank's account for mhink $100 (increasing the bank's liability)
    - DEBIT bank's cash account $100 (increasing the bank's assets)
> It has borrowed that cash from you. So that cash is both an asset (because having borrowed it from you it can turn around and loan it to someone else) and a liability (because the bank is in debt to you for the amount of the deposit).

The cash itself isn't both things- it's only ever an asset.

When I transfer cash from my wallet to the bank, I'm converting $100 of value from one type of an asset to another: from cash, to "a debt I'm owed by the bank".

The bank also didn't change its total position. It took on "a debt owed to mhink" (a liability), but gained "cash" (an asset).

mhink··on Interview with Senior JavaScript Developer 2024 [video]
And the complaints hold less and less water each year.

In 2015, people were still complaining about trying to build stable layouts using CSS. Congratulations, we now have Grid and Flexbox (and subgrids just stabilized!) but they still complain. They complain about having to write custom Javascript (which is slow and has to be maintained) so functionality gets pushed into CSS and they complain that CSS is getting too complicated.

There were people complaining about the "Javascript framework of the week". Then React took over the space, and now they complain about how everything's written for React and we should just use htmx and Svelte. I thought we didn't want something new every week!

There were people complaining that Webpack was slow and required too much up-front configuration. Then Vite came out (which is fast and only requires minimal configuration) and they went back to complaining about having to learn new things.

The best explanation I can come up with is that most organizations must just treat FE as an afterthought and assign people to do that work who have no experience or interest in it.

mhink··on Double-entry bookkeeping as a directed graph
I've seen it used frequently in the context of data pipelines and stream-processing applications, which is even more applicable to the subject.
mhink··on Double-entry bookkeeping as a directed graph
That's not entirely correct- or at least, it's more complicated than that. The question of whether a debit/credit increases/decreases an account has to do with the kind of account you're talking about.

When I deposit money, it modifies two accounts at the bank:

- the account which represents how much money they owe me - and the account which represents how much money they have on hand.

The former is a liability, and the latter is an asset.

The meaning of debit/credit is reversed between these two types of account. So, when I deposit $100, the entries entered are:

    - CREDIT mhink's account $100 (increasing liability)
    - DEBIT cash account $100 (increasing asset)
Since we only see one side of this, we start to associate "debit" with "less money for me" and "credit" as "more money for me".

Oddly enough, another common financial situation reinforces this interpretation from the other direction: accounts with utility providers. Unlike the bank, your account at the utility company represents how much you owe them. So the meaning of debit/credit is reversed, but so is the direction of responsibility: your account at the utility provider is money you owe them, which is an asset. So when I pay them $100, the entries entered are:

    - CREDIT  mhink's account $100 (decreasing asset)
    - DEBIT   cash on hand $100 (increasing asset)
mhink··on Is It Dry Yet?
For what it's worth, my understanding is that it's generally not recommended to wash/dry different kinds of fabric together.
mhink··on A response to "Erlang – overhyped or underestimated" (2010)
> Some years later, when I finally got around to learning and using Node.js, I realized just how poor the async/await/promise design Node.js was (error-prone by design). But I highly doubt I would convince anyone to really give Elixir or Erlang a go.

IIRC Elixir got a pretty big boost early on thanks to folks making the jump from Rails to Phoenix, so it could kinda be the case that most of the folks who *would* give it a shot *have already* given it a shot.

That being said, I have the impression that you'd have more success with folks using Node.js than you might think. I expect there would be more pushback from folks using, say, Golang and Java... if only because the wider dev community doesn't look down its collective nose at those languages quite so much. (Unless you're talking about generics or null values, respectively!)

mhink··on Memory and new controls for ChatGPT
Looks like you'll be able to turn the feature off:

> You can turn off memory at any time (Settings > Personalization > Memory). While memory is off, you won't create or use memories.

mhink··on How to center a div in CSS
They *have* come together. That's why we actually have several options built for purpose and native to CSS these days, as opposed to relying on hacks and Javascript (which was surprisingly common even as recently as 5-6 years ago. The name of this article is more or less a joke, because most people with any experience will recognize that since Flexbox and Grid have become widely available, the answer is generally "use Flexbox or Grid". Knowing the nuances of these layout algorithms is table stakes for building UIs that don't feel "janky".

The problem with designing and implementing application layouts for the Web is that it's such a dynamic medium. You can't ever rely on a particular viewport size, and you generally can't rely on content size either.

Like, just as a simple example: build me a page which has a white background and three blue boxes in the middle which contain white text saying "foo", "bar", and "Supercalifragilisticexpialidocious".

How big should these boxes be? Should they even be the same size? How should text wrap within them? How should the text itself be aligned within the boxes? When you say the "middle", do you mean centered vertically or horizontally? Are they laid out in a row or in a column? What's the expected behavior when the viewport size is too small to accommodate them? Should the boxes themselves wrap in some special way? Should they resize themselves?

This isn't even really splitting hairs or anything, it's just sort of the mindset you get into when you start working within a domain that's governed by *constraints* rather than specific sizes.

mhink··on “Wherever you get your podcasts” is a radical statement
One of the most surprising things (from a modern perspective) about reading that daringfireball article is that a game could be successful on the App Store at a $5 price point ($7 in 2024 dollars).
mhink··on The Apple Vision Pro
If I'm being completely honest, I don't think the AVP itself needs to be as open as a Mac, but I'd absolutely expect it to be able to act as a window manager for a paired Mac. It's really, really surprising that it only comes with the ability to project an entire Mac screen into VR instead of letting the user break out individual windows within the space.
mhink··on The NSA Furby Documents
> I almost hate to say it because it's antithetical to the Internet and Hacker News ethos, but it's a testament to how well networked applications could work with a central authority and no anonymity. You don't need passwords. Accounts are provisioned automatically. SSO is global to the entire network. You only need one identity. But no, your office can't have Alexa.

I don't think it's necessarily a dealbreaker if you consider this: from a purely technical standpoint, there's nothing really stopping anyone from setting up a certificate authority- the only issue is getting service providers to trust it enough to accept those client certs as sufficient identification. I could easily imagine a world where I receive an "official" client cert from a government (which I can use to thoroughly prove my identity if needed) as well as several "pseudonymous" certs from various other CAs that I may use from time to time.

The main difference between CAs would be the kind of attestations they provide for a given certificate holder. For example, I could imagine a CA which (for example) is set up to attest that any holder of a certificate signed by them is a medical doctor, but will not (by policy) divulge any additional information.

Or perhaps a CA which acts as a judge of good character- they may issue pseudonymous or anonymous certs, but provide a way for application owners to complain about the behavior of a user presenting that cert.

I'm sure there are plenty of holes that can be poked in this model but I don't think it'd be completely out of the question?

mhink··on Is htmx Just Another JavaScript Framework?
For what it's worth, the landscape of Web development was much, much different in 2014 or so. Ruby on Rails was still a huge player, and most major OSS projects meant to enable Web development took a lot of cues from it.

Projects like Angular, Ember.JS, and Backbone (and for that matter, Rails' own Turbolinks system) very much expected you to build your entire application in them from the get-go.

The gist of React saying "this is a view library" at that point was largely about what you should expect React not to do for you. For example, two-way data binding, where the framework managed a fetch lifecycle as well as plugging the result into a view. "React is a view library" was letting devs know that they shouldn't expect React to do this- they should expect to make calls themselves and re-render the view accordingly.

I actually think it's still fair to say that React *itself* is still "just a view library", but that's sorta like saying Linux is "just a kernel". Sure, it's technically true, but at this point when we're talking about React we're talking about the whole ecosystem surrounding it as well. That wasn't the case 10 years ago (damn, has it really been that long?)

mhink··on Git Branches Have No Rules
Is this on Windows? Or maybe you're on an older version of Git? I make branches with slashes in the name all the time, never had a problem with it.
mhink··on Bun, JavaScript, and TCO
> Look, I’m all for being clever when cleverness is needed but some of you JS/TS folks take it to the extreme.

To be fair, this is very much a toy example meant to demonstrate tail recursion- although I do agree it's not a great example considering you need to understand the evaluation order of ternaries in order to understand why one version is tail-recursive and the other isn't.

That being said: generally speaking, one might argue that using recursion at all is "too clever" and algorithms should just stick to iteration altogether. ¯\_(ツ)_/¯

> While you can nest a recursive function in a tertiary statement coro, don’t.

Man, why do you gotta call me out like this? :D

mhink··on Mapping the semantic void: Strange goings-on in GPT embedding spaces
> Since there are ~5000 dimensions, in which of those dimensions are we moving k out ?

> Is the idea you just move, out, in all dimensions such that the final Euclidean distance is k ?

If I understand correctly, the basic idea is that in earlier experiments, they were able to use a relatively-simple technique to come up with what they call a "probe vector" in the embedding space which represents "the property of a word starting with <letter>". For any given token, the authors established that it was 98% probable that the token's embedding vector would be closer (by cosine similarity) to the "probe vector" representing the first letter of that word than any other probe vector. This is shown in the first graph of the section "A puzzling discovery".

With that in mind, the diagram below that should start making more sense: "emb" is a particular token's embedding vector and "probe" is the probe vector for the token's first letter. "emb_proj" is the projection of "emb" onto "probe".

What they're doing is tweaking the network weights by subtracting multiples of `emb_proj` from `emb` (where the specific multiple is the parameter K), and then seeing how it behaves differently for different values of K.

Their original observation when doing this was that it reliably caused the model to claim that the first letter of the tweaked word was not the letter in question. In this article, they're trying to figure out how far they can push the tweak and still get reasonably accurate definitions of a token.

What they discovered is that when they push a token's embedding vector further and further out along its "first-letter vector" and ask the network to define that word, the definitions it provides seem to follow particular themes during different regimes of K.

← PreviousPage 2 of 11Next →