electrograv

IMO a UI library landing page should always contain a screenshot example of the UI.

I can’t find a screenshot of it anywhere, let alone the landing page.

While the mental image of eating roadkill is also unappetizing to me, I have to admit my reaction here is irrational.

Eating roadkill isn’t much different from eating wild game you hunted — except with roadkill, it was someone else and their car that killed it accidentally, rather you and a gun intentionally.

So if the disk isn't alive, the file on it isn't alive, the inference software is not alive - then what are you saying is alive and thinking?

“So if the severed head isn’t alive, the disembodied heart isn’t alive, the jar of blood we drained out isn’t alive - then what are you saying is alive and thinking?”

- Some silicon alien life forms somewhere debating whether the human life form they just disassembled could ever be alive and thinking

In 2025, after a £12m investment, YASA opened the UK's first axial-flux super factory, in Oxfordshire.

It’s a little sad to me that fundamental innovations in electromechanical engineering like this get just a few million in investment, yet if this had been yet another derivative software startup with “AI” in the pitch, they’d probably have 10x+ or more investments being thrown at them.

Yes, but with the “revolving door” between private financial institutions and government financial policy/regulation, there’s little real distinction anymore between the two.

Those private banks can print that money out of thin air because government allows them to. And the government officials (many formerly financial executives) allow them to because they “have to” to prevent “disastrous” private banking/financial collapse.

But if you or I wanted to play the same games to print our own money they way they do? No, that would be wrong and dangerous and illegal!

So it’s pretty clear that both government and private financial institutions are tightly coupled partners in a mostly corrupt, intentionally obfuscated shell game that primarily serves to keep money and power steadily flowing into the hands of the already wealthy and powerful.

Just look at who is actually held accountable for financial crimes. Some individual trader that finds and exploits some glitch that allows them to profit from the wealthy? Straight to jail. High ranking institutional powers (government and private) that implement often illegal schemes that continuously siphon wealth from common people into their hands? Slap on the wrist at most.

I had no idea Keynes had similar ideas, so I definitely should read his work (and economics literature in general).

I probably should generalize my thoughts though to say “expectation of economic growth” (instead of just “money printing”) seems to me necessary to yield “opaque market insanity”, as opposed to “transparent evil sanity”.

As a thought experiment, consider a (practically impossible) scenario where there is universally no expectation for long-term economic growth/contraction — regardless of whether it’s “real” or just monetary. Then by definition, a long term market simply cannot exist at all. No amount of wealth inequality can cause market insanity if there is no (long-term) market at all.

Wealth inequality in such a situation can still yield hoarding, domination, conquest, control, scams, manipulations, etc. But I wouldn’t call that “market insanity” so much as “evil sanity”.

In practice, the real impact of wealth inequality on the common people would likely be the same either way. However, without long term economic growth/inflation, the “sane evil” of the greedy wealth can no longer hide behind the veil of “market insanity”.

The thing is, you can simultaneously be completely correct about the market being insane, while also entirely wrong in expecting it to behave in a sane way.

Cue the famous quote: “The market can remain irrational longer than you can remain solvent.”

I have a vague theory that as the amount of wealth inequality increases in a system along with “money printing” (lending, hypothecation, etc where the wealthy are permitted privileged leverage and risk), the more detached markets become from reality in general. In such a case, an increasing majority of the money circulating has no need to be grounded in anything close to the common basic needs and values that most normal people have to live with.

Instead, most important to such wealth is to tap into the source of inflation to be on the winning side of that. This becomes a game of its own, where an investment’s connection to reality or fundamental value is mostly irrelevant compared to how it leverages or monopolizes the state-created and privately created instruments of “money printing” (sketchy lending, rehypothecation, etc.) and other such “games” that only the wealthy are allowed in on.

Intelligence and consciousness are two different things though, and some would argue they may even be almost completely orthogonal. (A great science fiction book called Blindsight by Peter Watts explores this concept in some detail BTW, it’s a great read.)

That's roughly how it works in C, and I know that it's also UB there if you do it wrong, but one thing is different: It doesn't really ever occupy my mind as a problem. In Rust it does.

UB doesn’t occupy the author’s mind when writing C, when it really should. This kind of lazy attitude to memory safety is precisely why so much C code is notoriously riddled with memory bugs and security vulnerabilities.

That’s why I specifically emphasized “perception and preferences”. Believe it or not, the science covers both - both what people can perceive, and what people care about and value.

Then why do people keep buying random laptops and then complaining when Linux doesn't run on it? You buy a laptop from a vendor who designs them to run Linux out of the box.

Because:

(1) Laptop models designed to run Linux out of the box are very scarce, with very few options to choose from.

(2) Of the few that do exist, I’ve never seen any even remotely close to being competitive with Apple’s laptops (in terms of hardware quality, and good performance with excellent power efficiency / fanless / thermals / battery life).

Part of that is due to Apple’s monopoly on the superiority of their M series chips. But the rest I assume comes from less R&D investment generally in the Linux laptop space due to it being such a small niche, unfortunately.

I would also argue the crowd that insists everyone needs Pro Motion is doing exactly what you accuse me of -- assuming their needs and perception must also be everyone else's.

I am not seeing this alleged crowd of people insisting that everyone needs 120hz/ProMotion. This seems to be a red herring.

I am seeing a crowd of people (including myself) saying that we experience 120hz/ProMotion as a huge improvement over 60hz, so much so that we will never buy a product without this ever again (so long as we have the choice).

I furthermore claim that while not everyone is a member of this crowd (obviously), it represents a sufficiently large share of the device-buying population to justify steering billions of dollars of hardware and software industry to support this, which evidently has happened and increasingly continues to happen.

If this crowd were an insignificant minority as you seem to imply, then 120hz displays would be a fad that fades away in all but the most niche markets (e.g. pro gaming), and yet we’re seeing precisely the opposite happen — 120hz displays are growing in popularity by expanding broadly into increasingly non-niche consumer device products everywhere, from laptops to tablets to phones.

When clearly the market has said otherwise, given Apple's success for many, many years with 60Hz screens.

Arguing that the market doesn’t want/need it now because Apple succeeded without it in the past, is completely absurd — just as nonsensical as trying to argue that computers don’t ever need any more memory because they sold just fine with less in the past.

You have to understand that your own perceptual experience is not identical to that of all other humans. Without recognizing that, we will inevitably end up talking past each other endlessly and writing each other off as { hallucinating, lying, exaggerating, etc } for one of us claiming to perceive something important that the other does not.

It would be no different than arguing about whether we need all three primary colors (red, green, blue) with someone who is colorblind (and unaware of this). Or like arguing whether speakers benefit from being able to reproduce a certain frequency, with someone who is partially or fully deaf at that frequency. And I truly mean no disrespect to anyone with different perception abilities in these or any other domains.

Recognizing that large differences exist here is essential to make sense of the reality - that something that seems completely unimportant or barely noticeable to you, could actually be a hugely obvious and important difference to many others (whether it’s a certain screen refresh rate, the presence of a primary color you cannot perceive but others can, an audio frequency you cannot hear but others can, or otherwise).

It’s deeply flawed logic at best (or an intentional red herring at worst) to cite the existence of pseudoscience discussed elsewhere, as an argument against real science being discussed here.

There is a well-understood science to both auditory and visual perception, even more concretely so for the visual side. The scientific literature on human perception in both categories is actively used in the engineering of almost every modern (audible/visual) device you use every day (both in hardware design, and software such as the design of lossy compression algorithms). We have very precise scientific understanding of the limits (and individual variation) of human visual and (to a slightly lesser extent) auditory perception and preferences.

I have absolutely no problem paying a premium for an upgraded display. The problem is that Apple does not offer that option for the MacBook Air.

The MacBook Pro has an amazing screen, which is why I bought the MBP. But the MBP compromises increased weight (which I don’t want) in exchange for more performance (that I simply don’t need). And we know this compromise is not needed to host a better display, as evidenced by the existence of the iPad Pro.

Don’t get me wrong, the MacBook Pro is a fantastic product and I don’t regret buying it. It just feels like a huge missed opportunity on Apple’s part that their only ultra-lightweight laptop is so far behind in display tech vs their other non-laptop products (like the iPad Pro which is lighter still, just crippled due to iOS limitations).

I would gladly pay even more than the price of my MacBook Pro for a MacBook Air with a screen on par with the iPad Pro or MacBook Pro. Or even for an iPad Pro that runs OSX!

Yes, human visual perception exists along a spectrum of temporal, spatial, and chromatic resolution that varies from person to person — I’ve even met some people who can’t perceive the difference between 30hz and 120hz, while to me and most people I know, the difference between 60hz and 120hz is enormous.

So you could make the same argument against high DPI displays, superior peak screen brightness, enormously better contrast ratio, color gamut, etc. Also speaker quality, keyboard quality, trackpad quality, etc.

Where does this argument end? Do you propose we regress to 60hz 1080p displays with brightness, contrast, and viewing angles that are abysmal by modern standards? Or is the claim that the MacBook Air’s current screen is the perfect “sweet spot” beyond which >99% of people can’t tell the difference?

I think the market data alone disproves this pretty conclusively. Clearly a significant enough percentage of the population cares enough about image quality to vote with their wallets so much so that enormous hardware industries continue to invest billions towards make any incremental progress in advancing the technology here.

To be fair, I think there’s strong data to support that modern “retina”-grade DPI is good enough for >99% of people. And you can argue that XDR/HDR is not applicable/useful for coding or other tasks outside of photo/video viewing/editing (though for the latter it is enormously noticeable and not even remotely approaching human visual limits yet). But there’s plenty of people who find refresh rate differences extremely noticeable (usually up to at least 120hz), and I think almost anyone can easily notice moderate differences in contrast ratio and max brightness in a brightly lit room.

I’ve considered trying an ultralight PC laptop with a superior screen. But the sad state of reality is that:

(1) Windows these days feels like a constant battle against forcibly installed adware / malware.

(2) Linux would be great, but getting basic laptop essentials like reliable sleep/wake and power management to work even remotely well in Linux continues to be a painful losing battle.

(3) Apple’s M series chips’ performance and efficiency is still generations ahead of anyone else in the context of portable battery-powered fanless work; nobody else has yet come close to matching apple here, though there is hope Qualcomm will deliver more competition soon (if the silicon’s raw potential is not squandered by Microsoft).

Just because Apple’s competition has been complacent and lagging for many years, doesn’t render irrelevant any feedback to Apple regarding what professional laptop users would like.

I really wish Apple would make a MacBook Air variant with display quality on par with the iPad Pro or MacBook Pro (ProMotion/120hz and XDR/HDR, at least). The screen quality is the only reason I currently use the Pro despite its chunkier weight, since the local compute/memory of the Air is already plenty for me (and most users).

The iPad Pro proves that weight and battery life is no excuse here for the lack of state-of-the-art display tech in the MacBook Air. And as for cost — the base 14” MacBook Pro M4 (at $1600) isn’t significantly more expensive than the 15” MacBook Air M4 configured with same CPU/RAM/SSD (at $1400).

It’s really quite a shame that the iPad Pro hardware is in many way a better MacBook Air than the MacBook Air, crippled primarily by iOS rather than hardware.

You’re right that it’s not inherently unique to EVs, but it started with EVs and now this new dangerously fragile design (of having a single monolithic computer console handle display and control of everything from critical drive modes and gauge display, to non critical things like music and playing fart noise jokes) is infecting ICE cars too (e.g. BMWs new touch screen AC controls and unified touch screen dashboards rolling out to all new cars, Audi doing something similar now, etc. — all following after Tesla, but with crappier software).

I’ve owned and driven EVs from several brands. Prior to this, I could pretty much always expect the following from my car:

1. The drivetrain always operates normally and safely (aside from some actual mechanical failure) with no computer glitches.

2. I can always see my speed and gear selector state on a dashboard somewhere, even when (not if) the infotainment screen crashes and reboots. I’ve had (2010-2020ish era) Lexus, Audi, and others have infotainment glitches, crashes, and reboots, but the speedometer, drive train, and AC all had physical controls running on isolated systems and so they always continued to work through a reboot or glitch of the infotainment.

3. The AC is always operating (aside from some actual mechanical failure) with no computer glitches or lag to my ability to control it. I consider this a critical safety system given that many drive in climates with weather that can be dangerously hot or cold.

In pretty much every EV I’ve owned, none of these have been true except maybe #1, and that is pretty sad to say that the only thing that hasn’t happened is my entire cars wheels locking up on the highway (and yet still this is reported happening for many EV brands, Tesla, Audi, and Porsche at least come to mind where I’ve read stories).

It’s insane to me that it’s even possible for the cars computers rebooting to entail AC shutting down, not being able to see your speed, etc. If this EVER happens, the entire vehicle line should legally require a recall until it’s guaranteed this won’t happen. We have ways of guaranteeing computer systems don’t fail like this to extremely high probability — car companies only don’t do it because it’s expensive and more complex than just throwing all the same crappy software into one single system rather than designing multiple isolated fault tolerant systems.

Less horrible but still shockingly bad regression is how almost all modern cars AC is controlled through an often laggy computer system (not to mention the almost universally despised move of AC controls to touch screens, instead of physical controls). Maybe not so laggy on Tesla, but in my experience both BMW and Audi have AC control touch screens which sometimes respond but occasionally can have 1-10 second random lags before anything responds. Presumably due to garbage collector lag or something. But this is also a mild safety issue since the lack of predictable behavior from common controls makes it very distracting when trying to so something so common and simple as adjusting the temperature that should just be as simple as a simple physical button or knob.

I’m not claiming that quality or licensing concerns are irrelevant, but that:

1. It will not be possible to reliably detect whether code was LLM assisted or not.

2. Humans are not always 100% truthful.

3. All the concerns cited here also applies to human-written code anyway.

So attempting to treat LLM coding assist tools as a special case here is going to be a losing battle. To solve these issues, we’re gonna have to come up with code review processes and tools that apply to ALL code up for review.

Do they think Copilot/ChatGPT are usernames of contributor accounts which can be banned from a project/repo?

It seems the underlying issue here is bad code submitted ultimately by human contributors. Consistently thorough code review and testing will always be necessary.

Low fat diets are not healthier. The notion that fat is evil has been thoroughly debunked for quite some time now.

100% agree it’s a huge mistake to water down Rust’s `enum` and `match` when teaching Rust.

If anything, these concepts (algebraic data types and pattern matching) are by far the highest “ROI” features to learn of all: they’re super easy to learn, intuitive to read and write, immediately useful, and unlock a level of rich, expressive, inherently robust code not possible in many other languages.

However, I do agree with the author that the borrow checker is the first area of “friction” for most, and so suggesting more liberal use of cloning is a completely fair strategy to make the learning curve less steep for beginners.

I actually fed two GPT4’s into each other as an experiment and they very quickly devolved into just saying things like “It’s clear you’re just feeding my answers into ChatGPT and posting the replies. Is there anything else I can help you with?”

Achieving <80% on CIFAR10 in the year >2020 is an example of a failed toy model, not a successful toy model.

Almost any ML algorithm can be thrown at CIFAR10 and achieve ~60% accuracy; this ballpark of accuracy is really not sufficient to demonstrate viability, no matter how aesthetically interesting the approach might feel.

I certainly don’t disagree with you there (about ease-of-use vs safety). I personally believe the “right” choice in C++ is to err on the side of safety (at least when adding new stuff), and wish STL did too — especially given how notoriously difficult it is to write C++ code that isn’t plagued with memory corruption bugs

I know anything other than switching to Rust is a losing battle of compromises in some sense, but many of us working with huge codebases written in C++ don’t have the luxury of that choice being unavailable.

The root cause of the danger you illustrate here IMO is actually a fault in the way the `std::string` to `std::string_view` implicit conversion is designed, as opposed to something inherently wrong with the concept of view/span classes.

In my opinion, `string_view` and `span` are otherwise wonderfully much-needed concepts (that arguably should have been built into the language itself, like Rust slices) that should never have been designed to allow implicitly creating views of temporaries. Disabling implicit conversion/construction from rvalue references is quite possible in C++17 and beyond, and very effectively prevents accidentally implicitly viewing temporaries just fine (I’ve done this myself in some enhanced span-like classes I’ve written), so I’m really not sure why string_view was designed this way.

https://en.cppreference.com/w/cpp/string/basic_string/operat...

Every wonder why Steam has so few Mac OSX users? Might it have something to do with how few major games are available at all on Mac OSX?

It doesn’t exactly take brilliant insight to see that this is a bit of a chicken-and-egg problem, and so them taking this justification to avoid porting (and I think you’re probably correct that this is their reasoning) creates quite a self-fulfilling prophecy.

I was in a similar situation, and found a solution that works for me! TDLR; Topre’s feel is still best, but Cherry SILENT Red (not to be confused with Cherry Red) is “close enough”, while opening the door to the larger world of custom and specialized ergonomic keyboards.

Like you, I prefer the sensation of typing with Topre switches. But I cannot escape the fact that ergonomics is by far the most important factor for my health — and as you know, there are sadly few keyboards on the market with Topre switches, let alone those focused on extreme ergonomics. (Again to emphasize the importance of ergonomics; as pleasing as aesthetics and sound and feel can be, they do not cancel out RSI).

Fortunately, I found a keyboard (and switch) with incredible ergonomics, aesthetics, AND great tactile and auditory sensation! Specifically, I use the ZSA Moonlander with Cherry MX Silent Red switches. I’ll spare you the rant about how amazing the ZSA Moonlander keyboard is with its ergonomics and firmware customizability, and explain Cherry Silent Red versus Topre.

The main problem I have with most mechanical “spring” key switches is the jarring nature when they bottom-out (as well as the loud click of some, which I personally hate); it’s very clear when the switch bottoms out abruptly and violently and makes a “clack” sound (not to be confused with the loud “click” sound that clicky switches like Cherry Blue make that some people like). I prefer a smoother sound and feel.

Unlike any other Cherry switch (or any others I’m aware of), the Cherry Silent Red has effectively a rubber pad on both the downstroke and the upstroke, avoiding the harshest “clacking” you get from mechanical spring switches that also is shared by Topre.

Now with that said, Topre still has an overall smoother and more graceful “force curve” than any “linear” spring switch will provide. But the rubber padding makes a huge difference and is (at least for me) good enough.

You're right, I initially completely misinterpreted your point -- sorry about that. I edited my response accordingly.

My new point is simply this: You can make this argument about anything whose value declines relative to other things; it doesn't make it a useful interpretation of the actual market trends.

For example, I could say that the price of computational resources hasn't been falling over the past few decades due to advances in computer efficiency an speed; no, rather I claim that this is just an illusion -- the cost of computation has actually remained fixed, while the price of literally everything else in the world has been rising, as long as we denominate them in terms of computational resources! But this is silly and doesn't reflect what's actually happening, because I'm subtly making an axiomatic/unfalsifiable assertion here (that the 'true cost' of computational resources is fixed).

It might be easy to misconstrue as reflective of truth because many things do rise in cost, so it intuitively might seem like a profound alternate viewpoint on reality that logically checks out. But, in reality, the viewpoint is simply wrong / not useful -- a reflection merely of an axiomatic assertion yielding a tautological claim, with no external evidence to back it up.