DVDs are extremely robust against scratches, even more so than CDs. Unlike CDs, which have the data protected by only a thin layer of lacquer under the label, DVDs sandwich the data between two layers of polycarbonate. The error correction is improved too.

Unlike hard disks, they're practically immune to shock (e.g. being dropped). Unlike SSDs and unlike hard disks, they're immune to ESD. And even if you somehow manage to damage one, it's just one, not your whole collection.

Serifs are chosen intentionally to be harmonious with the overall letterforms. They provide a feeling of visual stability and additional cues for recognizing letterforms. They provide a kind of consistency. They're not a distraction.

Jaggies come from a limitation of the pixel grid. They arbitrarily make diagonal strokes and curves bumpy while horizontal and vertical strokes are perfectly smooth, an inconsistency that would otherwise have no rhyme or reason behind it. Before letterforms were constrained to square grids, nobody was making diagonals and curves bumpy because it was a desirable aesthetic effect.

Jaggies are a distraction from the underlying letterform we all recognize. We know they are an undesirable distortion. Serifs are not. They serve an intentional aesthetic purpose, proportioned in a carefully balanced way.

Serifs are a skeuomorphic artifact of stone-carved text. They're no more legible than sans-serif fonts (see https://news.ycombinator.com/item?id=47492894 ). The only reason people like them is because they're used to them. You can get the same feeling from bitmap fonts if you read them enough.

I very intentionally didn't say serifs were more legible than sans-serif.

There are reasons people like them more than just that "they're used to them", however. I named a couple of them. Just because they originated in stone doesn't mean we kept using them for the same reason. A lot of things originate for one reason and then become used for other reasons.

Believe me, I got "used to" bitmap fonts throughout the 80's and 90's. But I still always preferred the 300dpi version of a document from my LaserWriter and then inkjet. Getting used to bitmap fonts never meant preferring them for general computer usage. Jaggies that appear arbitrarily on some strokes but not others is not visually pleasing. Nostalgic, maybe, but virtually never anything you'd choose if you weren't intentionally trying to create a retro vibe.

If you want minimum latency, you want the input side of an traditional vocoder, not an FFT. This is the part that splits the modulator signal into frequency bands and puts each one through an envelope follower. Instead of using the outputs of the envelope followers to modulate the equivalent frequency bands of a carrier signal, you can use them to drive the visualizer circuit.

That can be done with analog electronics, but even half an analog vocoder needs a lot of parts. It's going to be cheaper and more reliable to simulate it in software. This uses entirely IIR filters, which are computationally cheap and calculated one sample at a time, so they have the minimum possible latency. I'd be curious if any LLM actually recognizes that an audio visualizer is half a vocoder instead of jumping straight to the obvious (and higher latency) FFT approach.

Using LLM to target an FPGA would be interesting, generate the Verilog/VHDL.

much faster, above all, but this would be electronic signal processing with quantisation etc. I am targeting zero latency purely analog circuit. I'm not sure if FPGA can do it...

I got it to produce me an 6x lm3915 + 10x TL072CP + 18 1N4148 solution. Need to order the BOM and try it in free time

Audio mastering is already automated to the level of a mediocre human:

https://github.com/sergree/matchering

(I haven't actually tried this, I just watched the linked Benn Jordan video.)

IMO, the ideal would be for all music to be supplied unmastered so the listener's playback software can apply this process to their own taste. Mastering is necessary for listening with garbage playback equipment (e.g. phone speakers) or noisy listening environments (e.g. cars, parties), but it makes things sound worse in good conditions. The best sounding music CDs I own are classical CDs on Telarc that have liner notes bragging about the complete lack of mastering.

> Mastering is necessary for listening with garbage playback equipment (e.g. phone speakers) or noisy listening environments (e.g. cars, parties), but it makes things sound worse in good conditions.

Eh? I listened to it on quite good nearfield gear, in a decent room, and the AI track linked above still sounds like it needs a good bit of help from a responsible adult to bring it up on this rig. :)

Good mastering helps everywhere -- on all systems. For instance: The sound of Steely Dan is pretty good on playback with about anything, I think, and that sound took a ton of work.

And while classical music is not my first preference, I do love me a good Telarc recording. I strongly suspect that the signal path that they use isn't necessarily quite as pure as they insist that it is. Everything is a tone control, including a microphone -- and money is money. They're not going to reschedule an orchestra to fix an untoward blip at 3KHz. They'll just fix it in post (hopefully, as minimally as possible) and send it.

But otherwise, I agree. The mastering process can be automated. Ultimately, it will be. And for sure, it will also be a customizable user preference.

Some of that work has already been in the bag for decades. Ford, for instance, has been using DSPs in their factory car audio systems to shape sounds in unconventional ways for over 30 years. This gives them a lot of knobs to turn, and to fix into constraints, to help shape a listener's chosen music to sound as good as it can on less-than-ideal built-down-to-price on-road audio systems.

Or at least: It sounds as good it can to a consensus of engineers, or of a focus group.

But the knobs exist. And they don't have to be fixed or constrained: They can (and will) be automatically twisted to suit a listener's preferences.

I'll try to make time to check out your link in a day or two.

>What he proposes is radical

It sounds entirely reasonable and moderate to me.

Also entirely ineffective. Banning individual behavior won't prevent collective dysfunction and will only harm honest actors. The only answer that makes sense is reforming the acceptance of work to be resistant to the inevitable (ab)use of AI.

It's the exact same anti-cheating method used to great effect in top level e-sports. You can't trust competitor-supplied hardware, so the only option is for the institution to ban it and supply all hardware itself. Higher education is primarily about competition between job candidates. Eliminating cheating needs to be top priority or the whole system will collapse.

> Eliminating cheating needs to be top priority or the whole system will collapse.

It's going to collapse regardless because of the replication crisis. You might as well tackle the hard problem and figure out how to integrate replication into acceptance, or the consensus publication is intended to represent is meaningless. This is true regardless of whether a human or a robot is performing the work.

It's neither reasonable nor moderate, which is why it'll never happen.

XFCE is saddled with its GTK requirement, and GTK gets worse with every version. Even though XFCE is still on GTK3, that's a big downgrade from GTK2 because it forces you to run Wayland if you don't want your GUI frame rate arbitrary capped at 60 fps.

For people wanting the old-fashioned fast and simple GUI experience, I recommend LXQt.

It makes it easier to treat the computer as part of your own body, allowing operation without conscious thought, as you would a pencil or similar hand tool.

> What use is there in display frame rates above 60 fps?

On a CRT monitor the difference between running at 60 Hz and even a just slightly better 72 Hz was night and day. Unbearable flickening vs a much better experience. I remember having some little utility for Windows that'd allow the display rate to be 75 (not 72 but 75). Under Linux I was writing modelines myself (these were the days!) to have the refresh rate and screen size (in pixels) I liked: I was running "weird" resolutions like 832x604 @ 75 Hz instead of 800x600 @ 60 Hz, just to gain a little bit more screen real estate and better refresh rate.

Now since monitors started using flat panels: I sure as heck have no idea if 60 fps vs 120 fps or whatever change anything for a "desktop" usage. I don't think the problem of the image fading too quickly at 60 Hz that CRT had is still present. But I'm not sure about it.

120 FPS vs 60 FPS is definitely noticeable for desktop use. Scrolling and dragging are night and day, but even simple mouse cursor movement is noticeably smoother.

Outside of gaming, not much. However, now that I'm used to a 144Hz main monitor, there is no world where I would get back. You just feel the difference.

So basically, no use when you've not tasted 120+Hz displays. And don't because once you do, you won't go back.

Mine supports variable refresh rate, which means for most desktops tasks (I.e when nothing is moving), it runs at 48Hz.

Incredibly, Linux has better support than windows for it on the desktop: DWM runs full blast, while sway supports VRR on the desktop. Windows will only enable it for games (and games that support it). Disclaimer: Wayland compositor required.

It’s not enabled by default on e.g. sway because on some GPU and monitor combos, it can make the display flicker. But if you can, give it a try!

Windows 11 idles at around 60 Hz in 120 Hz modes on my VRR ("G-SYNC Compatible") display when the "Dynamic refresh rate" option is enabled, and supports VRR for applications other than games (e.g., fullscreen 24 FPS video playback runs at 48 Hz* via VRR rather than mode switching, even with "Dynamic refresh rate" disabled).

* The minimum variable refresh rate my display (LG C4) supports is 40 Hz.

Ah, then Windows 11 has at least one improvement compared to Windows 10 as far as user experience goes. Good to know, as I never made the jump

I use KDE + Nvidia, and last I looked into it, it only worked if you had one monitor enabled. That's fine for gaming, not for working.

But it has been a while since I've tried it, maybe I should look into it again

AFAIK it’s very compositor-dependent. Works fine with disparate monitors on Sway + NVIDIA (1x144 VRR, 2x60 no VRR + 1x120 VRR when the TV is on).

The state of Wayland compositors move fast, so support may be here. Last thing I’m waiting for is sway 1.12 that will bring HDR.

I, for one, lose track of the mouse way less often at 165Hz.

I lose track of the mouse less often at 1024x768!

You need a bigger cursor.

If the human is killed every 5 seconds and replaced by a new human, they are indeed less conscious. The LLM doesn't even get 5 seconds; it's "killed" after its smallest unit of computation (which is also its largest unit of computation). And that computation is equivalent to reading the compressed form of a giant look-up table, not something essential to its behavior in a mathematical sense.

I'm not understanding how this is analogous to being killed every 5 seconds as opposed to being paused. Let's call it N seconds, unless you think length matters?

> And that computation is equivalent to reading the compressed form of a giant look-up table, not something essential to its behavior in a mathematical sense.

Sure, that's a totally separate issue though.

Because (during inference) the LLM is reset after every token. Every human thought changes the thinker, but inference has no consequences at all. From the LLM's "point of view", time doesn't exist. This is the same as being dead.

The "time" part is what I don't get. If you want to say that "resetting and reingesting all context fresh" somehow causes a problem, that I can see. If you want to say that the immutability of the weights is a problem, okay great I'm probably with you there too. "Time" seems irrelevant.

Pseudocode for LLM inference:

    while (sampled_token != END_OF_TEXT) {
    probability_set = LLM(context_list)
    sampled_token = sampler(probability_set)
    context_list.append(sampled_token)
    }

LLM() is a pure function. The only "memory" is context_list. You can change it any way you like and LLM() will never know. It doesn't have time as an input.

As opposed to what? There are still causal connections, which feel sufficient. A presentist would reject the concept of multiple "times" to begin with.

How can consciousness be possible without internal state? LLM inference is equivalent to repeatedly reading a giant look-up table (a pure function mapping a list of tokens to a set of token probabilities). Is the look-up table conscious merely by existing or does the act of reading it make it conscious? Does the format it's stored in make a difference?

For all practical purposes, calling it a LUT is somewhat too reductive to be useful here I think. But we can try: leaving aside LLMs for a second; with this LUT reasoning model you're using, would you be able to prove the existence of just a computer?

What state is lacking? There is a result which requires computation to be output. The model is the state. The computation must be performed for each input to produce a given output. What are you even objecting to?

It's plausible that LLMs experience things during training, but during inference an LLM is equivalent to a lookup table. An LLM is a pure function mapping a list of tokens to a set of token probabilities. It needs to be connected to a sampler to make it "chat", and each token of that chat is calculated separately (barring caching, which is an implementation detail that only affects performance). There is no internal state.

The context is state. This is especially noticable for thinking models, which can emit tens of thousands of CoT tokens solving a problem. I'm guessing you're arguing that since LLMs "experience time discretely" (from every pass exactly one token is sampled, which gets appended to the current context), they can't have experiences. I don't think this argument holds - for example, it would mean a simulated human brain may or may not have experiences depending on technical details of how you simulate them, even though those ways produce exactly the same simulation.

The context is the simulated world, not the internal state. It can be freely edited without the LLM experiencing anything. The LLM itself never changes except during training (where I concede it could possibly be conscious, although I personally think that's unlikely).

Right, no hidden internal state. Exactly. There's 0. And the weights are sitting there statically, which is absolutely true.

But my current favorite frontier model has this 1 million token mutable state just sitting there. Holding natural language. Which as we know can encode emotions. (Which I imagine you might demonstrate on reading my words, and then wisely temper in your reply)