superjan

Watch the video to the very end: the final splat is not a gaussian one.

The article didn’t go into this, but I suspect that a large part of the dark mode trend is due to evening/night computer use. If you don’t light your room, the screen is the only light source which is unpleasant.

I would argue that you don’t need to learn string theory as it currently does not predict anything we can realistically observe (as you need energies that occurred only at the big bang). If string theory is correct we could observe a “supersymmetric” twin of all known particles, however we haven’t seen these, and they could exist even if string theory is false.

String theory aims to explain all physics as manifestations of a mathematical concept best understood as a vibrating string.

Initially, the hope was that string theory could predict the particle masses we observe, but that hasn’t worked as it turns out there were many different predictions possible. String theory has also struggled to develop a version of the theory that does not contradict known properties of our actual universe.

Loop quantum gravity is not equivalent to string theory, except that it also tries to unify gravity and quantum physics.

As things stand, string theory is not falsifiable, while that is the case, you could argue it does not count as physics.

But, by multiple accounts, it is interesting math, which can be worth doing for its own sake, and it’s happened often enough that interesting math turned out to be useful somewhere. Just not for explaining physics.

Maybe the following helps: if you have a an analog signal where there are no frequencies above 22.05 khz, it is in principle possible to sample it at 44.1 khz and then perfectly reconstruct the original signal from those samples. You could also represent the same analog signal using 48 khz samples. The key to resampling is not finding a nice looking interpolation, but rather one that corresponds to the original analog signal.

This is a story about a unique 1947 Chinese typewriter prototype that never got built. It could type thousands of characters by combining keystrokes. The prototype was presumed lost, found this year and it is now at Stanford.

This is not news, but I read about it here today in a Dutch newspaper article (paywalled, untranslated): https://www.nrc.nl/nieuws/2025/12/27/lang-verloren-gewaande-...

It”s worthwhile to mention that in the US and EU EMRs are generally not considered Medical Devices and are therefore not subject to a lot of regulations.

https://www.medicaleconomics.com/view/what-if-emrs-were-clas...

It’s probably more phonetic than Chinese but significantly less phonetic than Dutch.

i have wasted several weeks worth of evenings on vectorizing heaps (4ary heaps: with SIMD, you’re not limited to binary heaps). It did not provide any speedup. I’d expect that halving heap depth would help but no. Still don’t know why.

Nice to learn about. It’s good to know the field has progressed, however the context focused on JPEG, where my point does apply.

Arithmetic coding decodes 1 bit at a time, usually in such a way that you can’t do two bits or more with SIMD instructions. So it will be slow and energy inefficient.

There are very few people visible in the demo’s. I suppose that is harder?

Yeah, it’s a funny coincidence that all those constraints to make it look random produces exactly one solution. I guess the OP knows this is not ‘random’ in the mathematical sense.

I want to flip coins so randomly that I never see the same face twice in a row.

Yes, or when a character moves across the screen. They are quite fine grained. However, when the decoder reads the motion vectors from the bitstream, it is typically not supposed to attach meaning to them: they could point to a patch that is not the same patch in the previous scene, but looks similar enough to serve as a starting point.

I love this. The widely used standards for video compression are focused on compression efficiency, which is important if you’re netflix or youtube, but sometimes latency and low complexity is more important. Even if only to play around and learn how a video codec actually works.

Well, it is not pretty to see how the sausage gets made, but extracting formatted text from docx is absolutely doable, no PhD involved. Source: I have done it as a little sidequest because it was useful to audit a set of word documents.

Ok this is correct for traditional JPEG. Other flavors like Jpeg2000 use a similar (but lower overhead) version of this byte-stuffing to avoid JPEG markers from appearing in the compressed stream.

Related: https://en.wikipedia.org/wiki/JPEG#Syntax_and_structure

And don’t forget that Date objects are mutable!

There is a particular pattern (block/tag marker) that is illegal the compressed JPEG stream. If I recall correctly you should insert a 0x00 after a 0xFF byte in the output to avoid it. If there is interest I can followup later (not today).

Because there’ll likely be a lot of string concatenation in a language whose job is to output HTML. Adding a single char to a 1k string will require allocating a modified copy, in mutable strings it’s just appending in a preallocated buffer.

It’s true that many languages have immutable strings, but for a web-focused scripting language it makes sense to default to mutable strings. I think the comment is about that you now need to choose mutable vs immutable, and that is framed as a consequence of broader adoption. Which is a development I have also seen before.

The article talks about pollution but I’d expect them to be very noisy as well. Gas turbines are essentially redesigned aircraft engines.

And crypto scamming is anonymous, low risk and can be automated. Scamming people for cash requires you to get close to each of your victims.

Furthermore this assumes the rental company chooses to repair. I doubt they will do so for a scratch like this.

This is from the channel “the Gaze”. The channel is not the project’s author, but it does a great job introducing the idea.

The author’s page: https://wernerdevalk.nl/en/songs-of-the-horizon/

No, that statement was not about Windows, but about the argument for telemetry in general.

You are right: it can be done with the same ALU, for sure. But the data dependency on the carry flag makes it a really different instruction from the point of view of the CPU: three data dependencies in stead of two. For the CPU it is beneficial to treat the instructions differently.

Oh sorry, I did not know that was the guy who hates being linked to from HN. Copy the link manually to read it if you still care to learn what he wrote.

In my recollection it was Netscape, this appers to confirm it:

https://www.jwz.org/blog/2017/09/html-email-was-that-your-fa...

I’m in europe so I am all in on the metric system. But “about a foot” per nanosecond is so easy to remember, understand and reason about that it is worth the exception. If you prefer something European, think of a sheet of A4 printer paper: the long side is 29.7 cm. “One length of A4 per nanosecond” is within 1% of the actual value of the speed of light.