jkabrg

OK, fair enough.

You're judging Nietzsche using circular logic.

You're saying that because he's respected by some academics, he is objectively worthy of respect. But that's contingent on those very same academics being worthy of respect. And I'm not sure that they are, personally.

This is a major problem in academia and other areas. The academics end up being judged by fellow academics, and any claim to merit becomes completely circular. That's how you end up with things like Brutalist architecture -- it's because the architects are seeking approval from fellow architects; and not the people who have to live in, and around, what they build.

Even worse, those academics who "suck up" to other academics end up having the favour returned to them. He scratches your back, you scratch his, then he scratches yours again... You end up with citation rings.

See Taleb's Skin In The Game.

If you're using Firefox, use the reader mode.

See the slides example

> repairing a cable is really easy task generally

Highly subjective. Some people have other hobbies.

> make the cable replaceable, i.e. have it also be connectorized on the headphone end.

I might try that.

Problem with non-wireless headphones is that they break. One side of the headphones stops playing sounds; you can sometimes spin the connector and get the sound back, but that gets annoying, and it stops working eventually. The less wires and connectors, the less things can break, no?

I'm wondering how many of the complaints here are to do with FrontPage itself, or the very idea of WYSIWYG web design.

To me, it seems that web design is a visual thing. And if so, then a WYSIWYG editor is the appropriate thing to use. I'm suspicious that some people here don't think visually, and don't see the value in thinking visually.

I don't know what people mean when they say "bad markup". I imagine it might include things like positioning elements using absolute coordinates. But I'm sure this is solvable by previewing the webpage on different resolutions, browsers and device types.

If you like quaternions, you should check out the split-quaternions. They're equivalent to the 2x2 matrices, but help you understand them better by using analogies to the highly familiar and intuitive quaternions.

Nassim Taleb had some negative things to say about GARCH.

"GARCH does not work out of sample. It is a good story, but I was unable to use it in predicting squared deviations or mean deviations"

I haven't found it in Rob J Hyndman's forecasting tutorial either.

How does it fare in the Makridakis competitions?

There was a Quanta Magazine article talking about predicting the evolution of a flame-front using ML. The ML algorithm remained accurate for eight Lyapunov intervals; eight times longer than the previous SOTA.

https://www.quantamagazine.org/machine-learnings-amazing-abi...

There is an interesting correspondence between red, green and blue geometries and hypercomplex numbers.

The quadratic form in Blue is the norm squared of the complex numbers.

The quadratic form in Red is the norm squared of the split-complex numbers.

The quadratic form in Green is the norm squared of R⊕R, which is isomorphic to the split-complex numbers.

There's also the dual numbers. They're the only remaining 2d unitary algebra over the Real numbers, but they don't feature in the above paper.

Can someone explain how this works?

We've got the following chip companies: Intel, AMD, Samsung.

And we've got a bunch of "fab" companies: Intel, TSMC, Global Foundries, Samsung.

What's the difference between the two sets of companies? Why can't Intel just buy a "7nm" processor from TSMC?

And what exactly is "7nm"? The distance between the closest transistors in a chip? Just guessing.

[edit]

From browsing Wikipedia, I gather that:

  - A chip company designs a processor. The design is in the form of a circuit diagram(?)
  - A fab company turns that diagram into a physical product.

So for example, ARM designs a chip, which then gets built by Samsung. ARM is strictly responsible for design, and Samsung provides a factory.

It's not clear what he's referring to. It could be garbage collection, libraries, frameworks, optimizing compilers...

  - Garbage collection make performance unpredictable because of collection spikes. 
  - If you use a big library, you now have to load the library, which is an extra time-sink. 
  - Likewise for frameworks. 
  - Optimizing compilers optimize the 80% of our programs that don't need optimizing, and under-optimize the 20% that do.
    (This is DJB's argument.)

That might be what he's saying. He didn't say anything about dynamic vs. static typing.

Also, there's a counter-argument to this, that writing non-garbage-collected code might result in security bugs (to do with memory management). And security is a non-functional requirement that some might say is more important than speed. It's hard to judge though; there are loads of requiremenets in software dev, and it's hard to balance them.

Explain?

I was referring to Kirchhoff's first law which only applies to circuits -- that example is enough to show that I=dQ/dt is wrong. I wasn't considering the cases you're talking about (lightning etc.) because only one example is needed to prove an equation wrong. I could easily produce a single specific example where dQ/dt=0 but there is clearly a non-zero current present.

And you can talk about charge in an area (sort of) if you multiply charge density by area; the result has units coulombs * metre^{-1}. I thought counting charge in a cross-section of wire was something people did, one way or another, even if the units aren't Coulombs but Coulombs*metre^{-1}.

I could very easily have misunderstood something. With the clarifications above, I'm not clear on what though.

The amount of charge in an area stays constant because of Kirckhoff's laws, so dQ/dt=0 always. Wouldn't a better definition be:

  I = q v . A

where q is charge density. A is an area (actually, a normal vector to some flat cross-section, with magnitude equal to size of area), and v is velocity of the charged particles. I'm using the dot product.

If we check the units: (coulombs * metres^{-3}) * (metres * seconds ^{-1}) * metres^{2} = coulombs / seconds

By the way, I fudged the above to make "charge density" have units $coulombs * metres^{-3}$. I'm not the best person at physics.

I never wrote down any notes at university. My process was:

  - Review material on Moodle before lecture.
  - Go to lecture. Don't write anything.
  - Review material again on Moodle after lecture.

I don't know how you don't lose your notes, or have material missing from your notes because you were sick one day.

Go fork? No.

The wording is misleading. For the following reasons:

  - Variolation was common in England at the time.
  - Wikipedia says that he *inoculated* the boy with smallpox. That means variolation.

He didn't infect him, he variolated him. Variolation normally causes a very mild form of smallpox -- but results in a person becoming immune to the disease. Variolation is the predecessor of vaccination.

There are cases where zero-based indexing is more natural than one-based indexing. An example is naming the centuries: It would be nicer if this were called the twentiethan century instead of the twenty-first. When people talk about the fifteenth century, I have to think for a bit to understand what they mean. As such, I propose adding a suffix "an" to the end of any ordinal number to indicate a zero-based convention is being used -- this century would then be the twentiethan century [twεntiθən] as well as the twenty-first. A bit like degrees vs radians.

Starting at 1 is more intuitive, and therefore IMO better.

Regarding some of the advantages of zero-based:

  - Indexing backwards from the "end". Your language can always add an `end` keyword like Matlab does, and this stops
 being an advantage.
  - Indexing cyclically using modular arithmetic. Yes, this is an advantage. Albeit a rare one for me.

I commit less off-by-one errors with 1-based, and I don't have to double-check as much -- so on balance, I prefer it.

[edit]

The amount of karma this comment is getting is undergoing something like Brownian motion.

Wouldn't it be nice if ELisp were implemented in RPython? What sort of difficulties would this face over the JIT compiler in the link? It would have the advantage that such an ELisp implementation would still be an interpreter (annotated and auto-JITted). Presumably, that would keep it "simple".

Apparently, Hamilton defined quaternions as the solution to almost this exact problem. Quaternions are effectively elliptical vectors. That's a lot nicer than defining them in terms of generators and relations.

[1] - https://en.wikipedia.org/wiki/Versor

Yes. In Russian, ty/vy etc.

What about Japanese typewriters? Kanji is similar enough to Chinese characters.

Just because something is OK -- and has been OK for a hundred years -- doesn't mean it isn't also massively brittle (not that I'm saying it is). You're committing the fallacy of induction when talking about economics. Dangerous move.

That's a nice way of looking at it. The trick of adding 1 to the real part doesn't work in the degenerate case where y=-x. In that case, as I said in my earlier comment, an alternative formula is PERP(y)(yx+|yx|). Both formulas need to be used in practice, depending on which one is closer to giving a degenerate answer.

It's a one-liner in quaternion algebra: vu-|vu|

followed by normalization. Proof can be carried out completely in quaternion algebra.

An interesting degenerate case is when u=-v. In that case, another expression can be used: PERP(v)(vu+|vu|) {where PERP(v) means an arbitrary vector perpendicular to v}

followed by normalization. Both formulas need to be used in practice, depending on which one is closer to outputting a degenerate result.

I guess it provides an objective definition of when something is simple or complicated. This could have applications in statistics or machine learning, where "simpler" hypotheses could be less likely to overfit. There are computable measures of "simplicity" (unlike Kolmogorov complexity) based on compression algorithms or computation time (Levin complexity). There are also apparently connections to statistical mechanics, but I haven't looked into it and I don't know if it's useful.

See this paper on how to cluster using compression: https://arxiv.org/pdf/cs/0312044.pdf