wohlergehen

Things like OS search indices, recently used items databases or thumbnail caches can store hashes of files to identify whether they are up to date and referring to the correct file.

Genuine question: All the commercial vessels continually broadcast their position, speed and bearing, right?

Even if the military vessel does not do so, it should be able to receive all commercial vessels telemetry...

So how can it possibly crash into one?

I dont think this will be adopted broadly unless it matches the price of real milk.

Given that oat milk cuts out the middleman (middlecow?), should this not be cheaper anyway, or are there scale effects preventing it from reaching price parity?

As an added bonus, the big coal exporting countries tend to be relatively more democratic than the big oil exporting countries.

I.e. electrical power is already relatively cleaner by not supporting horrible regimes (cf. Saudi Arabia or Nigeria).

Has anyone actually benchmarked the vector lookup for various vector sizes in scala?

That would straightforwardly determine if the runtime is "effectively constant" or "effectively logarithmic", e.g. cause it is really O(a+b*log(n)) with a >> b.

ParaView would be a possibility. Many tools can export VTK files for it.

Would you mind sharing that? That would be super useful to me as well.

Does anyone know why these apparatuses are usually "inverted", i.e. hang from the ceiling. I've seen the same thing for quantum computers. Is it related to the way they are cooled? Or is it easier to work with somehow?

Of course this has already happened :)

https://youtu.be/dhM_2ubR9Aw?t=213

There's also "Skunk Works: A Personal Memoir of My Years at Lockheed" by Ben Rich for a more intimate perspective.

In that case I'd recommend "Quantum Mechanics for Scientists and Engineers" by D. Miller. When I (a trained computer scientist) had to pick up a bit of QM, I first watched through Susskind's videos (saved my ass) and then took a lecture based on the aforementioned book.

Exactly. Going back to hunter-gatherer and living off the land like that is hardly sustainable with the current population: Earth has 5.7e7 sq mi surface, but 7e9 people, i.e. 8.1e-3 sq mi/person. Hunter/gathers seem to need (after a short google session) need around 1e-2 (Tropical) to 1e2 sq mi (Savannah) per person.

How is this a useful comparison? It's 10kcals of ingested food per 1kcal of energy the hunter expends (not including e.g. the energy of the game). By that measure, I'm super efficient cause I buy food at the supermarket, and don't really expend any energy for it. That number has nothing to do with the second 1kcal/10kcal modern figure.

"The Swarm" by Frank Schätzing has a semi-related plot circling around the clathrate gun hypothesis.

Which really is a shame. When I read the headline, I thought "Oh, like AOSA, great idea to let students do it". Then I clicked through the chapters, and was continually disappointed. I mean the selection of projects is quite promising... I'd love to read about each of them, but the content is an astonishingly accurate description of the aspects of "Software Engineering" as a university discipline that I passionately hate.

On the other hand, I don't really have found a good solution how to teach Software Architecture to students. I feel like it's one of these things that come primarily with experience.

It really is no wonder that the project turned out this way if it follows the "deliverable each week" scheme, it's a terrible way to do a university project. You absolutely have my sympathy.

Highly focused on metrics, code organization and project stakeholders, with little discussion of actual architectural choices. I particularly hate the "lets run static analyzer on this project, and use that to make a judgement" step they all seem to do.

For actual introductions to very specific architectures adopted by open-source software and the reasoning behind, see [1].

1: http://aosabook.org/en/index.html

The real name is actually painted at the top. It's BBC Volga [1]. Currently in the Bahamas.

1: https://www.vesselfinder.com/de/vessels/BBC-VOLGA-IMO-943632...

TLDR (I think): Due to Github, recruiters can poach employees easily. As such, hiring anyone who has not peaked is useless, because he might be poached after training/getting better and we will have to compete with other (possibly more wealthy) employers to retain him.

There is an incredibly powerful German anti-war poem titled "say no" [1], that I instantaneously associated with the title.

I think it's important for people involved with technology to have an ethical compass that of course includes criminal activities, but ideally also goes beyond that.

1: http://www.swans.com/library/art13/xxx123.html

Also: If you're asking me to subscribe in the middle of me reading the article, I still happily will enter my e-mail address and click subscribe, but please, please, please do not send me to a new page then to enter more details. I'll instantly navigate back since I actually want to read the text.

Does anyone know how much time a typical TensorFlow model spends in transposition routines?

In the case of physics, the name tensor is well deserved and used with purpose---the object are mathematically speaking tensors (and where called tensors before computers existed). Their naming is unrelated to their physical representation in memory, even though they end up being arrays. The same is true for various engineering disciplines.

For ML, it just turns out that they can talk about their objects and the corresponding operations also mathematically as tensors.

Python has numba, which is equally backed by LLVM. What makes Julia interesting compared to Python IMO is its macro system much more than its performance benefits.

Presumably the Tesla cards, which are targeted at mostly simulations/HPC and enterprise. These tend to start at around 3K a piece.

For anyone interested in Galileo's story and more generally the history of science, I think that "The Great Ptolemaic Smackdown" [1] is a great read. It is a multi-part series of blogposts that covers the story leading to Galileo's conflict with the curiae and the historical context.

Also, it is written in a really easy to read and funny style. If you have the time, it's great.

1: http://tofspot.blogspot.de/2013/08/the-great-ptolemaic-smack...

If interested in "hard" masonry problems that had to be solved in ancient times, there's also Coral Castle [1].

1: https://www.youtube.com/watch?v=nOoCuDnmtyM

> Coal plants emit much more radiation than nuclear plants - even old ones. There's lots of research about HOW MUCH more radioactive coal power is, but we're talking about anywhere from 50-200% more.

I don't fully understand this argument: A well-maintained nuclear power plant emits next to no radioactivity (compared to the natural background), so a coal plant that emits even a huge factor more radioactivity is still negligible.

I feel like the absolute number (x times more radioactive than nuclear) is really not helpful, if one of the numbers is (just as an example) 0.0001% of natural background, and the other is 0.001% of it.

On a related note, as far as I can tell the STOKE code seems to be incorrect w.r.t. aliasing between y and x, i.e. if x == y + 1. Makes me wonder what SMT solver claims are worth in the literature.

Respectively exponential or linear in the input size. The NFA in exchange is exponential in the states (or rather, it's deterministic counterpart is).

TLDR: When using an iterative solver, do the first few calculations in reduced precision, and only then switch to double -- It is more energy efficient.