PROGCRYPTO (programmable crypto) The capabilities of multiparty computations, fully homomorphic encryption and zero knowledge proofs are super interesting and will allow for new ways of using the internet. In the same way it was hard to imagine what the consequences of asymmetric encryption was at the time when it was conceived, it think the same is true for MPC and FHE today. These technologies are only recently getting feasible. Example use-case: Look for relatives (think ancestry) based on your DNA sequence with cryptographic guarantees, that everyone's DNA stays secret.
I really liked this article. It's the first time I see async Python used to solve a problem.
Suggestion: I think it would make sense to use softmax as the scaling function. This way you don't need to worry about outliers and you get to decide the running time.
Yeah totally agree, it would need a better scaling function to tackle outliers. The basic scaling was just to give an idea of how this could be improved. We can always devise better scaling methods and negative number handling. Thanks for your feedback, will try to incorporate it!
This week I am refreshing my front-end skills, so I did a little market research.
The best resource on CSS basics, I have found, is https://web.dev/learn/css. It includes flexbox and grid.
For design guidelines I recommend Refactoring UI. It is full of actionable advice, which books such as The UX book is lacking.
In my project I use TailwindCSS (by the author of Refacturing UI) with Daisy UI components, so I don't actually write raw CSS. At first TailwindCSS feels like a step backwards to inline CSS, but it is just better than the alternative. I recommend giving it a chance.
Possibly. There are race conditions with `wait` for `$somepid` because shells kinda have to reap processes as soon as possible, but that means that your `$somepid` can be stale and even refer to some other process. If you have a pipe from the process' stdout and/or stderr then when you can try to read from that pipe to know if the process exited -- you might still lose the exit status, unless you run the program like `(the_program ...; stat=$?; printf '%s\n' $stat 1>&2; exit $stat) &` and then read the exit status from the stderr pipe.
(zk-)SNARKs are essentially succinct (sub-linear / O(log)) arguments of knowledge. Essentially the arguments (proofs) boils down to the Schwartz-Zippel Lemma which states that two non-identical polynomials disagree on a random x-value w.h.p.
Its a rabbit hole but it is an interesting niche if you like this kind of stuff.
To get you started, you can read about the KZG scheme which is even O(1):
The best tradeoff for me is (2)+(3) sacrificing (1), so I use a local password manager (named 'pass'). That said, I would never trust a 3rd party like LastPass.
Choose a complex base password + a memorizable _algorithm_ that combines (eg) domain names with the base password. That satisfies all of your requirements, and it's not super difficult to create an algorithm that is easy to perform mentally and hard to reverse.
I started doing this in college, and it's been super easy to keep up and my passwords are pretty secure.
I did this before using a password manager. I found that it does work but it's still a pain.
For example, what do you do when you're forced to change a password? You need to remember that this website doesn't follow the rule. There's also the case of password being used by several websites with different names or urls, you need to remember which is the base one.
Also it's really convenient to store other things than passwords in the password maanger. Like some notes, your customer number, which email you signed up with, some PIN they give you...
Not being able to do #1 can be a feature. If you can’t easily login to your accounts it adds just the right amount of friction to buying things online.
I enjoyed Branching Paths, which gave provides a unique window into the Japanese indie game developer scene -- If you liked Indie Game: The Movie 2012 and are interested in Japanese media, then watch at least the trailer:
180 degree rotational symmetry happens because almost all roads are bidirectional -- if there's a road pointing north, it's also a road pointing south.
90 degree symmetry seems like a consequence of the road system consisting more of straightness and 90 degree turns, with no other specific angle being common.
Rio fits the bill. There's no overarching grid, but if you zoom into any small area, you'll usually find a pattern of rectangular blocks with 90 degree intersections.
As far as I can see, the only viable way out rotational symmetry would be to use concentric 'ring-blocks' either for local areas or for the city as a whole. The rings would cancel themselves out, and the 'spokes' could be places relatively freely.
So... Canberra might have an interesting distribution?
