It had some good results but made up a command called `cargo publish --workspace` when asked about publishing a Rust workspace project to crates.io -- that command would be handy, but doesn't exist. Overall 7/10
I agree with the article. The SAT is a chance for any student, regardless of background, to demonstrate academic skills and the ability to succeed. I was one of those students -- my SAT score got me a full merit scholarship. I wasn't as polished or accomplished as other students around me, but I was able to show that yes, I am smart, and that I can succeed academically at the college level. I did go to a korean after-school program to help study for it though.
I feel like instead of dropping the test they could provide free resources to students to prepare for it. At least get them familiar with the structure of the test and encourage them to do the best they can. A lot of kids like those in the article would stand to benefit.
The only winner from getting rid of the test is rich, stupid kids that don't do well regardless of how much prep they get. They then go on to cheat and BS their way through college and use their network to land jobs. Poor and disenfranchised students will continue to be marginalized.
Just find a new gig. I also worked from 2019-2022 for a company without a raise for 3 years. Arguably, I stayed six months too long, but we were doing some greenfield stuff that was actually fun.
There's no need to over-analyze or justify it either. Get a 20% minimum bump by switching employers, like every other person here would do.
The papers are all based in cryptography and zero-knowledge proofs specifically. I would argue they are in the first category you described. They seem to be descriptive and walk through a problem but there are no experiments or anything, just theorems and proofs. I get lost pretty easily
I think I'm on the "just trust that the papers you receive are the state of the art in this field and think about how to implement their conclusions in software" path. I actually prefer that, because I'm not really much of an academic. I would prefer to focus on the code and implementation details versus the theory stuff.
Distilling it down from a really abstract level to something to implement is hard though. To be fair, I'm just starting out with this so it should get easier from here I suppose.
I would agree. I'm actually quite a few years out of school though, I have around 5 years of experience. I do have a bit of graduate-level experience but basically never did any real kind of research. That's why I think I'm struggling lol.
I find it a bit surprising that only a few people knew about this code change. I expect other engineers found that piece of code and probably didn't ask any questions. It's normal after working on a codebase for a while to explore all the different branches. It takes a few years though, from what I gather the codebase was only 2-3 years old, so it's possible no one noticed.
I think this revelation points to the further guilt of SBF and his excuse that "he was incompetent" will fall face down under scrutiny.
Containers are great tools that can encapsulate an application and its dependencies, allowing apps to run anywhere in a streamlined way. Some container images contain commands to start a long-lived binary, whereas others may simply contain data that needs to be available in the environment (for example, a Kubernetes cluster). For example, operator-framework bundles and crossplane packages both use container images to store Kubernetes manifests. These manifests are unpacked on-cluster and made available to end users.
One of the downsides of using container images to store data is that they are necessarily opaque. There's no way to quickly tell what's inside the image, although the hash digest is useful in seeing whether the image has changed from a previous version. The options are to use docker cp or something similar using podman or containerd.
Using docker cp by itself can be cumbersome. Say you have a remote image somewhere in a registry. You have to pull the image, create a container from that image, and only then run docker cp <container-id> using an unintuitive syntax for selecting what should be copied to the local filesystem.
dcp is a simple binary that attempts to simplify this workflow. A user can simply say dcp <image-name> and it can extract the contents of that image onto the local filesystem. It can also just print the contents of the image to stdout, and not create any local files.