selecsosi

Why would you type the wrong letter when you mean the other one? Authentication means proving what something says it is, authorization is allowing someone to do something.

The concepts are different so it's not like a dial you are turning to make a measurement, the terms are for different domains

This gels so much with my experience, thank you for sharing. The lack of down-regulation is exhausting at times.

Your work on gratitude mirrors my experience as besides that and some psychedelic experiences to help process from a less traumatic dissociative state were some of the big keys for me helping to process the grief and anxiety coming from those experiences. I wish you the best! Breaking a trauma cycle is beyond difficult and tiring.

AuthN and AuthZ are appropriate and succinct ways to express the concerns when brevity is required

That is exactly my perceived experience and has been useful to me to read but I have immense difficulties spelling. Essentially can speed ready by shape and contextual grammar clues but cannot form the internal shape of the word. I had thought it was an adaptive response to dyslexia, fun to see others with it as a non-adaptive response. I also have a similar response to dense text where it's necessary/useful to slow down to fully grok.

Orthogonally I have excellent memory and pattern recognition for numbers so it's a fun mystery. Vision itself is such an interesting sense and it's super interesting how languages can feedback into the perception mechanisms.

I have a very similar personal experience. Perchance, are you dyslexic? Part of my applied /intuitive reasoning comes from my inability to perceive direct language but early ability to read based on contextual extraction that applied to problems solving and communication.

The brain is so interesting at what point certain pathways activate. The blocks/shapes of text piece is especially similar to my experience.

This gels with my experience (purdue physics/math grad 2011). It also reflected in the composition of staff for the departments. I think to field a robust math department you don't need the staff levels that you do to field a dual theoretical/experimental program in something like physics or chemistry.

e.g. at MIT the ratio of grad students in physics/math is 2.4 even though the undergraduate programs in Math (especially if you include Math+CompSci) are much greater

https://registrar.mit.edu/statistics-reports/enrollment-stat...

It depends on the resolution of the timestamp. With 1000s of posts a second coming in, the fallback would be that it is at the same "timestamp" and then we fallback to a greater identifier. My guess is that the resolution of the timestamp allows for more efficient postgres index usage then the id index which my guess is larger and a string so not as efficiently searchable as the underlying identifier

(my guess is time based index offer faster search performance or lower overhead than a string based search index which doesn't understand it is representing encoded time data)

They could be using something like application generated time sortable UUID7s or some other sortable key

https://uuid7.com/

[edit] CreatedAt timestamp could be something from the client when the post is submitted or tagged from the ingest server and not when they actually are processed and hit the database

The only constant is that things will change, good architecture accommodates by only making the necessary decisions to provide safety and usability with the hooks and studs to build what's needed now.

I think the problem is that you are also trying to represent the biases of directionality in the connection e.g.: [-1,0,1]. In the adjacency matrix, if you store the sign (direction) of the edge, then I think that would map fairly isomorphically.

Momentum can be a gift or a curse. Keeping average speed above 0 is often a more important metric than top speed when you are trying to go far

This would be fun to wire up into django middleware for some quick VM binding for react frontends. Appreciate the approach, as other's have said kudos and keep up the good work! Any work in this space to help alleviate the boilerplate is great

You are going down the path of theoretical particle physics! It is the ultimate question of that to answer what is the fundamental element that makes up matter and what should we "name" that has a useful property that can either be used or helps to explain how other things work.

In reality, "protons" do not "exist" but are semi(very) stable collections of energy that interact in an interesting enough way in a group that it is useful for us to retain the name, rather than refer to it by its constituents.

Electrons don't really glob up into things like atoms due to repulsion (no moderation by the stron/weak nuclear forces) so we don't have a really useful reason to keep going beyond the definition of the electron so we just stop trying to find additional constituent parts.

I would argue we utilize symmetry of rotation and balance along with holding a blade at a fixed point (laythe) or rolling hot metal between two bodies (ball bearings), the avantage / creation of that was one of the crucial advances of humanity (being able to make actual circles / cylinders / spheres, since most objects you mentioned are also 3d)

What makes the circle unique (or a copy / scaling of the unit circle) is that it exists defined by a relationship that is true on the euclidean plane, something itself which is ideal, and only exists in our imaginations.

With mater being quantized at some level, we are always approximating, and for my car's sake, things rolling at several thousand rpms, we have some pretty circular things.

Not to say this addresses the concern but in that case a computation system is being driven by a pure line signal (the power) and is ultimately both dependant on that for operation, and requires a high level of stability of that line to be "effective".

The alternative in this case is imagining that you have a biological process that has adaptive energy response controlling the movement from short term, to long term and itself optimizing based on the performance trade-off of input stimulus overload.

In that case, the process of deciding what to, and when to move data from short term to long term memory (in a feedback loop) could start to approach more of the appearance of consciousness (choosing what to process, what to remember, adjusting the input rate and types of input based on the limitations of processing and retention).

Many animals have consciousness as well so I think it's important to be more open to reducing the bare essentials of what a process of "awareness" is or could be rather than looking at it from a hyper evolved language/information influenced state of humans.

Note I agree that it is not consciousness, but i think it's an interesting exploration to look at the modeling of it from computer forms since we are made of non-conscious atoms and still made it here

Your experience matches my own with TA'ing being a fertile training ground for solving problems from a variety of perspectives on your toes. I also appreciate your resultant interview approach with which I've also found great success.

The point for me that is most revealing is the adaptation to the sticking point, as the questions are engineered to zig, when a candidate zags more than from solving any particular software problem.

A specific but satisfying use case, my apt building was being stingy* with handing out RFID tokens so I used it to copy and program a cheap RFID token for lending to a trusted visitor.

* Stingy => security protocols that I agree with in sentiment but unfortunately I need to let my pet sitter in and it's nice to allow them to keep the keys as I travel frequently and key exchanges are less than optimal for my spouse and I

Beam divergence, while happening at greater rate not in a vacuum, is a property of how photons propagate, so it's not something that can be eliminated

Many years ago, I used the python kernel / API to build a headless asset production pipeline for creating custom 1080p photo slideshows from user images set to produced music. The entire project took less than a month from keyframe generation with Pillow and then using Blender VSE to stitch the static assets together into what was a very polished product for such limited dev time start to finish. I had only briefly used it in college playing around while getting comfortable with linux and the fact it was OSS, and linux compatible was always a shining light for attention to cross-platform development best practices.

By virtue of exposing quality, "developer" first capabilities, Blender in my mind continues to set an extremely high bar for technical product quality and makes it a top candidate when I consider any digital asset project. I'm inspired by the continued dedication of the development community and sponsors and wish the team another 30 years of continued success. Things seem to be picking up every year and massive rewrites / expansions of performance capabilities come every new version.

As part of the definition of the monad it includes "applying" computation on the underlying dataset (wrapping and then applying a function is defined). The mondad also defines the resultant type of the operation (unwrapping) and so in another language, something like mapping a function over a dataset is a built in "consequence" of the definition.

This makes it natural to express a program as a pipeline of operations over a variety of dataset inputs by wrapping, mapping, and unwrapping operations.

An example is something like the IO monad in haskell allows a programmer to incorporate non-functional (side-effect) computations (e.g. writing to disk or to console) naturally where as a 1d list (more like a buffer) doesn't naturally express the same semantics

Being _hyper_ reductionist, the standard model of physics, which describes, to an extremely high degree, the interaction (creation, elimination, and chance of one thing turning into another) of particles that constitute "stuff" (matter / energy) does not account for the interaction of matter via gravitation.

Most things interact with each other, they create "particles" to exchange force / energy (actually in the case here much of this is actually virtual particle interaction / creation). An example of this is when two magnets are interacting with each other, the repulsion/attraction that the magnets / and magnetic fields, the force carriers there actually described by best by "virtual photons" https://en.wikipedia.org/wiki/Virtual_photon#:~:text=Virtual....

The whole standard model is like this (protons/neutrons exchange gluons via the strong nuclear force).

For gravity, there isn't a good "particle" we've found that can accurately describe how gravity behave. It's good to note here a "particle" is what is used to "quantize" something, and so quantum mechanics is the study of how these subatomic particles interact and are created when things like protons are smashed together in a particle accelerator.

Gravity, on the other hand, behaves more like a classical theory (in the sense that it is a field, rather than a discrete quantized energy exchange). In the parlance of the field, mass, appears to cause a curvature in space time according to the rules (quite accurately predicted) in general relativity.

This appears to be some sort of more reconciliation by acknowledgment of the different aspects of the behaviors of the models (GR vs QM), and chooses a path to verify this by looking at the breakdown of matter (in this case, entangled quantum particles) in a asomtomic limit that might be imposed on a quantized theory when that is subatomically reconciled against a continuous field like gravity.

(My own notes) This might be indicating that there is a continuous nature (or at least continuous at the scales that these subatomic particles experience) such that when the particles interact with the continuous field, their movement creates some snapback of the field (thinking how movement in water creates microscopic cavitation) which disrupts what would be a continuous laminar flow in the water case, creating macroscopic turbulence (in this case disturbing the entangled particles)

* Edit: a little clarity / notes on the SM/QM/GR overlap

Not for sqlserver, but for postgres working w/ python, the psycopg library has a API for getting the bound parameter expression

https://www.psycopg.org/docs/cursor.html#cursor.mogrify

Most of this I would expect to be more on the client side than server (though w/ sqlite you are working w/ both)

One of my favorite research projects on the subject. This research publication from the 2014 siggraph conference covers a ML backed design optimization approach to design novel paper airplanes designs, and simulate flight paths.

[Research]: http://www.nobuyuki-umetani.com/publication/2014_sigg_pterom...

[Video]: https://www.youtube.com/watch?v=-KJUVJAUY8o

Also as someone with a physics degree, it's difficult for me to think of taking courses beyond sophomore year that didn't involve SVD to some extent or were using proximal solution strategies (solid but not crazy tough public state school, late aughts). It's not something skipped for time, it's a basic tool used in multiple branches of physics/math. I'll need to look further to validate some of the content/capabilities but as with most things, buyer beware.

Hear, hear

I don't think he was arguing against paying taxes, I think he was arguing that there are _some_ transactions which it is reasonable to expect both a level of privacy, as well as obscurity to the government for purposes of dissent or lack of relevant jurisdiction (private, ad-hoc transactions are still a thing).

e.g. purchasing large amount of print materials / equipment or other distribution tools that are then flagged / targeted by a less than agreeable local political group to repress participation

While federal laws "sound" nice, enforcement and targeting of most laws are still done by local, less arguably democratic practices

Thank you for doing the often uncomfortable act of reminding people of the historical facts and record vs. revisionist subjectivists. There is plenty of historical record and primary sources to back your line of discussion here.

You're addressing the shame and internal conflict of someone who "believes" they are a "good" person, but have met reasonable evidence to show otherwise, and instead of address the conflict, they would prefer to retreat into denial (not unlike Southern Revisionists!)

It was similar at Purdue in the 2000s

I've always appreciated the approach taken by https://www.aosabook.org/en/index.html which has different depths of analysis and scope and gives the user the option of diving into different types of programs. I usually recommend it to friends /colleagues as a foundation when they've had to write networking or other tight performance applications. Evaluating design decisions and process of app evolution is just as important as where a modern piece of software is because it can help provide similar context around getting the product out there now, rather than perfect later.

Also the world is constantly changing so trying to make a perfect product is partially an evaluation of how long your solution will even be relevant or how isolated it is from external change.

Discussion and advice for rotating secrets: https://news.ycombinator.com/item?id=34255319