HN user
Idk if you ever shared this view with art people. It must have been hard because there is a sort of obligatory necessity that people MUST like art embedded into their worldview. But also, there is a basic universality of art, and I wonder where it comes from, and what would make some people into it, and others, like you, not into it.
HAHAHAH don't even get me started on how bad anaconda is. On how slow the installer + interpreter, how they avoided being a part of the usual pip workflow, bloated environment, cross platform inconsistencies, extremely slow dependency resolution, etc etc etc...
I'm not familiar with the medical side of this study, but I am very familiar with the statistics.
### They did not do a non-vaccinated vs vaccinated people comparison
These studies are observational, and in that sense no manipulation (such as random treatment assignment) was made. So they need to rely on observational studies techniques.
From what I could gather, one of this observational techniques in inverse treatment probability weighting when they say
"we adjusted for all baseline covariates using stabilized inverse probability of treatment weights"
The IPW technique basically tries to solve the problem that some groups of children might have a higher prevalence of the treatment by weighting them in a manner as to mimic random assignment. This is as close as we can get to comparing treated and untreated in observational settings.
Also, the division of the doses of vaccine is also a better characterization of "was vaccinated". If you think getting vaccinated causes something, then getting more vaccinated should increase the incidence of that something.
### Wouldn’t you need a sizable cohort of non vaccinated children?
Yes, but as you pointed out, that might not exist. So by creating comparable groups (via IPW) + treatment intensity/dose we can still arrive at some conclusions.
### The researchers discarded from their cohort 34,547 children for receiving too many vaccinations/too much aluminum (...) Wouldn’t that data be relevant to look at?
They probably did. A few things are important here. The data comes from administrative data sources, so mistakes can happen. While you need to trust the sources, there could be imputation typos, or just weird cases. So the researchers probably went after a notion of "clinically relevant vaccine dosage for this study" to know up to which point to consider data points, because from that point onward, it either is not interesting because its a rare treatment incidence, or just seems like a mistake.
### “adjusted hazard ratio (95% CI) per 1mg increase in aluminum received……” and in the chart Asperger’s is listed at 1.13 (.89 - 1.44) so they do potentially see an increase for Asperger’s
Important to say that with hazard ratios, a HR of 1 means that there is no change because we are in the multiplicative scale (as opposed to 0 meaning "treatment does nothing" in the additive scale).
So in this setting, after adjusting for the treatment probabilities, a HR CI of (.89 - 1.44) just translates to 'no effect'. Nothing out of the ordinary in terms of the interpretation.
just use it on pc. It gets most of the benefits for contact maintenance. You can still watch the reels you are sent. But no reels, and the feed scroll gets very boring suddenly.
I feel I could use it more. I use the link jumping all the time, but that's basically it. What other features do you guys use?
Genuinely curious, how you manage to derive value from this kind of LLM-assisted reading. At least in the areas I know, the gpt summaries are wrong and generic enough to the point that a read of the book summary, or a read of the preface, is more useful than what the AI returns. If I want to summarize something I provide, it works great. But for these more general overviews I don't trust it.
Ill mention a few books and how they've helped me become a better programmer.
- Structure and Interpretation of Computer Programs: this helped me understand the different types of recursion; how a problem can be modeled for the code to solve and how those designs can evolve and how the paradigm I'm using (functional or OOP) can help me
- Refactoring: this book helped me understand the different patterns we can use to alter code. Giving them a name and symptoms is a great way to apply and communicate with other coders. I recommend this book to everyone.
- The little Lisper: this really helped me get recursion. Like really. Although I could use a refresher.
- Clean code: look... this book is bad. Like, really bad. But, at the time I read it, and in the languages I used (Python, C, R) we didn't have many rules on how to write and organize our programs. And I wanted to improve on that skill. And this book gave me pointers and made me think about those ideas. I don't recommend it around, because for many people Clean Code can become a dogma really easily. But if you can question and critique and reach out to other resources and is an advanced beginner to coding, I can see this book bringing value.
- An Elegant Puzzle and Staff Engineer by Will Larson: those helped me get clarity on the role of the technical team in an organization. A team that produces code as the main driver of impact in a product/service. I lump them both because I read them back to back and don't remember, exactly, which is which. But une is more focus on a tech leader and the other a tech specialist.
- Unix and Linux System Administration Handbook: whenever I need to interact with some unix technology, this book delivers in terms of giving me a nice historical view and presenting the common tooling around the tool. This helps me interact with stuff programmatically much better. Be it GRUB, X, SAMBA, this book gave me the necessary introduction I needed, much more organized and thought out and with useful references than the average blog post you will find in the wild.
And I'd like to recommend this Ask HN on problem solving books
I think math exists to solve real problems, so we should give examples of solving real problems.
This is a view that, depending on the lecturer, can be the total opposite of their view. See, many mathematicians don't have this instrumentalist view on math and some even despise it with their heart. They don't care about applications and don't get particularly excited when a problem was solved in another domain with their technique.
Also, solving a particular problem is that, a particularity. They look for generalizations and for rules and methods that can be applied regardless of the domain. That is why weird and contrived examples are the bulk of their logic, since when you look for generalizations, those are the cases where it breaks. Does continuity implies differentiability? For almost all practical cases yes, but when you want to be general, thats when you get the Weierstrass function [1], a weird function that needs to be studied if you want to develop a theory of real valued continuous functions. Yet, such knowledge is extremely impractical for you average engineer or deep learning researcher.
And ok, you might argue that "I'm not at such level, I don't need those particular theoretically interesting cases since I'm not going to study advanced mathematics at that level". Ok fine, I respect that. But that is not part of the "culture" of mathematics (a culture which your lecturer likely comes from). Since the subject is learned and investigated with this "quest for generality which makes us study weird examples" this is how it is passed along, and this is how it is studied and structured.
All that to conclude: I think its best to respect how math is structured. First, it is precisely the type of thought that I think very other applied fields lack, but that actually helps solving problems. It is by seeking generality, and looking at corner cases that stuff gets built. And mathematicians are the best at formalizing those because this is how they advance their craft. Second, and last, its by creating your own investigation and development of examples that you can stride for a more applied and practical version of the course. That way you'll get the most out of the course.
symptoms can change through time, and be triggered by life events. It can be that a person had the coping mechanisms for one manifestation of symptoms, but due to external events, they don't work anymore.
The book I'm reading, Taking Charge of Adult ADHD, mentions that for an evaluation and diagnosis to be reliable in adulthood, the current form of symptoms should be happening for the past 6 months.
great book btw, highly recommend it.
I just use https://hckrnews.com/ for a better glance at a given's day posts. Sorted by upvote, I think, and then I can choose to see the top 10, 20, or 50% of posts. It's an ok curation, but I much preferred one over a newsletter that the author stoped delivering a couple of years ago.
Yep. Just got hired out from an YC backed tech giant here in Latam and I'm joining a large IT consultancy firm. I think its gonna be fun.
Or, OR, maybe, not every bit of code needs to match all the common good practices and standards. I bet you don't write formally or in Elizabethan English all the time. Context matters in the form of common language. Why should it be any different with code? Just because LLMs can pick those practices up?
Let me tell you how I did it. I remember I wanted to start using Linux full-time no BS and have had some exposure to the terminal. I took the opportunity of a new position, I said I had to use Linux for the day-to-day activities and they bought it. So I had to use the terminal/Linux without disrupting any of the necessary activities.
Meanwhile, I read the beginning half of "The Linux Command Line" [1]. I highly recommend it. Looking back, I might have read it until part 2 and skimmed the rest, and read it as I needed it.
Try to do everything through the terminal. File creation, reading, managing stuff, changing wifi, configuring editors, writing files. Try to find the conversion tools with the rest of the "world". Some examples of what I mean are pandoc for md->docx conversion, python pandas for csv -> xlsx. Build small scripts for your tasks. For example, I work in a multilingual company and I have a tool for translation which at first used Google Translate and now uses gpt. At some point in the journey I even had my own custom dwm build. Fun times. Do read about suckless software, a fun philosophy with some genuine principles, and a nice (but not friendly) ecosystem.
So my recommendation is to read something to give you breadth on linux/terminal knowledge so you can navigate around and stick to it as your main driver.
Another book one I recommend is the "Unix and Linux system administration handbook"[2]. It will tell you more about the history and inner workings when compared to the previous one.
if you are into ML libraries, take a look at fklearn, a scikitlearn-like lib, but written in a functional. Fun read to compare both side by side.
I've followed the Strange Loop conference on yt for a long time, and your mention led me to search more about it. Just found out the 2023 edition was the last one. Sad to see such a nice gathering not having another edition.
1) free
2) able to be around if a single CEO isn't around (Wolfram)
3) able to continue if the supporting company is not profitable anymore
4) possibility of greater oversight if popularity rises
5) extensible if one puts the effort into it
uuu cool stuff. Ill def try and start using it
Honestly, memory leaks are common in software, so I really don't buy it when people say they don't need to do it. Unless it is a server where the memory footprint is arguably simpler due to not having graphics + UI + different interfaces and peripherals, I can't convince myself that a reboot won't freshen things up.
hey meatjuice.
Any guides on using nix-darwin for system configuration and etc? Ive been using it for a while and Id love to learn more!
Every week or so. I have a feeling stuff starts to run smoother. And I also tell macos to not recover the windows. I like a fresh start.
hate it.
IDK, I read some reviews about this book and it seemed to be written for a time before remote work. It looked overly worried about office organization and dynamics. Is it true? Because those ideas made ti fall in my priority list form management reading.
idk, feature flags for the entire application seem ok to me. The issue is when we want to add feature flags for part of the users, or only on a part of the application flow. Then it will become a mess.
Where I work, those feature flags are used to deploy a feature for a single application "run time", i.e. a country, which has its own cloud and everything.
BTW have you read the article on feature flags on Fowler's website? Could help you form some opinions https://martinfowler.com/articles/feature-toggles.html
Uuu, wasn't aware that Sage is built on top of Python. This seems promising :D
Yeah. I don't think I'll be after achieving a dense knowledge of the language. I'm at the age (and mindset) for some pragmatics in my choice of learning tools.
I've dabbled with SymPy in the past. Maybe I should follow the Cassela and Berger appendix with both side by side. Might help me understand if there is some value in it.
And yeah. Proprietary software sucks. I suggested Mathematica because its what is mentioned in the book, and I don't hear much from others.
Thank you for the resources! Given that I already can handle myself around viz and algorithmic necessities, I can see the benefit of a centralized source of algorithms and plots. (But, after working a buck with Python and R I still can't find a better plotting library than ggplot2 :D)
My main interest is in algebraic manipulations. like trying to integrate symbolically or mapping between complicated sets, some optimization, etc. But I'm ahead of myself... I should check those resources you sent me, and then form a better notion of what it is that I want :D
Thank you for those, btw!!
They are rockstarts because they have technical skills AND soft skills. Sure, John Carmack was a technical beast when it came to computer graphics. But one can't work for Facebook/Meta from 2013 to 2019 (and as a consultant until 2022) without soft skills.
What a complete view of what one should strive for. I know this is not possible all the time, but it is a nice goal
I’ve been using reading to keep thinking about things I want to focus on. Getting involved with some different thoughts around a theme eases my anxiety around it and allows me to have a more ordered "vocabulary of ideas" around those problems.
This leads me to read some "boring" books, some simple self-help stuff whose sole point is not to engage and make it extremely important in my life, but to think and organize those ideas.
I'll read other nonfiction books because the subject interests me, such as history or math/probability.
However, my approach is insufficient.
I bought a cheap course from a local intellectual in my country (math and philosophy) on "how to have an intellectual life". My main takeaway was that I lack a fundamental question (or a set of) that I should use to guide reading and studies. For you to research, think, and confront your fundamental question you should have a solid base and read many angles, tangential including, around your fundamental question.
This framework should guide the readings since if something is way above your head or skill, there is some base knowledge lacking and you should study those first. The way to organize all of this is through writing and reflection, where you summarize and intentionally spell out shortcomings and conclusions, even if sporadically.
So yeah, I liked the approach and have, during this year, slowly adding to my study routine.
One final add is the latest Ali Abdal video which sort of talks about this https://youtu.be/AvKGYyowFK4?si=Ah_6SM-nhv9uQ55O