Sergey Antopolskiy

Dependency

The original source is the famous XKCD "Dependency" xkcd.com/2347/

Universal Task Estimation Table

easy-peasy – 1h
easy – 2h
simple – 4h
kinda simple – 6h
fine – 8h
sorta fine – 12h
idk – 16h
idk, kinda weird – 20h
kinda hard – 24h
hard – 30h
very hard – 40h
oh fuck – 48h
total fuck-up – 60h
some next-level fuckery – 80h
easy, I think – 100h

day 7 on my no vape journey. there is a nicotine-shaped hole in my soul no amount of Kiev Cake can fill

I got pretty good, but it wouldnt have made any difference if, during a computational neuroscience summer school in Germany, one of the instructors didnt do a small workshop on Jupyter notebooks. I immediately fell in love with Python, which made it so easy to transition from academia to SE later.

I love the smell of refactoring in the morning

Funny story. I haven't done any coding up until my mid-twenties, at which point I was doing PhD in neuroscience and I had to learn how to analyse collected data in Matlab (yes, Matlab was my first language, which is both hilarious and sad).

after watching the video on it, I kinda get the idea better. it is more about just pushing the UX forward, which is by itself laudable

it's really fast, which is awesome. but am I the only one thinking that quality of responses far outweighs the speed? I'd wait for a response for 1 min any day if the quality is even 20% better. the time you waste working with bad responses would compensate for any speed boosts tenfold

watched it yesterday, it was pretty cool! I tried raycast after @t3.gg's video, it is really something. thanks for your hard work!

yeah, same. so much exposition I started to watch some parts on 1.25 otherwise my adhd brain goes crazy. thinking about transitioning to the audiobooks, but it will most likely render the show boring for me

man drag and drop functionality is a beast. a dev on our team spent 3 weeks trying to transition from the deprecated beautifuldnd to dndkit.com and we're still wrapping our heads around how to do feature parity. I had to step in to implement a minimal example, and it is still tough 🤔

From the reactjs community on Reddit: Is this too hard/niche a problem for a Sr React position? Explore this post and more from the reactjs community

it is inspired by this: www.reddit.com/r/reactjs/s/...

yeah, the first this is a must, but often people overlook having a solid understanding of the life cycle. I routinely test React lifecycle knowledge on the interviews with some tricky useEffect side effects exercises

this is for parsing JSON strings, not objects? i.e. in this case you don't need to parse at all?

true, I was thinking about that as well, but I thought the intention was to enforce the presence of specific fields in some places in the application :)

that's genius )) 10 slimes out of 10

the podcast lmao

2024 was really productive for me; I've learned so much both in terms of hard skills, management and being a founder. I am glad I also had the time to contribute individually on a regular basis.

It would have been better to implement `PersonWithAddressAndEmail` as `PersonWithAddress & PersonWithEmail` in case the email type gets changed so that you don't have to modify it in 2 places?

but like... why?

pity you can't experience it twice. the next best thing is watching someone's playthrough on YouTube. it is pretty fun watching people immediately figure out things which took you ages and yet getting blocked in the places you barely noticed. really highlights how people have different perspectives

yeeeep. epic, finished it a year ago, still thinking about it from time to time.

yeah, seems like AI hate/hype is a goldmine of views right now. your video is really thorough and thought-provoking, thanks for making it. it inspired me to write my first thread ever 😅

"Thinking in Systems" by Donella Meadows
"I Am a Strange Loop" by Douglas Hofstadter

5/5

"Superintelligence" by Nick Bostrom
"The Art of Learning" by Josh Waitzkin
"More Than Two" by Eve Rickert and Franklin Veaux
"The Power of Vulnerability" by Brené Brown
"Nonviolent Communication" by Marshall Rosenberg
"A Thousand Brains" by Jeff Hawkins

4/5

"The Selfish Gene" by Richard Dawkins
"The Extended Phenotype" by Richard Dawkins
"Guns, Germs, and Steel" by Jared Diamond
"Why Nations Fail" by Daron Acemoglu
"The Little Book of Common Sense Investing" by John Bogle
"Statistics Done Wrong" by Alex Reinhart

3/5

disclaimer: some are very dependent on the period when I've read them, so definitely not a general "books you must read" type of list. 2/5

t.me/danokhlopkov asked to list the books which have transformed your understanding of the world. went through my audible and goodreads to jog my memory, here are the results. 1/5

#BookRecommendations #Reading #Books #PersonalDevelopment #Science #NonFiction
#BooksToRead #Booklist #BookThread

to summarise -- I believe within 1-2 years we'll see tools and products (and maybe even a general shift on the Internet) towards prioritising human-generated content. We'll get the good of both worlds -- easy access to high quality expert content and AI to help us understand and use it.

and -- I think it is harder to make steps in this direction than to make the metrics of "AI-generated-ness" less orthogonal to "usefulness". (mostly because as a general principle it is easier to make critics than creators).

a more geometric way to phrase this -- how orthogonal are the metrics of "usefulness" and "AI-generated-ness"? my intuition is that they are not that orthogonal. i.e. if you try to lower your "AI-generated-ness", you'll most likely get more "useful" content.