wallscratch

Can anyone comment on how efficient the Warp code is compared to manually written / fine-tuned CUDA?

They literally just saved the world, and we’re not going to give them any credit for it…?

I’m sorry, but what is the play on words in the title “Attention is all you need” here?

Going to echo the suggestions for the Art of Problem Solving books, particularly I recommend the contest books (vol 1 or 2). Several very talented people have said to me that these books taught them how to think. Maybe a bit exaggerated, but they’re very good.

That’s true, but my view is even if the original article is unbiased, the selection process by which it was chosen to appear on this periodical with the reach it has introduces some statistical bias.

Does anyone have suggestions with where to find more exercises? I notice the book pdf has like ~38 total, but was curious where one would find more.

Human Revolution was wonderful, but Mankind Divided had such bizarre pacing… I was ~15 hours into Prague, waiting for the transition to the next locale like in all of the other Deus Ex games, and then the story ended… sigh

You mean Pembroke? Probably more like 5-10 kids a year. Even my public suburban kc high school had 7ish kids going to ivies my year.

Could someone explain why refactoring is so much easier in functional languages?

I’m confused — why can’t the sex workers just claim to be independent masseuses or something else innocuous like that?

Can you give an example of a system that’s in place to ensure workers are near max capacity most of the time?

Linux kernel has a bunch of gotos, primarily used to just jump to exit / deallocate on kmalloc failure, without having to unnecessarily complicate the control flow to handle that.

This is super cool, and helps fill a hole I had always been frustrated by during my own self-education.

I am curious to know approximately how many hours the courses take, and if you can share anything about which future courses you have in the works?

This take seems a bit pessimistic about the longevity of society to me. If we still have tens or hundreds of thousands of years to go, just going with whatever was best by 2020 seems a bit ad hoc. Not that I know the future.

It should be totally deterministic with the use of random seeds, I think.

I thought modern computers are primarily cache limited and even running an os entirely on ram (on a system with no hdd/sdd) isn’t that much faster. I would be curious to see some data.

Why? Aren’t all regular expressions equivalent to some finite state machine?

I went to a mid-tier state school and my friends were just people I happened to be lumped together in housing. Is it "optimal"? No, but I enjoyed it. My friend crew consisted of a stoner, a very religious dude, a meat head and a Brooklyn hipster. I don't know, I liked it like that. None of them really taught me anything (wtf does an 18 year old know), but it might be different in technical fields. Both have trade offs.

Yeah, my ug friend group was similar. It’s completely valid to have different preferences in what one wants out of life. However, I personally did feel much happier / interested later on in life being in friendgroups where everyone had some kind of serious career aspiration.

I strongly encourage you to contact MIT about your situation and family’s situation and try to negotiate out more financial aid.

As someone who went to a flagship state school for undergrad and ivy for phd…

A) Most of the people I know who turned down higher-ranked schools for lower-ranked ones because of money regret it. You will make a lot of life-long friends in college, and you will just be exposed to a different caliber of person on average at mit. Random people you meet through friends of friends at brunches or happy hours will be weirdly accomplished and teach you things.

B) Your analysis seems to hinge on doing a phd at a top-n school. What if it turns out after a few years of college that you don’t want to do a phd after all? Then instead of being either mit phd, or mit bs, you are z school bs. This may not be terrible, but not optimal.

D) On the other hand, I think the differences in career outcomes on average are small, although I wouldn’t be surprised if the probability of starting a company with any given level of success x is 10 times higher for mit alum than Z school alums. Anyway, median mit cs alum has some faang-y job throughout their careers, and these companies all know/understand that many top students can’t all afford top colleges, and so they recruit from state schools as well. So the tail outcomes can be quite different, mean/median isn’t that much.

E) For careers like management consulting or investment banking, some top firms only recruit at top-n schools. However, eg mckinsey even does on-campus recruiting at places like ut austin or georgia tech now, so then it doesn’t matter. You just need yo be at an on-campus recruitment target school.

In my opinion, the lifetime of friends and network effects is most of the benefit, and not to be underestimated. You only live once.

I rewatched lotr after having seen GoT, and I was struck by how simplistic the good vs evil narrative of everything was.

As someone who went to a state school for undergrad and an ivy for phd (and took some undergrad cs classes there), and I can say that yes, the undergraduate classes are definitely much harder at ivy league schools, at least in stem.

Sure, but I don’t think that devalues his criticism.

Do you remember what library it was?

There are a few techniques:

Patch clamp recording: you are literally sticking a needle into a single neuron and measuring it’s activity

EEG/MEG: using electrodes to measure activations of many thousands or millions of cells

At the day job, its Python everywhere. I've been told people would laugh at using Perl. Kind of a shame, as I see multiple pages of code that could be easily reworked into a far more readable, easy to reason about and comprehend small set of Perl code.

Could you give an example? I would be curious to better understand python’s limitations.

Any thoughts on anvil? It’s a full-stack python framework. I think it features a python to js transpiler:

https://anvil.works/blog/client-side-python

Chocolate also has caffeine, which may be what’s keeping you up.

It’s striking that every org and every team is hiring.

What’s up with wikipedia results being deprioritized?

Couldn’t there just be some google maps / waze API that jurisdictions can enter speed limit information into for different days and times of the week, and just have the car query that?