int3

> After all, most ML in industry today seems to be lightweight models applied to heavily engineered features

I assume "lightweight models" are those that don't have too many parameters, and "heavily engineered features" mean that the data fed into the model has undergone significant pre-processing via potentially complicated UDFs -- hence the motivation for the project. Is that right?

> Quokka is an open-source push-based vectorized query engine ... it is meant to be much more performant than blocking-shuffle based alternatives like SparkSQL

Does anyone have pointers to what push-based vs blocking-shuffle engines are? Any good papers?

> It should work on local machine no problem (and should be a lot faster than Pandas!)

So I understand why Quokka is faster than Spark, but I'm a bit uncertain as to why the author is also making a comparison with Pandas on a single machine. Is it because the streaming pipeline design means that Quokka can better take advantage of multiple cores?

> If there's no semaphore, good luck crossing

Do they actually call traffic lights "semaphores" over there?

Also, Google eventually lifted this limit on newer versions of Android because other apps were running into the same issue. It's not just FB :)

> People who keep on gaining weight until they die must have a broken thermostat set on "high". It can't be psychological.

> Because conscious thought is not a replacement for automatic regulation.

I skimmed the article, but only because I've already read an entire book on the topic. I understand that there are hormonal effects on weight. That's no reason to throw your hands up and declare that you have no agency in the matter.

I used to be slightly underweight when I was younger, and could eat endlessly with no consequence. These days I have to consciously watch my calories. It was a bit difficult at first, but I've gotten used to it. Conscious thought is not a replacement for automatic regulation in itself, but you can consciously work to form new habits.

How is this different from hedge funds who do the same thing with equities, and outperform the retail investor as a result?

The advice to retail investors is "buy and hold", and I think the same applies here.

> The semantics of LLD’s archive handling are different from the traditional Unix’s. You can observe it if you carefully craft archive files to exploit it. However, in reality, we don’t know any program that cannot link with our algorithm so far, so it’s not going to cause trouble.

Because a society where we can have less partisan opinions and more fruitful discussions is a better one.

But I'm not attacking you, and I don't think most of the other commenters are, either. You expressed a disagreement with its thesis and explained how identity has been useful to you. I just pointed out that perhaps there is a way to achieve the same goals without expanding your identity. I was trying to have a discussion about the essay. But it appears I misunderstood, and you did not actually wish to engage in a discussion, so I will stop here :)

I can respect that "all other things being equal" does not hold here. There are always tradeoffs. But there are real benefits to not identifying with things (as the essay discusses).

Also, LLD in trunk can link for Apple Silicon too :)

Also the marginal value of forests and wildlife have changed significantly over time as they have been depleted. The tradeoffs that made sense then don't make as much sense now.

Pretty sure it's a pun for the French `tout suite` -- "right now / ASAP". Pretty clever actually given how much value people place on immediacy on social media...

- Barring very degenerate inputs, prefix tries are usually faster than `std::set<std::string>`. And for longer strings, prefix tries are definitely faster than `std::unordered_set<std::string>`.

- The same ideas that apply to strings apply to bitvectors too, so we can handle arbitrary binary data the same way. (This gives us things like Patricia trees.)

- They're great for implementing immutable sets. In fact most functional languages use some variant of prefix tries for their set/map data structures.

And some real-world use cases:

- Symbol lookup in mach-O (macOS) binaries use prefix tries

- One of the fastest string sorting algorithms (burstsort) uses prefix tries

- Program analysis / abstract interpretation really benefits from immutable sets of bitvectors with fast union operations

- Pretty sure every typeahead / autocomplete implementation uses some kind of trie

- And for something a bit more esoteric: scrabble/boggle AIs use them too