CaptainNegative

Why is this a bad case for using the median, which is normally pretty resilient to influence from outliers?

If you truly like animals, you probably want to learn all about it and see what you can do to stop it.

Those definitely existed in the past, but there's been a dramatic uptick of terrible office etiquette in my current situation relative to pre-pandemic times, even though I'm in the same building working for the same company. I now expect any rando who sits next to me to enter a meeting.

I don't like working from home, but I hate mandated RTO because of all the antisocial behavior people bring into the office from their WFH days (loud meetings at their desks, brushing their teeth at their desks, Das keyboards, etc.).

Animal Rescuers, e.g. https://bemagazine.org/direct-action-everywhere-trial-finall...

The appropriate notion here seems to be Differential Privacy, which is a mathematical definition informally saying "a scrambling of the dataset that is information theoretically indistinguishable from that where one arbitrary person is added or removed". It's a surprisingly deep topic, with entire (very good!) textbooks dedicated to it.

PDF (entirely legal): https://www.cis.upenn.edu/~aaroth/Papers/privacybook.pdf

By that notion my home kitchen also has something for everyone because I have a tomato in the fridge. Where's my HN praise for being so culinarily flexible?

There are a few more options when generalizing beyond entrées, but still five of the items listed on that page are salads, with the other four being a side of beets, asparagus, artichoke, or edamame. Which, at double digit Calories, is only as much a reasonable meal for vegans as it is for carnists (that is to say not much of one).

"Something for everyone" is stretching it when they have one vegan entree: The Impossible Burger. The closest second option, if it even counts as a separate option, is the "Skinny" Impossible Burger, which comes with mayo that can be removed.

Nothing else is close to vegan, unless you're into ordering Chicken Lettuce Wraps without Chicken.

Sears is an amusing example given their mail-order history. Ultimately they shut down as a direct result of making no attempt to innovate or even mimic the innovators since those glory days.

Walmart, Target, Home Depot, etc. embraced the 21st century and are doing just fine. Others like Chewy only came about as a result of the new trends.

BB&B bought its way into bankruptcy by taking on debt to fund a multi-billion dollar stock buyback; external competition (including Amazon but primarily Walmart) was a secondary factor.

This reply feels a bit tangential. The amount of land taken to raise livestock is a small fraction of the land used to grow their feed. You discuss the lack of nuance with respect to the former, but OP was talking about the latter. The land taken for raising animals hadn't even entered OP's equation yet, and can only strengthen their point.

I think the confusion comes from the term free-range meat, which (at least in the US) is a marketing term far removed from what any normal person would consider "free range". It could be -- and often is -- just a barn with a fenced outdoor opening small enough that it amortizes to a few square inches of "free range" land per animal. Many of whom will never get there because they die of various kinds of nastiness before they make it to the opening.

The percentage of human-raised animals who are truly "free range" and roam grasslands in any meaningful sense begins with a decimal point. Cut the per-head subsidies and you cut the incentive to singularly maximize body count at the expense of all potential benefits of free-roaming animals.

The Collatz conjecture isn't a decision problem. The natural corresponding decision problem (given an x, does the conjecture hold when starting at all positive integers k less than x) is decidable in constant time.

We don't know what specifically the deciding machine is, but it is either the one that always returns Yes or the one that returns whether or not x is below some finite value. Either case can be solved by only reading a bounded number of bits from the input.

With some work, one can eventually generalize Collatz to make an undecidable problem; this is more or less the basis for FRACTRAN. https://en.wikipedia.org/wiki/FRACTRAN

The <pre> tag has existed since HTML 2 for displaying preformatted/whitespace-sensitive text, and HTML 1 had the mostly similar <LISTING> tag (plus <PLAINTEXT> which is a little different).

Since it had stickers, why not just move them around instead of smashing the clock?

There's a long-standing effort by Nation of Islam (the American hate group that is only nominally Islamic) to rewrite the history of slavery in the US as largely orchestrated by and for Jews. Nonetheless the actual numbers show that they "possessed far fewer slaves than non-Jews in every British territory in North America and the Caribbean" [1] though the reason why is still not well understood.

[1] https://www.jstor.org/stable/2674370

You're never concentrating the PFAS, so at worst you're bringing poor sap's new PFAS concentration closer to your own pre-transfusion levels.

So the question is whether they need the blood more than you did pre-transfusion. If they did, then you're still commiting a net positive. If not, then it serves them right for requesting a blood donation despite needing it less than did an individual who was ready to donate their own.

Unless the blood is going to research, in which case it is a net positive regardless.

Not even surrounding Paul with water emojis could stop the fire.

I'm not using it. Same goes for the rest of the junk operating system it's built for.

Not the person you replied to, but from a privacy perspective Apple won't allow Westerners to disable all network access per-app. Something that's been around in Chinese iOS releases for nearly a decade.

I'd love to use many offline note taking apps, but not if they're liable to randomly upload my notes to a potentially adversarial third party. And the actively privacy-hostile behavior of maintaining a separate branch just to disallow access to this feature is not appreciated.

(sorry, that's the sum of 2n+1 terms, but the analysis doesn't appreciably change)

Saying (2n)! ≈ 2^(2n)*(n!)^2 means he's approximating the central binomial coefficient (2n choose n) as 2^(2n). In reality, 2^(2n) is really the sum of all binomial coefficients (2n choose k) for k=0..2n. So clearly this is going to be an underestimate, as we're approximating the sum of 2n terms by only one of them.

However, since the biggest term in the sum of 2n positive numbers is clearly within a factor 2n of the entire sum, the ratio of the logarithms is at least log(2^(2n) / 2n) / log(2^(2n)) = 1-log(2n)/(2n) which converges to one pretty quickly.

And, in fact, a tighter analysis would show that the error decreases faster than what I outlined.

Many humans also evolved to tolerate ethanol but it's also bad for you unless you specifically need the calories.

I'm not sure even the first statement is entirely fair. As far as I can tell this result is indistinguishable from "a sufficiently powerful model can distinguish data generated by two different popsynth models".

I would have had more faith had they scrambled the labels and run the experiment of distinguishing one random subset of songs from another random subset. Anything greater than 50% would indicate the model's ability to overfit; I'd like to know how accurately their model would perform.

Netcraft now confirms: Beowulf cluster memes are dead.

The diameter of the cube group is 20, but speedcubers go for natural/memorized sequences of moves they can execute extremely quickly ("fingertrick") over technically shorter but awkward move sequences that require grip adjustments, have hard to visualize side effects, and so on.

Why is it Justine posts and seemingly only Justine posts that always get this type of comments? Do people regularly comment on the authors of other content, for better or for worse, and I miss it?

Why not use the MATE edition?

Heap sort's worst- and average-case running time are both provably within a constant factor of optimality (both in running time and number of comparisons), but it's either likely or straight up provably not actually optimal in any of the four categories.

Light bulbs are also nearly 100% efficient, minus whatever fraction of light seeps out through windows or similar. As is just about anything else electric. Unfortunately that heat is usually suboptimally distributed, and pales in comparison to the heat pump effective efficiency which as you mentioned can go well above 100%.

Not that I envy the ducklings thrown into the sewers for those videos, but I'd still rather be that duckling than a ferret into whom scientists have been stuffing colloidal tungsten to see if its LD50 differs from that in hamsters.