Please help contribute to the Reddit categorization project here


    + friends - friends
    24,850 link karma
    225,664 comment karma
    send message redditor for

    [–] Sex education, made fun! Astrokiwi 2 points ago in CasualUK

    "Its's Herpes Genititahlis, not Herpes Genitalis"

    [–] Redditor explains common core math thought process Astrokiwi 1 points ago in bestof

    Because we work in base ten, so if you can reduce something to a power of ten, it simplifies the problem.

    [–] Relevant. Astrokiwi 8 points ago in TheSimpsons

    He then went on to work for Dr House, and when that fell through, became an unethical therapist who dated one of his court-mandated clients.

    [–] Relevant. Astrokiwi 24 points ago in TheSimpsons

    Didn't he have a PhD in computer science? He made a perfect tick tack toe AI on punchcards

    [–] Redditor explains common core math thought process Astrokiwi 2 points ago in bestof

    I don't know the details of how it's supposed to be taught. But I'd still split it up - e.g. 9+1=10, so it becomes 50+97+180 and so on

    [–] Redditor explains common core math thought process Astrokiwi 3 points ago in bestof

    It's "maths" in the UK, Australia, and New Zealand, and I am cute thank you very much

    [–] Well thank you, random villagers from the North Pole. Astrokiwi 9 points ago in civ

    Aurora Borealis! At this time of year? At this time of day? In this part of the country? Localized entirely within your kitchen?

    [–] Redditor explains common core math thought process Astrokiwi 112 points ago in bestof

    This is how people who are good at maths do it. But people who weren't good at maths were not directly taught this until fairly recently. You had kids who figured it out for themselves, and kids who felt that maths was black magic and memorisation.

    [–] Deadpool 2 poster art by Rob Liefeld Astrokiwi 1 points ago in comicbooks

    Because 12 year old boys in 1991 didn't care about accurate anatomy and mature story telling. They wanted big dudes with big guns saying badass one liners.

    [–] Can confirm. Astrokiwi 3 points ago in ProgrammerHumor

    Fortran is great for when numpy doesn't quite do exactly what you want. There are times when I would even choose Fortran over C or C++, just because a few things can come out a bit tidier in Fortran

    [–] One of our iconic species Astrokiwi 3 points ago in newzealand

    For some reason their enclosure was in a building that was mostly the Commerce department I think.

    [–] One of our iconic species Astrokiwi 2 points ago in newzealand

    If you had another one, they would be Ruatara

    [–] Anyone else find Ubisoft games exhausting? Astrokiwi 1 points ago in truegaming

    This made Oblivion almost unplayable to me. Morrowind and Oblivion both had a pretty unbalanced levelling system, where non-optimal choices at early levels (when you're still learning the game) could hurt you a lot in the long run. But in Morrowind you could get around it by just leveling up some more. On the other hand, the enemy leveling in Oblivion meant you'd always be playing catch-up. It wasn't so bad in Skyrim because the leveling was more streamlined and poor early choices were not so catastrophic.

    [–] Invincible Iron Man vol 2 #1 Astrokiwi 6 points ago in comicbooks


    [–] Invincible Iron Man vol 2 #1 Astrokiwi 14 points ago in comicbooks

    As a huge Bendis fan, this was okay, but didn't really go long enough for Bendis to do his long-game character development properly.

    [–] Invincible Iron Man vol 2 #1 Astrokiwi 23 points ago in comicbooks


    [–] Progression of my Master's Thesis [OC] Astrokiwi 1 points ago in dataisbeautiful

    Yeah, I guess what I was trying to get at is that you generally only delete stuff if you're trying to make things more tight and concise (and not for e.g. fixing errors, which usually ends up taking up the same number of words) and that's not usually a big focus for a thesis.

    [–] [HFM] Suez Canal decision problem Astrokiwi 1 points ago in paradoxplaza

    Related question: is there a name for this type of data format, and are there standard ways to parse it in popular languages? I've been thinking sometimes about trying to make things I write a bit more scriptable and less hard-coded. I know Civ IV uses XML, but it seems like JSON is the "cool" one to use now. Is this some proprietary thing that Paradox has set up?

    [–] Machine Learning can predict evolution of chaotic systems without knowing the equations longer than any previously known methods. This could mean, one day we may be able to replace weather models with machine learning algorithms. Astrokiwi 1 points ago in Physics

    The problem is that you can only match things in a statistical way. You can run a cosmological simulation and compare your simulated sample of galaxies with the observed sample, but you can't make and test predictions for a single galaxy, because the time-scales are long enough that you essentially only have a single frozen snapshot per galaxy. This means that you can't get fine constraints like you can in meteorology. They can say "our models predicted this bank of clouds would go here, but in reality in went there". We can't say "the SED of this part of the galaxy evolved to this in our models, but to that in the observations".

    So, because we can only compare statistical samples of galaxies rather than individual galaxies, we can't constrain the full 3D evolution of a galaxy - we can only constrain the general bulk properties of a sample of galaxies. This just gives you far too much degeneracy to play with, and not enough to train an ML algorithm. So we have to build models "by hand", and, as you say, this is a pretty tricky and difficult process.

    Of course, the other part is just the time it takes these simulations to run. You can't really do an iterative process like ML if each simulation takes 6 months on a large cluster.

    [–] Progression of my Master's Thesis [OC] Astrokiwi 5 points ago in dataisbeautiful

    "Four people" = "your supervisor and up to three examiners" basically.

    But in my experience - I finished my PhD about 5 years ago, and I'm a postdoc working in astrophysics simulations - a PhD thesis is just too big and broad to be useful as a reference or as a way of presenting original research. It's a combination of an elaboration on research articles you published during your thesis, and chapters that will form future papers you publish. (Although sometimes you can do a "paper staple thesis", which is three published articles, plus an introduction and conclusion). Nobody reads PhDs in this field, but people do read research papers that came out of PhD work. I suspect it's very similar in the pure computer science field.

    And yes, you do need to present yourself clearly. But there's so many things you have to do well that you do need to prioritise a little. It's just a much better investment of a PhD student's time to try to craft a journal article as well as possible instead, to catch people's attention, disseminate their research, and prove their worth as a researcher. With the PhD thesis, it does need to be reasonably well presented, but you could spend months or more editing it down to something beautiful, without really changing your changes of passing your examination. A lot of the examiners might not care about that sort of stuff anyway - remember that the discovery of the Higgs Boson was announced in Comic Sans.

    [–] The difference between the USA and New Zealand Astrokiwi 2 points ago in videos

    In New Zealand, we tend to call all types of squash "pumpkin". So when Americans talk about "pumpkin pie", we're imagining a butternut squash being put in a pie, which is something you'd probably have meat & potato in too.

    [–] Progression of my Master's Thesis [OC] Astrokiwi 9 points ago in dataisbeautiful

    A thesis has a lot more detail than a peer reviewed research article. You can record almost everything you did in great detail. It's also only going to be read by like four people so you want to concentrate on including everything you can to prove you deserve the PhD or MSc, rather than spending a lot of time working on flow and clarity.

    [–] Anyone else find Ubisoft games exhausting? Astrokiwi 3 points ago in truegaming

    Horizon Zero Dawn kinda has this. Different regions definitely have different difficulties, and while you can technically go anywhere once you get out of Tutorial Canyon, you're going to have a tough time in much of the map until you have the high-level equipment, and you've levelled up both your in-game skills and the actual player's abilities and knowledge of the game.