zwaps

28 ms·

331.
by zwaps 4y ago
The argument about a third party is trivial in my opinion. Someone responds correctly in Chinese, and the onus just falls on that element of the system to be conscious or not. It's another argument and I don't even see how this e
332.
by zwaps 4y ago
The point of the experiment is to think about the individual in the room. You can not say it's irrelevant, because it's the entire point. The system's response is trivial: Sure, if the room+person combination leads always to
333.
by zwaps 4y ago
Of course people can have preferences for race - just as they can preferences for height, weight, disabilities, gender, sex, and any other attribute. This is not discrimination, it is personal preference. It is an area where these propertie
334.
by zwaps 4y ago
Yes there is, and it is quite related to the article. But it is not practical - both men and women should rank their preferences and then an optimal matching can be found. In online dating, I would not expect people to have complete and tra
335.
by zwaps 4y ago
I fully agree with the article that discussing sentience - at this point in time, at least - is not very useful. Perhaps it never will be when it comes to statistical models. That being said, there are at more than two academic factions dis
336.
by zwaps 4y ago
Palm Pre was the best phone I ever owned edit: like, by a lot.
337.
by zwaps 4y ago
You don't seem to offer viable GPU options - or maybe the pricing page is not up to date? The GPU add-on cost is about 7.50$ per hour for a 12 GB K80 (which has 24GB, so you share it? Also, it's from 8? years ago?). As an add-on,
338.
by zwaps 4y ago
If the person creating this website are reading: Please do not pop up a "sign up for newsletter" pop up before one can even take one decent look at the article. There could not possibly a more clearer message saying: "we don&
339.
by zwaps 4y ago
I struggle to understand precisely what you mean by user experience and ‘often improved performance‘. Previously, there was no actual support for crucial features of the TPU related to data loading when using PyTorch, say. In turn, using a
340.
by zwaps 4y ago
Matlab solvers have been the most stable and performant for me, sadly not open source nor free.
341.
by zwaps 5y ago
In statistics, there's a joint distribution of x and f(x) to be learned. In this piece, there's a distribution over trained models that naturally depends on the initialized parameters. It's interesting when data is considered
342.
by zwaps 5y ago
ceteris paribus that is some pretty good drone video footage
343.
by zwaps 5y ago
Conversely, switching from Windows/Linux to MacOS is also very painful. For instance, why is there no way to minimize to desktop? Having to use Command+Button to go to beginning/end of a line? A horror for coding. My favorite pain
344.
by zwaps 5y ago
It should also be noted that (as a product of their time) each of these affiliated economists had a vastly different approach and view on economics. In light of this plurality, ignoring any quote from Schumpeter by default seems unwise. He
345.
by zwaps 5y ago
For my mathematical bird-brain, the most straightforward way to capture the situation you describe would be in terms of equivalence relations: Each person in your example defines an equivalence relation and each is "valid" and use
346.
by zwaps 5y ago
Decisions Theory (so, statistics and economics) have also always relied on the idea of equivalence (I believe the notion is the set-theoretic one) for purely practical reasons. So for instance (drawing on the other reply in this thread abou
347.
by zwaps 5y ago
The problem is that UI has reached its pinnacle with Win2k/xp classic. So if you are a designer, you still have to do your job, you gotta design something. But it is never going to be on that level, in efficiency, looks or usability. I
348.
by zwaps 5y ago
Lots of things run on Linux, which was Torvald's thesis project at a European university. Which, by the way, took 8 years... typical length for European's who really don't feel like graduating (back in the day). It's a b
349.
by zwaps 5y ago
There are plenty of people who are willing to forgo a lot to be able to do research. It is hard to overstate the difference in mentality. Few who are primed in academia would even think about salary as a main argument. Sadly, the trajectory
350.
by zwaps 5y ago
this will be costly for them
351.
by zwaps 5y ago
If my model is used to profile a given user such as to maximize revenue from them (my objective is generally increasing with a more accurate classification of a user to the degree that such categories are revenue relevant), does this model
352.
by zwaps 5y ago
Looks cool, but let me offer one point of critique: the website is very sparse. Indeed, too sparse for me to hack in my e-mail and risk getting it into yet another newsletter loop. I'd probably try to provide info on a few more benefit
353.
by zwaps 5y ago
So then, when will common ML frameworks work on Apple? I guess compiled Tensorflow works with some plugins or whatever, where afaik performance is still subpar. Apple emphasizes that they have this many Tensorcores... but unfortunately to
354.
by zwaps 5y ago
Okay so people with who are less smart tend, on average, to drink more? Jokes aside, I am shocked that a study without a causal identification strategy is still published in such a journal. The authors straddle the line in their discussion,
355.
by zwaps 5y ago
Relevant to mention that since today, mentioning the war carries an up-to 15 year jail term. Whether this excuses anything is not my point.
356.
by zwaps 5y ago
I went to a Michelin-starred vegan restaurant in Berlin (on the wall was a mural spelling the F-word and you had to walk through a warehouse to "find" the place) and it was really good. The wine was also very good. I wasn't
357.
by zwaps 5y ago
To give a more neutral context than the paper: Causality, identification and counterfactual analysis is an increasingly important topic in "all things statistics". For example, it is crucial in AB testing, "lifting" appr
358.
Heckman on Causality, Do-Calculus and Econometrics(nber.org)
2 points by zwaps 5y ago|1 comments
359.
by zwaps 5y ago
Crazy how evil Google is. Just wow. Since this runs entirely on the domain of the website, it can easily ignore your privacy rights, with Google more or less washing their hands clean of it. Indeed, if we take blocking trackers as expressio
360.
by zwaps 5y ago
This is simply not true. It's obviously the same company and will treat your privacy right the same way (i.e. as nonexistant) Further, you need a facebook account to log in (at least, for now) and it needs to be your real name (if you

More ›