fnl

9 ms·

61.
by fnl 9y ago
What about YouTube? I hear people can make a living off that with 100s of thousands of viewers or more - note sure that's Funny or Die's league, though?
62.
by fnl 9y ago
While I love Python, I think all UIs I've ever seen that are based on it feel pretty clumsy. The QML+JavaScript path Qt is taking makes more sense to me - still snappy, but it unifies the web, app, and desktop worlds.
63.
by fnl 9y ago
Those JS ninjas could just go with QML and JS. Shouldn't be any difficult for them. And much leaner and more efficient than an Electron app, too. http://doc.qt.io/qt-5/qtqml-javascript-expressions.html
64.
by fnl 9y ago
You won't be using Python with QML at all, but rather JS - so you still can enjoy the speedy development vs. a traditional Widget app, and you get a far more efficient runtime, too. http://doc.qt.io/qt-5/qtqml-jav
65.
by fnl 9y ago
Not that most companies would care, as dev cost seems to be all that matters these days, but the upsides are much faster and leaner applications than what you get from Electron.
66.
by fnl 9y ago
Are you aware of QML, which is part of the Qt framework? https://en.m.wikipedia.org/wiki/QML
67.
by fnl 9y ago
As nobody mentioned it yet, I think: I use the smart referrer plugin. https://addons.mozilla.org/en-US/firefox/addon/smart-referer...
68.
by fnl 9y ago
Maybe so. Yet, I can't find any features in Drake that I don't get with Make, too - in fact, it looks to be rather the opposite. Indeed, for some of the tools I listed, they barely have any more functionality than I'd get out
69.
by fnl 9y ago
To amend, why I'm even bringing this up, what worries me about Drake is this: https://github.com/Factual/drake/pulse/monthly Its GitHub pulse is - dead; For two years now. Makes me think one of the other
70.
by fnl 9y ago
Given the number of tools around, and as this book promotes Drake, has anyone got "comparative experience" with some of the following tools: Cookiecutter: https://drivendata.github.io/cookiecutter-data-science Dat
71.
by fnl 9y ago
> If you're a book editor and the author adds, edits, or removes a section, you need to update the index But all the page references would be wrong, too, after the author updates the text. That would not happen with a DB index and i
72.
by fnl 9y ago
A very relevant "must see" if this stuff interests you: Kate Crawford's NIPS 2017 keynote on bias in machine learning https://nips.cc/Conferences/2017/Schedule?showEvent=8742
73.
by fnl 9y ago
Bad title (this is all about text classification/mining, not NLP), but a very nice introduction at that. Maybe a tad optimistic - I'd never even consider applying a classifier with 80% accuracy to the Twitter firehose (unless extr
74.
by fnl 9y ago
I'd agree - and also, it should be made clear that a Twitter classifier with 80% accuracy is probably a long way from something that you would apply to the actual Twitter firehose. But the article is a very nice introduction to text
75.
by fnl 9y ago
Well, the 80% accuracy the author reaches isn't anywhere near what you need in production for text classification. Imagine classifying all of Twitter with that... I don't want to make this look bad - the author did a great job a
76.
by fnl 9y ago
Good point - there we have our margin for exceptional language stability: ~20 years, set by both Java and C++. I guess Go v1 still needs a few more years to even be half that old... :-)
77.
by fnl 9y ago
Python 3 is older than Go, and PHP 5 is even older than either, at least as for Go version 1. C++ is fully backwards compatible, as is Java. I don't see this point. And, we'd first need to see if a hypothetical Go 2 will be backwa
78.
by fnl 9y ago
What part of my statement makes you think I said anything to the contrary of that?
79.
by fnl 9y ago
Java and C++ are virtually fully backwards compatible. Java 1.6, 10 years old, is still commercially supported. C++ compiler support depends, but can go back far, too. Golang only seems to support the latest two [minor] versions, however (s
80.
by fnl 9y ago
I don't see why this isn't just another example of a broken system - copyright is wrong, always. It was invented by the Inquisition, after all.
81.
by fnl 9y ago
Well, the US wasn't always like that. But if I look at the prison population statistics, someone seemed to have figured out how to make big bucks off the prison system sometime in the early 80s...
82.
by fnl 9y ago
Given the OP says he was 22 when getting hired the first time again, I doubt it can have been a serious crime, unless he had committed it at a very young age, which would put everything into another light, too.
83.
by fnl 9y ago
Not wanting to dampen your high hopes, yet, the rules of Go seem a lot simpler than the rules and grammar of the current hypertext markup "language", particularly if taking the "browser dialects" into account, which are
84.
by fnl 9y ago
The upside here being that societal evolution seems to be accelerating, just as nearly everything else.
85.
by fnl 9y ago
Excellent summary of the core issue. Though I think there should be room for some nature vs. nurture debate. Exclusively attributing this behaviour to our genes seems a bit like the "lazy excuse" to me. I don't see why proper
86.
by fnl 9y ago
I don't think you need to out-complete these systems in processing speed. Even if you can only do 10 documents/hour, that probably can be scaled to handle a few million. Where you need to out-complete cTakes & co (on a large m
87.
by fnl 9y ago
Yet proving that reliability in a fully automated fashion is hard. And it's even harder/more costly when the core problem is reproducibility...
88.
by fnl 9y ago
You call your problems "low hanging fruit", yet they are real problems that multi-billion dollar enterprises are facing and know of. If they really were as "low hanging" as you say, don't you think they'd have
89.
by fnl 9y ago
Perfect answer. This is exactly the case, from Swanson's seminal text mining in biology paper on fish oil in the 80's with Arrow, to modern tools like Leach's Hanalyzer: You need to focus on a very narrow target application o
90.
by fnl 9y ago
Designing a program that can read (all) papers and understand their content conceptually, then comprehend the question it is asked, and finally reason about all that to generate a sensible output is equivalent to creating an AI (a real one,

More ›