fnl

16 ms·

121.
by fnl 9y ago
The problem with maximizing efficiency is that it kills diversity and creativity. In my home country, Austria, everyone can study anything for free, and I think it has not had that negative impact either. Most people are reasonable and will
122.
by fnl 9y ago
Agreed. Nobody takes into account the total loss of social security in the US. Healthcare, retirement, child aid, education, .... If anything, I'd be upset if I were living in the US about other jobs not earning at least that much, to
123.
by fnl 9y ago
It would be cool to have some equivalent of a complete BLAS implementation in pure Rust, but I am not sure how sensible/economically viable that would be. But there do exist Rust creates built on top of existing BLAS libraries, e.g. h
124.
by fnl 9y ago
Interesting article, but too many uncertainties: 100 examples only, the 30 code bases w/o unit tests might have tests the author didn't check or see, and there is likely a cluster effect in this sample to be found. But interesting
125.
by fnl 9y ago
If working in C++ land, I'd agree this is the nicest approach. It does, however, require linking against a specific libpython version [1], unlike Milksnake. But I'm not sure that's a bad thing... [1] http://pybin
126.
by fnl 9y ago
Thank you for the very nice, educative article, Bruno! If performance comparison of counting character pairs really were the issue here, in addition to the already suggested numpy approach, an implementation I'd dare wager to be as com
127.
by fnl 9y ago
Well, a model you fine tune to your specific corpus/domain works even (in fact: much) better... And gensim there gives you the tools to build the best possible embeddings. But you do need a use case and an economic reward for the subst
128.
by fnl 9y ago
- Text generation and dialogue systems
129.
by fnl 9y ago
I'm always astonished how little mention gensim gets, considering that it can basically be used for all the listed tasks, including parsing, if you combine it with your favorite deep learning library (DyNet, anyone?).
130.
by fnl 9y ago
Hmm. We do get more information, obviously, but I am not sure that means the ratio of true or correct information over false or incorrect information has changed. Formerly, there have been far more myths and lore around that are easier to f
131.
by fnl 9y ago
Paxos is about ensuring consistency. So why would the original arguments hold for a write-available database like Cassandra - or about log/journal storage, then?
132.
by fnl 9y ago
> But in an era when the truth is so easy to distort, what should you believe? What makes you think that in some time before it was easier to find the truth?
133.
by fnl 9y ago
Does your explanation also hold when you look at read and write availability separately? I'd only see that to hold for the read "view" here. E.g., Cassandra is used because you can write your data almost always to Cassandra,
134.
by fnl 9y ago
Fully agreed. Running a smartphone and worrying about its security seems at odds. Rather, we should treat our "phablets" like public, insecure terminals, with "spy" sensors anyone can access given sufficient effort.
135.
by fnl 9y ago
That sounds like a great idea. So no more importing of implicits in Scala 3 to use them on methods of another package (or namespace)? That would make them quite a lot more "wieldable". What Odersky warns about is defining implici
136.
by fnl 9y ago
Right. "class fields" as opposed to "instance fields" had me confused.
137.
by fnl 9y ago
Sadly, I agree on your beef about implicits. They are a nice idea, but are far to easy to abuse to "hide" poor design.
138.
by fnl 9y ago
Indeed, if people would only use implicits the way Odersky recommends, I'd love Scala. The problem is, too many people do exactly what he warns about in this video. And so my love-hate relationship continues... :-)
139.
by fnl 9y ago
Indeed, Guice even recommends not to use (static) final fields. https://github.com/google/guice/wiki/Injections
140.
by fnl 9y ago
Would class fields not be a solution based on mutable state? Isn't that exactly the problem Odersky addresses in the beginning that your state can get mutated while your transaction is ongoing?
141.
by fnl 9y ago
NewsBlur, the reader I mainly use, can fetch that linked web page for you with a single click.
142.
by fnl 9y ago
the (good) old reader and newsblur are my favorites https://theoldreader.com/ http://newsblur.com/
143.
by fnl 9y ago
And you can push out news on RSS as they arrive, not just "weekly (newsletter) updates" or so.
144.
by fnl 9y ago
I'd say the life form that dominates this planet are single cellular organisms, i.e., mostly bacteria. They have run this place for billions of years, deciding who gets to stay in symbiosis with them or not and most likely will continu
145.
by fnl 9y ago
As said, with Putin, I think things have only gotten worse. And China seems to be settling on a benevolent dictator, too, these days. So at least recently, I don't think things are getting any better in those two countries.
146.
by fnl 9y ago
I never place a photo, martial status, or birthdate on my CV. Yet, I even got hired by Novartis (in Austria). If a German company of any size really were to sort me out sure to those "omissions" on my CV, it's probably for th
147.
by fnl 9y ago
Your assumption is that economic cooperation makes "evil" countries become more like the "good" countries they cooperate with. I don't think that holds true. Just take China and Russia, who have very strong economic
148.
by fnl 9y ago
Source(s)?
149.
by fnl 9y ago
> Even if you would need this money to keep your own wife and children alive? If I'd live under circumstances where I indeed had no other choice, maybe. Luckily, I live in Western society, so the answer to your question is "can
150.
by fnl 9y ago
How do evolutionary hyper-parameter optimizations compare to Bayesian ones? My impression always was that Bayesian optimization is more targeted, and therefore more efficient and ultimately finds optimal parameters faster. However, evolutio

More ›