rspeer

25 ms·

181.
by rspeer 8y ago
Rent control isn't the solution to everything. If housing is consistently available at a discount to market price, it will fill up and stay full. New entrants to the housing market, such as the younger generation, are still screwed. It
182.
by rspeer 8y ago
It sounds like you may have missed one of the GP's points when you're blaming renters for not voting. I'll reiterate it with a specific example, although I am on the east coast. Lots of people want to live in Cambridge, MA. C
183.
by rspeer 8y ago
You'd think it wouldn't take so much effort to defend the claim "if you build enough housing, people won't be homeless", but it does. So let's go with a prominent example: Tokyo. Neighborhoods in Tokyo have no
184.
by rspeer 8y ago
This sucks for everyone, but I understand why Facebook did that. A very real threat against users who delete their accounts is that a scammer will re-create the account, and re-friend your former network, in order to defraud your friends. T
185.
by rspeer 8y ago
In addition to what novia said about your first sentence, your comment is so weirdly off-base. You have very strong opinions about pen computing, but where in reality are they coming from? No, we are not talking about the 1940s, why would a
186.
by rspeer 8y ago
> The implication is that police and courts are racially profiling or being racist. But you shouldn't believe that without evidence. It sounds like not even all the evidence that exists would make you believe it. Racial profiling is
187.
by rspeer 8y ago
Language codes made up off the top of someone's head are my peeve. If anyone else wanted to support Kansai dialect, they wouldn't necessarily (and probably shouldn't) make the same decision to pretend that it's a country
188.
by rspeer 8y ago
BCP 47 language codes [1] have three major parts: the language, the script, and the region. Canonically, the language is two or three lowercase letters, the region is two capital letters or three numbers, and the script is four letters in t
189.
by rspeer 8y ago
> With Rust I've been on the other side of that exchange, too. That is really the only region of Stack Overflow where I've seen the pedantry and unhelpfulness that others complain about---mostly from one person actually. And I
190.
by rspeer 8y ago
I've also had to parse Wikitext. The fact that there are 54 parsers in various states of disrepair listed here (and I have written a 55th) is not because people really like reinventing this wheel; it's because the complete task is
191.
by rspeer 8y ago
Let me spell it out for you. It's fine to be unaware of a particular trend, but it's not fine to continue to be unaware and to disbelieve it when you read an article that directly points it out. * Neo-Nazis invented the "echo
192.
by rspeer 8y ago
Oh my god, you can't possibly be that sheltered.
193.
by rspeer 8y ago
HTML is pretty repetitive, but if you want to archive HTML data, you don't get to redefine what HTML is. Compression is useful.
194.
by rspeer 8y ago
So you're on board with destruction as long as you get to do it your way? This is going to be a radical idea, so bear with me here. What if instead of destroying civilizations, we didn't destroy civilizations?
195.
by rspeer 8y ago
Everyone capable of understanding this thread knows what overfitting is, and the difference between training and test data. That is the topic we are talking about. Because test data is unseen at training time, it cannot possibly affect the
196.
by rspeer 8y ago
I'm glad that Google is part of this conversation, and they're now applying tests for bias to new models that they release. (Some of their old models are pretty awful.) If you want to see a further example, in the form of a Jupyte
197.
by rspeer 8y ago
People who make word embedding models are not doing it to find out whether men are better than women. That is not the analysis they are doing. You always have to look at your results with a critical eye to recognize when something has gon
198.
by rspeer 8y ago
> A model may never reduce any probability to zero. That's the entire point. You can't teach a model not to reduce probabilities to zero just by showing it more data. There will always be more unseen data. And you, a human, kno
199.
by rspeer 8y ago
That still doesn't mean that your training data is right for the predictions you're going to make. Here's an example detached from any sort of politics. If you are training a language model, the maximum-likelihood model is on
200.
by rspeer 8y ago
If you understood the scientific method, you would understand that you always need to account for bias. Just because something comes from data doesn't mean it's objective.
201.
by rspeer 8y ago
Ethereum simply promising to switch to PoS, at some undefined future time when it's extremely convenient for them, provides absolutely no moral redemption. They have to actually do it.
202.
by rspeer 9y ago
Can you blame them for not being happy with that? Why put "superior" in quotes if they were doing a better job? It sounds like Mobileye knows what it takes to make self-driving work given present technology, and Tesla didn't
203.
by rspeer 9y ago
2to3 is a bad and outdated tool, please don't recommend it. You want http://python-future.org/ .
204.
by rspeer 9y ago
Reification is at least part of the solution to provenance, but none of the Semantic Web ontologies you've ever heard of use reification, and there is no upgrade path to make an ontology based on plain un-reified triples use reificatio
205.
by rspeer 9y ago
Most technologies that were specific to the "Semantic Web", such as OWL and SPARQL, failed to scale and failed to solve realistic problems, and therefore died. (I always maintained that running a SPARQL endpoint amounted to runnin
206.
by rspeer 9y ago
The Semantic Web as originally imagined is also incompatible with privacy, and with malicious fake data. So is the decentralized Web you imagine, I think. Let's keep using FOAF as an example. The facts about who knows who in FOAF are j
207.
by rspeer 9y ago
It's true that the OP page skips a lot of characters (including all the Han Chinese characters -- it occurs to me that I have no idea if Unicode says anything about how these characters should be sorted). But the page you're linki
208.
by rspeer 9y ago
YouTube video recommendations, which converge on recommending conspiracy theories and shocking videos to easily manipulated people, especially children. I am not worried about literal paperclip maximizers, but this may be the closest real t
209.
by rspeer 9y ago
I don't think you answered 3pt14159's question. What if you publish the research and nobody who can do anything about it cares?
210.
by rspeer 9y ago
It saddens me that IBM is setting the expectations so low for natural language processing that they're basically reducing it to a Myers-Briggs test.

More ›