rspeer

26 ms·

211.
by rspeer 9y ago
That makes a lot of sense. Ultimately, credit card breaches are merely inconvenient to consumers, because fraudulent charges can be reversed. It's the card companies who care, because they eat the cost.
212.
by rspeer 9y ago
If you "democratize social data", you should expect privacy to quickly get out-voted.
213.
by rspeer 9y ago
I don't know why you'd think I'd be surprised that MailChimp has customers who use it for its purpose, which is marketing. It seems you're saying that MailChimp should be entirely neutral and should let you promote anyth
214.
by rspeer 9y ago
If you're paying to send people "Ethereum news this week", it's because you have a financial interest in promoting certain things that happen on Ethereum. There are plenty of places you can discuss Ethereum without payin
215.
by rspeer 9y ago
Fill me in here: if you're really not selling or promoting anything, what's the use case of MailChimp? MailChimp didn't make all other mailing lists and forums stop existing. It's just the service that's optimized f
216.
by rspeer 9y ago
Wow, great to see how responsible Salon is being with their ad network, after they decided to inflict cryptocurrency-mining malware on their visitors who use ad-blockers. It's like the Forbes thing, but stupider. salon.com should be co
217.
by rspeer 9y ago
Okay, but a lot more people own pets than own cattle. In the computer sense as well.
218.
by rspeer 9y ago
The quip by Babbage is clever, but Babbage was too early and didn't know what we know. He could not anticipate the digital abstraction, error-correcting codes, compressed sensing, and machine learning. Modern computing systems have com
219.
by rspeer 9y ago
> You can't remove it with a software bandage Turns out you can. It just takes effort. And you should, because amplifying shitty things about people is not the ideal state of software.
220.
by rspeer 9y ago
The page is definitely poorly engineered, but the font looks nothing like Arial or Helvetica.
221.
by rspeer 9y ago
It kind of sounds like you're trying to think of tax law as some sort of CTF programming game where you're supposed to find the exploit. That's definitely not what tax law is. You should report your taxes as honestly as possi
222.
by rspeer 9y ago
I use Spotify and never connected it to Facebook. It's not actually required. They just make the "social" features of Spotify super hard to use if you don't connect to Facebook, which... doesn't matter.
223.
by rspeer 9y ago
This is just being atavistic for no reason. The telephone system is a horrible UX. I hardly ever use Facebook, but I'm definitely not replacing it with the telephone.
224.
by rspeer 9y ago
The current Ubuntu long-term support release is Xenial, which comes with only Python 3 installed. If you're talking about Trusty, its support is ending fairly soon. On Xenial, you have to separately install an optional package for Py
225.
by rspeer 9y ago
If a group has a security policy that makes it hard for them to install Python 3, of course there's no way in hell they're installing Docker .
226.
by rspeer 9y ago
Yep, so there's a lot of work to do. Bolukbasi, Cheng, et al. will straightforwardly admit that their model only applies to the one axis of gender bias. Fortunately there are other people working on this too, and I am one of them. The
227.
by rspeer 9y ago
You've given an example of where a human decision in a data-processing pipeline harms the result. There are many examples where human decisions are a benefit. Look at every other decision in a successful speech-recognition pipeline for
228.
by rspeer 9y ago
I'm glad we could at least get you to read it, but you're still not caught up. The effect is many orders of magnitude larger than floating point error. The notion that implicit bias could be corrected by quantization (not "qu
229.
by rspeer 9y ago
That's possibly a fair description of the kind of as-late-as-possible de-biasing that I find insufficient. I think there is not nearly enough research into de-biasing at training time, perhaps because it would be discouraged by the rew
230.
by rspeer 9y ago
Ah. In that case, I agree with where you're coming from, but not on the details: I think there's no reason to believe that bias from multiple layers of ML end up neatly rolled up into one wad of bias that can be extracted at the e
231.
by rspeer 9y ago
Ah yes. This approach to ML bias is popular in some large corporate research groups: - Say a lot of nice things about fairness - Talk about the importance of de-biasing and doing it as well as possible - Show (legitimately) that the farther
232.
by rspeer 9y ago
The corpus alone can't be sufficient. The corpus displays associations that are disproportionate for how we want to use it, because corpora are from the past . But the point of ML is to build something that will be used in the future
233.
by rspeer 9y ago
Where did you get the idea that data , of all things, is objective? Data is shit! Data is the noxious raw material that we have to process with great difficulty into something useful! If you are familiar at all with machine learning, you s
234.
by rspeer 9y ago
Thanks for being willing to discuss it. And sorry if I'm complicating your task. But any interesting release of NLP data has the potential to affect the way the field progresses, so take it as a compliment that I consider this an inter
235.
by rspeer 9y ago
There may be a way to build it into the loss function so that it happens before the quantization, right? (and holy crap, look how fast the HN conservatives are getting to my comment)
236.
by rspeer 9y ago
It's interesting and slightly uncomfortable that the illustration for similar words uses the word "man" as an example, given the gender biases that result from learning word vectors solely from word distributions. To explain,
237.
by rspeer 9y ago
Geocoding is probably simpler if its users aren't "customers". Lots of systems have checksums. It appears to me that what3words's complexity is designed to make it intentionally difficult for someone else to reimplement.
238.
by rspeer 9y ago
The other advantage of this over w3w is that it the encoding makes any kind of sense. w3w's feature where it moves an address onto a different continent because you forgot to pluralize a word is the worst idea for an address encoding.
239.
by rspeer 9y ago
You've shown that you misunderstand the term. Your misunderstanding includes nonsense examples like "those dictionaries are a dumpster fire". Nobody else would say that. The "dumb-store-burning" thing is something y
240.
by rspeer 9y ago
If you think the phrase "dumpster fire" is the sum of its parts, then you don't know what it means. And if you don't know what it means, a dictionary can help.

More ›