Gimpei

21 ms·

211.
by Gimpei 7y ago
I'd argue that the upfront model leads to higher quality games. They're actually fun to play rather than simply being addiction machines.
212.
by Gimpei 7y ago
Also they dropped nvidia so I can't train models on the GPU. I'm switching to PC + ubuntu for by next laptop.
213.
by Gimpei 7y ago
Exactly. Artificially reducing the supply of diamonds doesn't reduce profit. It increases it.
214.
by Gimpei 7y ago
Not saying I disagree with you, but west coast cities do have temperate climates that I would think attract more of the chronically homeless. I'd much rather live on the street in LA in the winter than SLC.
215.
Tensorflow 2.0 Beta 0(github.com)
109 points by Gimpei 7y ago|43 comments
216.
by Gimpei 7y ago
Not sure if I would call him the greatest writer of the 20th century, but "Life and Fate" is fantastic. It's a shame that he isn't more recognized in English. Great that NYRB books published him again.
217.
by Gimpei 7y ago
Not so convinced by this. The "evidence" he links to is an opinion piece, not an empirical paper. The anecdotal evidence of prices going up in Toronto despite supply increases is equally unconvincing. For all we know prices would
218.
by Gimpei 7y ago
Psychopaths are actually supposed to be very good at reading other peoples' emotions. They just don't feel much empathy. Autism is associated with difficulties in reading emotions.
219.
by Gimpei 7y ago
I started off in journalism just out of college in 2000. Even then, future prospects were pretty bleak. People with 10 years more experience than me (plus PhDs in comp lit) were basically deciding on comma placement. After two years, I fled
220.
Is daycare bad for affluent children?(journals.uchicago.edu)
32 points by Gimpei 7y ago|15 comments
221.
by Gimpei 7y ago
Feels like this kind of thing would be useful for generating better jazz backing tracks. The programmatic backing tracks you get these days are good, but can be really repetitive. I wonder if there's some way to force the model to foll
222.
by Gimpei 8y ago
Looks like it was an issue with a script. I wouldn't be surprised if this isn't a bigger issue than we know in science. Good scripting practice is not taught in most programs. On top of that it's not really possible to write
223.
by Gimpei 8y ago
Not sure if your point holds if you read more of the article, which implies that plenty of economists are critical of wage and wealth disparities just for different reasons. My personal read is that Piketty doesn't bother to come up wi
224.
by Gimpei 8y ago
Will this help spark performance at all? I remember having to make a lot of GC tweaks in order to get some of the larger jobs to run.
225.
by Gimpei 8y ago
What was the philosophical difference between the two? Short-termism vs what?
226.
Tensorflow 2.0.0-Alpha0(github.com)
81 points by Gimpei 8y ago|38 comments
227.
Tensorflow 1.13.0(github.com)
3 points by Gimpei 8y ago|0 comments
228.
by Gimpei 8y ago
No model is perfect, but some models are better than others. I think it's fair to blame models that don't work very well. It's true that they are presenting neoliberalism as a boogieman. I don't even really understand wh
229.
by Gimpei 8y ago
I think because some people can't train themselves up through the gate. This is not because they can't figure out the leetcode problems, but because they have more difficulty dealing with the kind of pressure you see in interviewi
230.
New NBER Literature Review on UBI(nber.org)
1 points by Gimpei 8y ago|0 comments
231.
Engagement with Fake News Is Declining(nber.org)
2 points by Gimpei 8y ago|0 comments
232.
Facebook use decreases well-being(nber.org)
3 points by Gimpei 8y ago|0 comments
233.
Tensorflow 1.13.0 Released(github.com)
13 points by Gimpei 8y ago|1 comments
234.
by Gimpei 8y ago
23andMe told me that I was a quarter Italian. Seemed to have mixed that up with what I actually am: 50% middle eastern. For a while I wondered if there was some funny business in the family, but then my sister got the "right" perc
235.
by Gimpei 8y ago
What I'd hope to see is randomized studies or studies that exploit some random variation for identification. The studies mentioned in the article all appear to rely on regressions controlling for observables. But of course there could
236.
by Gimpei 8y ago
Ultra brief summary of findings: teaching Protestant values to ultra-poor Filipinos raises their income by 9%!
237.
Randomizing Religion: The Impact of Protestant Evangelism on Economic Outcomes(nber.org)
3 points by Gimpei 8y ago|1 comments
238.
Historical Legacies and African Development(drive.google.com)
2 points by Gimpei 8y ago|0 comments
239.
Silicon Valley's Keystone Problem: A Monoculture of Thought(nytimes.com)
122 points by Gimpei 8y ago|162 comments
240.
The Big Disruption(medium.com)
2 points by Gimpei 8y ago|0 comments

More ›