ericjang

62 ms·

241.
by ericjang 9y ago
Does de-oxygenated blood have a measurably different refractive index w.r.t. radio waves than oxygenated blood?
242.
by ericjang 9y ago
The analysis in this article was based on data from Thomas Piketty and collaborators, who have also compiled an extremely detailed and comprehensive analysis of the subject of capital & wealth & inequality. I highly recommend Pikett
243.
by ericjang 9y ago
Why are you so bullish in crypto? Would love to hear your thoughts here (or send me a DM if you prefer).
244.
Fair and Balanced? Thoughts on Bias in Probabilistic Modeling(medium.com)
1 points by ericjang 9y ago|0 comments
245.
by ericjang 9y ago
This. I wrote a crypto arb bot in 2014 and decided to stop working on it / open-source it for the same reason after a couple exchanges that I had cryptos in mysteriously vanished.
246.
Why Smelling the Opposite Sex Can Age You(aging.nautil.us)
1 points by ericjang 9y ago|0 comments
247.
Expressivity, Trainability, and Generalization in Machine Learning(blog.evjang.com)
2 points by ericjang 9y ago|0 comments
248.
by ericjang 9y ago
If you don't mind me asking, what social media are your family members using? Are their friends on that platform too?
249.
by ericjang 9y ago
If it were memorizing training data, it would due poorly on a Wasserstein critic trained on a validation set drawn from the same data distribution.
250.
by ericjang 9y ago
This is a general problem for generative models, regardless of whether they are explicit latent density (variational autoencoders) or implicit density (GANs) or fully-observed (PixelCNN). To date no consensus has been reached in the researc
251.
by ericjang 9y ago
This is a very wise answer! This reminds me of the movie "Iron Sky" which involves Nazis on the Moon. I found the film surprisingly intelligent (at least more than I expected) and had some visually nice political commentary.
252.
by ericjang 9y ago
Maybe the laws of our math and physics are not invariant to space and time :)
253.
Undercover at the Bullshit Factory (Dan Lyons)(currentaffairs.org)
2 points by ericjang 9y ago|0 comments
254.
by ericjang 9y ago
Is this an American cultural thing, or are other cultures just as averse to public swearing?
255.
by ericjang 9y ago
He mentions at 14:00 the "Dark Brown" experience, where he went to Brown University to make the HyperText system. He says "he had to work under a torrent of insults from someone who he thought was his friend and HyperText got
256.
by ericjang 9y ago
I met Inigo at an internship when I was just starting out my CS career and he was kind and patient enough to mentor me on all sorts of cool computer graphics. A quote from "The Prestige" comes to mind: Oh, no, sir. This wasn&#x27
257.
by ericjang 9y ago
I would guess "environment simulators"? Environments are to RL as datasets are to supervised/unsupervised learning.
258.
by ericjang 9y ago
she's
259.
Can a Divided America Heal?(ted.com)
1 points by ericjang 9y ago|0 comments
260.
by ericjang 9y ago
this is true. Source: I work at Google
261.
by ericjang 9y ago
Interesting points. What would you consider an adequate task that demonstrates a program/robot has acquired "semantic understanding"?
262.
by ericjang 9y ago
Murray makes a number of nuanced comments about AI philosophy that I enjoyed. I've reproduced them below, but please see the full transcript in case I have taken quotes out of context: - Is she conscious at all? Is she conscious just
263.
by ericjang 9y ago
Does that imply that lifting, building up one's body and consuming excess calories to bulk up is somehow stressful for the body's metabolic aging process?
264.
by ericjang 9y ago
I'm grateful that there is such a wide spectrum of how to interpret the ambition of these companies between yours and the parent post's comment. Thanks :)
265.
by ericjang 9y ago
For those interested in this topic, I highly recommend "The Song Machine", by John Seabrook. Music is no longer written by single songwriter-singer, it's manufactured on an assembly line to optimize for 'hits'. Hits
266.
by ericjang 9y ago
No quote in particular, but it seems unlikely that they would miss such a blatant dataset bias and fail to instrument their ML models with metrics besides accuracy. But then again, welcome to Silicon Valley :) 1 - 200/6158 = 97% is in
267.
by ericjang 9y ago
I think the article is not precise with their wording, but the 97% is actually recall (i.e. detects 97% of the positives).
268.
by ericjang 9y ago
My eyes glazed over at the vast number of priors introduced in this work. Why should I believe the assumptions hold? In Machine Learning, the standard expectation of research papers is to empirically demonstrate the increased statistical po
269.
by ericjang 9y ago
Meta question: is there a "Hacker News of China" forum where I can find this article discussed from the Chinese perspective?
270.
by ericjang 9y ago
I'm reminded of this famous historical figure who inspired numerous vampire stories. https://en.wikipedia.org/wiki/Elizabeth_B%C3%A1thory

More ›