Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
angel_j
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
22 ms
·
181.
▲
by
angel_j
8y ago
The companies have been funding the research anyway, it's about time they hire them too. The bigger problem will be a brain/knowledge drain at the universities.
182.
▲
by
angel_j
8y ago
They all want to own a rental too, since it only costs a down payment to possess the property iff somebody else pays the bank. But it's more than NIMBYS. There's foreign capital holding land and property, dog-eat-world constructio
183.
▲
by
angel_j
8y ago
The US has an annualized growth rate [0] of ~.7%, or ~2M/yr, and this an historically low rate (probably b/c of economic problems) Annualized "new housing starts" is 1.28M/yr [1] And we already have a housing defici
184.
▲
by
angel_j
8y ago
In California, you must have a special license to do almost any business, but be a landlord. You need a license to do people nail's (nail technician). And in some cases, there a special taxes applied to that industry (like legal cann
185.
▲
by
angel_j
8y ago
Could you... modulate the amplitude of the PD?
186.
▲
by
angel_j
8y ago
If you think China is your competitor, it makes economic and political sense to help them do something stupid. Don't interrupt your enemy while they are making a mistake, unless you can assist.
187.
▲
by
angel_j
8y ago
What if you used the noise as the carrier wave for the signal? Like, if in an urban area you have a "stable" probability distribution of noise, modulate that.
188.
▲
by
angel_j
8y ago
From a frog.
189.
▲
by
angel_j
8y ago
I'm trying to think of some way a large equity firm could make money on shares that are almost certain to go from high to low...
190.
▲
by
angel_j
8y ago
> Firefox's letterboxing support doesn't only work when resizing a browser window but also works when users are maximizing the browser window, or entering in fullscreen mode Brilliant. All you have to do is change your window
191.
▲
by
angel_j
8y ago
We should destroy phone numbers instead, not only for spam, but for privacy and security. You might as well wear an RFID on your collar. Replace phone numbers with DNS and spam blocking. Change it when you want instead of asking your prov
192.
▲
by
angel_j
8y ago
https://www.hessix.com/?show=math.AP https://www.hessix.com/?show=math.NT https://www.hessix.com/?show=math.NA https://www.hessix.com/?show=math.OC etc..
193.
▲
by
angel_j
8y ago
Does it test against and prevent over-fitting?
194.
▲
by
angel_j
8y ago
Tensorflow.js is pretty good.
195.
▲
by
angel_j
8y ago
I've seen carpenter bees in California as big as that.
196.
▲
by
angel_j
8y ago
They did release the model [0]. A smaller-scale version of it. If you want the big one, all you have to do is scrape 45M web pages, and increase the parameters to scale. What they didn't release is the training corpus and the trained
197.
▲
by
angel_j
8y ago
I always thought the obvious route is to replace a couple lanes of interstate with rail.
198.
▲
by
angel_j
8y ago
That's if you go about it from low-dimensional embedding to feature classification, and predictions based on probabilities for those features. As you mention, that works if you have metric data, or can juice metrics from abstraction.
199.
▲
by
angel_j
8y ago
How many reddit users here hit the [-] to minimize discussions?
200.
▲
by
angel_j
8y ago
2nd "The Alphabet vs The Goddess"
201.
▲
by
angel_j
8y ago
https://github.com/deepmind/hanabi-learning-environment
202.
▲
by
angel_j
8y ago
Some Fundamental Theorems in Mathematics are bigger than others. Some Fundamental Theorems in Mathematics' Mothers are bigger than other Fundamental Theorems in Mathematics' Mothers.
203.
▲
by
angel_j
8y ago
Pssh, all you need to do is install docker, convert static files and blog post into lambdas, then route each unique http request to a serverless IPv6 address located on the nearest continent.
204.
▲
by
angel_j
8y ago
Pretty sure all this says is to minimize KL-Divergence instead of log-likelihood (for the encoder), or your latent variables are garbage. Judging by many examples I've seen of VAEs in ML, this is not news. This post [0] does a good jo
205.
▲
by
angel_j
8y ago
Maybe they undertook pyramids too fanatically altogether, and that that was their downfall. The thought-leaders thought pyramids were the best investment, instead of astronomy. The downfall may have been an effect of the great thing. Pyra
206.
▲
by
angel_j
8y ago
"I can take some fictional people, make everyone betray them for all their own petty and cowardly reasons, and our protagonists will just have to deal with it." Birth of a genre.
207.
▲
by
angel_j
8y ago
The first author has done similar studies with lower p values: https://www.ncbi.nlm.nih.gov/pubmed/24740588 https://www.ncbi.nlm.nih.gov/pubmed/26632023
208.
▲
by
angel_j
8y ago
Fair editorial suggestion, no doubt. Will change to 'yelling "terrorists!" on a busy day at Disneyland'. H/e, that the example was used to uphold an immoral conviction does not make it an "immoral standard"
209.
▲
by
angel_j
8y ago
Terms of Service are really out of control, and probably need a democratic checkup. One cannot sign away constitutional rights, which means somewhere there is a line to draw. When is a company abusing it's first amendment rights and d
210.
▲
by
angel_j
8y ago
Tech companies should use these to screen applicants instead of code challenges.
More ›