Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
harias
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
23 ms
·
181.
▲
Beauty is truth, truth is beauty, and other lies of physics
(aeon.co)
3 points
by
harias
8y ago
|
0 comments
182.
▲
by
harias
8y ago
India. But having classes on AWS seems fishy, why stick to Amazon? Will have those classes in the future, but the syllabus is pretty outdated.
183.
▲
by
harias
8y ago
I would consult glassdoor, payscale or similar websites to find out the wages offered to full-time employees. Then, depending on the status quo (my experience, project fit etc), will quote a similar amount.
184.
▲
Ask HN: How do I switch to CS industry(Web) from college?
3 points
by
harias
8y ago
|
5 comments
185.
▲
by
harias
8y ago
Ripple signed on some 100 banks to use their tech, not their tokens though.
186.
▲
by
harias
8y ago
I tried this once. The client ran away with the wireframe, hired a couple of unpaid interns to complete it. Just saying.
187.
▲
by
harias
8y ago
Same here. It works for paywalls too! I wonder what their business model is though.
188.
▲
by
harias
8y ago
On a side note, would amazon echo or google home be a better option?
189.
▲
by
harias
8y ago
Vectordash has some : https://vectordash.com/pricing/
190.
▲
by
harias
8y ago
Are you the same guy : https://www.reddit.com/r/gpumining/comments/86ofw2/rent_out_...
191.
▲
by
harias
8y ago
Can somebody please verify the authenticity? Why didn't arm use a subdomain of arm.com?
192.
▲
FaceForensics: A Large-Scale Video Dataset for Forgery Detection in Human Faces
(niessnerlab.org)
2 points
by
harias
8y ago
|
0 comments
193.
▲
by
harias
8y ago
CUDA libraries fit your description I guess.
194.
▲
by
harias
8y ago
Do you mean to say it was better than he depicts it to be? or is it something else?
195.
▲
by
harias
8y ago
Thanks for that. It seems safer than I imagined (Schools, theatres, and bars), might have something to do with his financial status though.
196.
▲
Learning Rich Features for Image Manipulation Detection
(arxiv.org)
3 points
by
harias
8y ago
|
0 comments
197.
▲
Spotting Image Manipulation with AI
(theblog.adobe.com)
1 points
by
harias
8y ago
|
0 comments
198.
▲
Ask HN: How do you manage to stay social?
36 points
by
harias
8y ago
|
8 comments
199.
▲
by
harias
8y ago
What about clustering?
200.
▲
Ask HN: How do you decide the priority of tasks?
1 points
by
harias
8y ago
|
0 comments
201.
▲
Theory hacks for machine learning research – Ian Goodfellow
(twitter.com)
1 points
by
harias
8y ago
|
0 comments
202.
▲
Develop Blockchain Games Without Knowing Blockchain Programming
(medium.com)
2 points
by
harias
8y ago
|
0 comments
203.
▲
HOWTO: Get into grad school for science, engineering, math and computer science
(matt.might.net)
2 points
by
harias
8y ago
|
1 comments
204.
▲
Dynamical Bias in the Coin Toss [pdf]
(statweb.stanford.edu)
1 points
by
harias
8y ago
|
0 comments
205.
▲
How I Fail – Ian Goodfellow
(veronikach.com)
153 points
by
harias
8y ago
|
36 comments
206.
▲
by
harias
9y ago
I think speed and camera should be more intuitive. Speed doesn't actually make much sense, it could be divided into classes like social media, gaming etc. And what about phones not rated by Dxomark (assuming you have used their scores)