Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
lowercase1
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
31.
▲
by
lowercase1
6y ago
Too many comments focused on whether Google was justified in firing her/the appropriateness of the email and not enough on the impact/veracity allegations. Much more interesting questions that might get revealed a bit: Does Google
32.
▲
by
lowercase1
6y ago
The anger is a distraction. The allegations are most important. "Google silences minorities/women/AI researchers" is way more interesting/important than "Google/AI researcher did/didn't misbehave
33.
▲
by
lowercase1
6y ago
Too many people expecting too perfect of behavior from someone apparently put it a shitty situation by much more powerful forces.
34.
▲
by
lowercase1
6y ago
Based on the contents of the email things were already beyond repair https://www.platformer.news/p/the-withering-email-that-got-a...
35.
▲
by
lowercase1
6y ago
I mean if you are a bias/ethics researcher and your bosses ignore/limit your work then there are reasons to complain/whistleblow.
36.
▲
by
lowercase1
6y ago
Slandered, how? https://twitter.com/timnitGebru/status/1331757629996109824?s... seems to be the complaint. This seems plausible given her work to push towards less biased data/algorithms got squashed and she
37.
▲
by
lowercase1
6y ago
Once you factor in selection it probably does.
38.
▲
by
lowercase1
6y ago
I'd say obfuscation is creating a level of indirection just to look modular. Moving things complex dependent side effects into separate functions just hides what is going on instead of clarifying it. void MajorFunction( void ) {
39.
▲
by
lowercase1
6y ago
Facebook provides a service that people use. Improving the likelihood of finding accurate information may increase people's likelihood of using it. Taking down groups decreases that likelihood. Which is more important isn't always
40.
▲
by
lowercase1
6y ago
Planned from 2010? Give me a break. https://www.nais.org/magazine/independent-school/spring-2010... And nobody has contradicted the story? C'mon.
41.
▲
by
lowercase1
6y ago
Biden is a studderer and that can cause merged and slurred words.
42.
▲
by
lowercase1
6y ago
Blatant in annotation false or outrageous claims by Trump? Trump claiming victory when votes aren't counted is something to be careful about reporting.
43.
▲
by
lowercase1
6y ago
Hydrogen is very cheap to store. Renewable electricity is free once the panels/turbines are built. There will be long periods with zero price energy and other times with limited energy. Making hydrogen to use for the energy into the da
44.
▲
by
lowercase1
6y ago
I've never really seen a good explanation of why REST is a meaningful concept. Why REST? Why these requirements? What do you mean stateless? What is bad about deviating?
45.
▲
by
lowercase1
6y ago
Have you moved on from Cadence to this? What differences are there? Looks interesting.
46.
▲
by
lowercase1
6y ago
This is similar to AWS step functions which kinda descended from more general/powerful AWS Simple Workflow.
47.
▲
by
lowercase1
6y ago
Ooh interesting, I'd be interested in that code/work. I'm somewhat familiar with Tornado codes but haven't looked in years
48.
▲
by
lowercase1
6y ago
Reed Solomon might run into issues with it's n^2 runtime if you make too many pieces. Fountain codes or other codes though could do the job.
49.
▲
by
lowercase1
6y ago
Yeah I skimmed through and was like "there is no content here"
50.
▲
by
lowercase1
6y ago
Note that as much as 30 percent of car pollution is not from the fuel and would remain.
51.
▲
by
lowercase1
6y ago
Yes you can! Apply an All or Nothing Transform before RS encoding so that you need all of the original data to decrypt. Then you can only decrypt if you have k of n pieces. You save by a factor of k in piece size over SSS. http://
52.
▲
by
lowercase1
6y ago
You could also look at http://web.eecs.utk.edu/~plank/plank/papers/FAST-2011.pdf You can combine the reconstruction properties of Reed Solomon (you need k of n pieces) with the All or Nothing Transform. Encry