Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Estragon
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
31 ms
·
271.
▲
by
Estragon
12y ago
Letting people know that you're absolutely OK with them taking a huge dump on your work is sometimes a good way to invite cogent criticism, when you don't have to defend your credibility for social reasons.
272.
▲
by
Estragon
12y ago
It's not wrong, but it would surprise me if anyone ever had to appeal to it when reasoning about a useful model of a specific real-world situation.
273.
▲
by
Estragon
12y ago
Yes, but the mean number of times was less than 1.5, with maximum 4 if you exclude that outlier. The vast majority of cases are perfectly compatible with curiosity, a scenario the paper ignored completely, and which could have been at leas
274.
▲
by
Estragon
12y ago
> people with lots of code on github and what looks like > lots of accepted pull requests to many projects who still > can't seem to reason How does that happen? Do you think they were lying about their github ide
275.
▲
by
Estragon
12y ago
> the interviewers have some other work to get done Work they'd otherwise be doing during the 3-hour coding interview?
276.
▲
by
Estragon
12y ago
Is there a business model here? Why are the talented engineers running the practice interviews devoting time to this?
277.
▲
by
Estragon
12y ago
Not sure why this was downvoted. I'm curious too.
278.
▲
by
Estragon
12y ago
In my experience, it's already less than 5 minutes for Lyft in SF.
279.
▲
by
Estragon
12y ago
Are there any experiments where the pilot wave theory and the Copenhagen interpretation predict different results?
280.
▲
by
Estragon
12y ago
Thanks.
281.
▲
by
Estragon
12y ago
Learning Deep Learning. I want to recapitulate the results in this paper: http://www.cs.toronto.edu/~vmnih/docs/dqn.pdf
282.
▲
by
Estragon
12y ago
You would probably get a lift / good advice from listening to this: http://www.econtalk.org/archives/2014/04/mcardle_on_fail.htm...
283.
▲
by
Estragon
12y ago
I saw a news article about suicides among 50-something men in the US going up by 50%. I'm curious, is this statistically significant? I'm having trouble finding information about the variance of suicide rates.
284.
▲
by
Estragon
12y ago
"Playing Atari with Deep Reinforcement Learning" http://www.cs.toronto.edu/~vmnih/docs/dqn.pdf
285.
▲
by
Estragon
13y ago
Is there any risk of legal action from being an airbnb guest in SF?
286.
▲
by
Estragon
13y ago
It would also prevent the interviewer from forming any prejudices based on the appearance of the candidate's screen. Esoteric window managers, that sort of thing.
287.
▲
by
Estragon
13y ago
I learned BASIC, Forth and 6809 assembly on one of these. Fond memories.
288.
▲
by
Estragon
13y ago
I've been using python since the 90s, and I learned a few things from this post.
289.
▲
by
Estragon
13y ago
Maybe the patch wouldr have gotten more attention if he'd broken it up into smaller chunks.
290.
▲
by
Estragon
13y ago
> utterly infuriating changes in the Google Maps interfaces Anyone know what Google's plan is here? I've stopped using Google Maps on my phone for the same reason. Doesn't seem like a win for them.
291.
▲
by
Estragon
13y ago
You get this for free if you use magit
292.
▲
by
Estragon
13y ago
What is a ray startup?
293.
▲
by
Estragon
13y ago
Yes, I often read the comments, decide that the article is uninteresting on that basis, and leave it at that.
294.
▲
by
Estragon
13y ago
If it's not too personal, I'm curious to know what you get out of the relationship.
295.
▲
by
Estragon
13y ago
I haven't yet heard any good music done that way, but I'm looking forward to it.
296.
▲
by
Estragon
13y ago
"Once the two hypotheses have been defined, the first hypothesis is scarcely mentioned again -- attention focuses solely on the null hypothesis. It makes me laugh to write this, but it's true! The null hypothesis is accepted or
297.
▲
by
Estragon
13y ago
This is very interesting. What actually is the hard part and what drives success, in your experience?
298.
▲
by
Estragon
13y ago
How hard is it to roll your own version of AWS's security groups? I want to set up a Storm cluster, but the methods I have come up with for firewalling it while preserving elasticity all seem a bit fragile.
299.
▲
by
Estragon
13y ago
Might be a bit of a yak shave, but you can tunnel arbitrary traffic over http. http://dag.wiee.rs/howto/ssh-http-tunneling/
300.
▲
by
Estragon
13y ago
Thanks, that example makes the issue very clear.
More ›