argonaut

71 ms·

751.
by argonaut 10y ago
Whether Tesla has a lower accidents/mile rate (or is some factor better or worse) is a red herring [1]. The key factor in whether this is a legitimate flaw in Tesla's system is: could an alert, competent driver easily avoid the a
752.
by argonaut 10y ago
They (reasonably) claim the joint model doesn't have to be as big, but, for example, it would be interesting the see an ensemble of 2 models: a wide model of the same size as the wide half of the joint model, and a hierarchical model o
753.
by argonaut 10y ago
Linguistic relativity is quite controversial and nowhere near as clear cut as you are making it out to be - there are studies that don't support the claims around color naming, for example. When I last spoke to someone who studied ling
754.
by argonaut 10y ago
Companies are not doing an "end-run" around taxation by using the tax codes as they are intended to be used when granting stock options. In fact, I was pointing out that the poster I was replying to's complicated scheme was
755.
by argonaut 10y ago
It's not an outright falsehood. He's referring to the (correct, but also insidious) fact that compared to a 90-day exercise window, more people will end up exercising under a more extended window - and this will dilute. In other w
756.
by argonaut 10y ago
Read the article more carefully: "Palantir did not impose restrictions on employees or stockholders selling shares in the private market."
757.
by argonaut 10y ago
On the contrary, the OP is at the exact right time in life to be thinking about these things. If the OP were 35 with a job, that would be the wrong time to be thinking about a PhD. The OP is a senior in HS, which means they can 1) tailor
758.
by argonaut 10y ago
This is one of the exceptions to noncompetes in California - when you sell your interest in a company. They are a Delaware corp. They have tons of employees outside of California (New York?).
759.
by argonaut 10y ago
One of those does not belong.
760.
by argonaut 10y ago
Yeah, this is a pretty good point. IMO it's a major flaw in the paper that they didn't empirically compare their new model to other models (like just an ensemble, as you mention), though they do discuss it. I wouldn't be surp
761.
by argonaut 10y ago
Given that the buyback price is above market price , I don't see a problem with this. Palantir is compensating you for signing a noncompete/NDA. If you don't want to sign those things, you are free to sell on the private mar
762.
by argonaut 10y ago
There's a huge tax hit. Giving shares is taxed as ordinary income.
763.
by argonaut 10y ago
First, realize that this logic applies to any conflict of interest. For example: of course the judge would rule in favor of the company he's invested in. He wouldn't invest in a company if he didn't think it was a company w
764.
by argonaut 10y ago
Yes. Those are peer reviewers. And people still publish negative result papers.
765.
by argonaut 10y ago
If they have a good research portfolio otherwise, then the first candidate's chances are unaffected, and the second candidate's chances at an interview are probably reduced. The first candidate's replication studies won'
766.
by argonaut 10y ago
This does not reflect hiring practices at reputable US research universities.
767.
by argonaut 10y ago
The original research accounted for the effect of the audition year. http://www.nber.org/papers/w5903.pdf
768.
by argonaut 10y ago
I wonder if you actually read the article. The author concludes that masking the gender of the voice has no effect on performance.
769.
by argonaut 10y ago
Did you even read the article? Nowhere does the author blame the tech industry. In fact, at the end, she is basically claiming that women tend to have less self-esteem when it comes to evaluating their performance.
770.
by argonaut 10y ago
That doesn't relate to the actual content of the emails or data. Although I agree that they dealed with the FOI requests poorly.
771.
by argonaut 10y ago
Neither of the candidates you describe will even get a faculty interview. Reply to below: I think it's valid. The first researcher has not demonstrated they will be leader/pioneer/inventor/theorizer/discoverer of ne
772.
by argonaut 10y ago
One flaw in your analogy is that if all you do is actually just delete code (you don't replace the code), that's not usually valuable. It is easier to publish null results than it is to publish legitimate positive results (o
773.
by argonaut 10y ago
Being the most prestigious null-result journal doesn't rise to being a prestigious journal, though. Regardless, this is the reality today. I don't think trying to boost a null-result journal to the level of Nature or Science is a
774.
by argonaut 10y ago
At the risk of further derailing this thread, there was nothing wrong with the content of the climategate emails, other than poor choice of words - a poor choice of words that were all too easy to take out of context.
775.
by argonaut 10y ago
A journal dedicated to negative results is IMO not the solution here. Academics must publish in reputable, prestigious journals in order to advance their career (a publication in a non-prestigious journal basically counts for nothing). A jo
776.
by argonaut 10y ago
PLoS one is an open access journal, which she talks about at the end.
777.
by argonaut 10y ago
She's talking about point 1) when she talks about open access journals (plos one is an open access journal).
778.
by argonaut 10y ago
I mean, I don't see it quite so negatively. It's good to have a diverse mix of funding sources: academia, government grants, industry research, and private research (non-industry oriented) like YCR.
779.
by argonaut 10y ago
If you think about it, YC Research is how academic research used to be funded centuries ago - patronage by wealthy elites.
780.
by argonaut 10y ago
This is called domain adaptation and transfer learning in the literature. There are ways to do that. It is an active area of research. Basically the idea is to run a few real world dogfights (you could conceivably collect a few hundreds

More ›