Fede_V

23 ms·

181.
by Fede_V 11y ago
My experience is that the entire value of an MBA is in the network of people you meet through business school. I asked a friend who had a technical degree from an excellent university, and later got an MBA what useful skills he learned, and
182.
by Fede_V 11y ago
Scott Aaronson is fucking awesome. Whenever there's a new paper claiming breakthroughs about quantum computing or P=NP, I go to his blog and wait for him to clearly explain it. He is the very rare case of a top notch scientist with a
183.
by Fede_V 11y ago
Christian Gohlke distributes free NumPy binaries built against MKL.
184.
by Fede_V 11y ago
Chapeau Mr. and Mrs. Zuckerberg.
185.
by Fede_V 11y ago
The pendulum has swung quite far in the opposite direction now, but it used to be a common cliche` that a business person would come up with a possible business idea, than look for a 'tech person' to actually build it. When techie
186.
by Fede_V 11y ago
You got the journal completely wrong.
187.
by Fede_V 11y ago
If the error is proportional to the measurement - can't you just take the log to make it additive again?
188.
by Fede_V 11y ago
The best they can do is measure the sample with zero measurement noise - but that's no panacea if the variation is in the sample itself, and not the measurement.
189.
by Fede_V 11y ago
If this study is correct, and blood drops have a high variability in certain important clinical features, then, no matter how accurate the instruments are, it's impossible to actually get a clean measurement from low volumes of blood.
190.
by Fede_V 11y ago
Absolutely no. However, as the parent comment pointed out, the root problem is the broken incentives - the lack of philosophical sophistication of scientists is at best a 3rd order term.
191.
by Fede_V 11y ago
It would be if what got published were plenty of incremental papers. That's the case in certain fields (deep learning comes to mind) - but in other fields, such as the life sciences, you instead get a bunch of results that claim to be
192.
by Fede_V 11y ago
Yeah, pretty much exactly this. In the abstract, I fully support teaching science PhDs more about Quine, Russell, Popper, Carnap - they are all great thinkers who had very intelligent ideas. In practice, PhDs at top tier institutions are
193.
by Fede_V 11y ago
They really came to the forefront with alex.net - GPUs plus lots of labeled examples made them suddenly very practical.
194.
by Fede_V 11y ago
Random forests are very difficult to interpret especially once you start having lots of trees.
195.
by Fede_V 11y ago
Sure. Random forests are a very powerful machine learning method that has proven to give excellent performance with very little tuning. Neural networks with many layers (aka deep learning) is the new frontier of machine learning - they are
196.
by Fede_V 11y ago
Scientists applying for academic positions do not count against the H1B limit. That only applies for academic position though, not work in private industry.
197.
by Fede_V 11y ago
One small note on the paper itself. I love that the paper's author are in alphabetical orders, without any stupid jostling over first/last authorship. I think that's fantastic and wish academia went that way too (yeah right)
198.
by Fede_V 11y ago
This is awesome, thanks so much for sharing this - and I love how whimsical the docs are: Our pond is a perfect 500 x 500 square, as is the case for most ponds found in nature. A feedback on the PDE aspect: is there any nice way to write yo
199.
by Fede_V 11y ago
No, those are the deaths cause by the extra pollution due to the cheating. In practice, as a society, we indirectly decide on pollution standards by voting in politicians who nominate people in regulatory roles. If we are unhappy with th
200.
by Fede_V 11y ago
What infuriates me about this is that the people who made that decision were responsible for at least 50 deaths due to increased pollution, and probably even more. If a serial killer murdered 50 people, there would be a public outcry - in
201.
by Fede_V 11y ago
I agree with the second part completely, but the story did mention the ethnicity of the accused.
202.
On the proliferation of ninjas, evangelists, and hackers(mobile.nytimes.com)
2 points by Fede_V 11y ago|0 comments
203.
by Fede_V 11y ago
This is remarkably humble and honest. Most people are willing to take far too much credit for their own success, and far too little responsibility for their own failures.
204.
by Fede_V 11y ago
PyPy is an absolutely amazing piece of technologies and the core developers are brilliant people - however, their main use case is not numeric python. Their framework is (theoretically) able to trace any Python operation, no matter how dyna
205.
by Fede_V 11y ago
I blogged about a very similar topic (in biology) over a year ago: http://federicov.github.io/Blog/deception-by-mathematics.htm... People give a lot of authority to mathematics, because mathematics is immensely powerfu
206.
by Fede_V 11y ago
Credible research institutions have no problems publishing white papers or research reports discussing their own research. For example, Microsoft Research is fully financed by Microsoft, but they have a sterling academic reputation which t
207.
by Fede_V 11y ago
Hi sama, First of all, thanks for the 10 million dollar donation for YC research. That's an amazing gesture, and I'm really looking forward to seeing the scientific research model get completely disrupted - it's in a terribl
208.
by Fede_V 11y ago
I think this is absolutely amazing. sama has written multiple times about the 'distractions' that pull founders away from building beautiful products. Things like attending conferences, networking, fundraising, etc... Conservativ
209.
by Fede_V 11y ago
Julia does very little tracing, except to infer types, which it then passes on to LLVM. All the optimizations in Julia are because Julia stresses type stability, and once types are known, LLVM can generate very good code specific for those
210.
Uber, but for cancelling Comcast subscriptions(geek.com)
1 points by Fede_V 11y ago|0 comments

More ›