HNHacker News
TopNewBestAskShowJobs

haffi112

274 karma · joined November 18, 2010

submissionscomments
haffi112··on Eigg – A small Scottish isle that runs on 90-95% renewable energy
On a national level have a look at Iceland:

https://en.wikipedia.org/wiki/Renewable_energy_in_Iceland

haffi112··on 'Healthiest hearts in the world' found
> One idea is that intestinal worms - which dampen immune reactions - could be more common and this may help protect the heart.

In case you are interested in this point, then have a look at helminthic therapy (https://en.wikipedia.org/wiki/Helminthic_therapy).

haffi112··on UC Berkeley Center for Human-Compatible AI
This thought experiment on a paperclip maximizer made it very clear to me:

https://wiki.lesswrong.com/wiki/Paperclip_maximizer

haffi112··on Ask HN: Any recommendations on resources for learning one algorithm a day?
https://algo.is/ is great
haffi112··on Ask HN: How do you decide what to learn next?
I keep a list of things that interest me with sublists about interesting observations I make about each item on the list (I use workflowy for this task - no affiliation). The observations can be anything from ideas, to blog posts, books, online courses or articles about the subject.

When I want to learn something new I pick an item from the list and work on it (usually some idea I came up with). My preliminary research efforts often help me realise an idea or it gives me a chance to compare two different learning sources. I also try to create something using what I learned. Through craft I feel like I draw more from the learning experience. The outcome can also be that I need to find better references or that I simply want to learn something else.

The most difficult thing is getting started. I find it useful to be systematic about it by explicitly devoting time to it. Once you have a system in place you like it eventually becomes a habit. Also note that it is helpful to break tasks into small subtasks. Having a feeling of accomplishment leads to a more positive experience of the learning process which further leads to increased learning drive.

Note that my process is not much more sophisticated than "this is interesting". However, instead of acting on some hunch in the moment I act on observations which I gather over time.

haffi112··on AlphaGo Beats Lee Sedol in Final Game
Next challenge: Can AlphaGo beat 10 or 100 humans playing together against it?
haffi112··on Vitamin C kills tumor cells with hard-to-treat mutation in mice
Well, he was 93 years old. That's still kind of beating the odds in a way.
haffi112··on Eating Spicy Food Linked to a Longer Life
The more spicy food you eat the less crappy food you eat...? That could actually be the explanation (not a variable I saw corrected for in the text at least).
haffi112··on What happens to your brain when you stop eating sugar
You could apply this argument to anything addictive, such as smoking...
haffi112··on [dead]
The guy who doesn't work with eaters. Second order eater.
haffi112··on Facebook Fraud [video]
I meant likes on particular posts on the page. Not the page itself.
haffi112··on Facebook Fraud [video]
Well, there is one solution which wasn't mentioned in the video, get the bots to like your posts so they get visibility leading to actual users seeing them.
haffi112··on Researchers Develop Method for Getting High-Quality Photos from Crappy Lenses
I wouldn't be surprised that you could make the convex optimization algorithm independent of the lens. That would just be prior information for the algorithm which you won't necessarily need.
haffi112··on Researchers Develop Method for Getting High-Quality Photos from Crappy Lenses
I would think not. By lossy compressing the image you have already lost information necessary to reconstruct it again. The method uses information from all three color channels to reconstruct the image. So all the information is there, the image is just blurry because the color channels are shifted (chromatic aberration).
haffi112··on Ask HN: Alternatives to HN
I'm surprised no one has mentioned this but a really good option is Prismatic (getprismatic.com).

You basically just follow topics/sites and it suggests to you more sites/topics which are similar.

Browsing from there I get all the big articles on hacker news and from other sources which I wasn't aware of previously.

haffi112··on How Starcraft2 Legend TLO keeps himself organized
Prizes from competitions.
haffi112··on Automatic Generation of Regular Expressions from Examples
Then you might have the problem of over-fitting the training set.

See here for example: http://en.wikipedia.org/wiki/Overfitting

haffi112··on 20 lines of code that beat A/B testing every time
You can instead of selecting according to the largest expectation select according to the largest value you achieve with a certain variance deviation from the calculated expectation. Law of large numbers will apply and you will have tighter and tighter bounds around the expectations. That is a superior and more efficient method used in monte-carlo simulations, e.g. for AI playing the game Go.
haffi112··on Poll: What is your favorite programming language?
You missed the Morpho programming language (http://morpho.cs.hi.is/index.html).
haffi112··on R.I.P. Good Times: The Startup Bubble May Be Coming to an End in 2012
So, where will the next bubble be?
haffi112··on Ask HN: What books are you reading
I'm currently reading Gödel, Escher, Bach an eternal golden braid. I definitely recommend it.
haffi112··on The Importance of Abandoning Crap
Reason I quit using Facebook.
← PreviousPage 4 of 4