ntoshev

88 ms·

811.
by ntoshev 18y ago
They will get a lot of training and assessment data for their algorithms in this way.
812.
by ntoshev 18y ago
There is no test etherpad on their page where you can actually test it on a document with others editing simultaneously, so I created one for HN: http://etherpad.com/a4jtHYAoF9 Just join in and try it!
813.
by ntoshev 18y ago
A lot of expertise is not based on scientific statistics, but rather on the sample of problems the expert encounters and his intuition.
814.
by ntoshev 18y ago
The other article is actually much better, never mind the revenue model references.
815.
by ntoshev 18y ago
NLTK uses statistical approach, which it seems he doesn't consider to be "NLP". TF-IDF is also part of that statistical approach. He seems to be talking about the old school rule-based natural language understanding.
816.
by ntoshev 18y ago
http://netflixprize.com Statistical language processing is also cool: http://nltk.sourceforge.net/index.php/Main_Page http://norvig.com/spell-correct.html
817.
by ntoshev 18y ago
A smart country? It is better to think smaller here: put yourself (and possibly other engineers) in charge and reap the windfall of profits. I am. Well, still working on that "windfall of profits" part. ;)
818.
by ntoshev 18y ago
Such overgeneralizations are useless. How much are they tied? How the effect of recession compares to the already known trends?
819.
by ntoshev 18y ago
What is wrong with their interpretation? The data are incomplete as they don't account for possible seasonality as brfox pointed out, and because they don't include small ad networks.
820.
by ntoshev 18y ago
Online advertising changes. Product ads are headed towards blending with recommendation engines (this is what the whole "targeting" is about), branding ads can stay where they are or (more likely) go in the directions of "your friends like
821.
by ntoshev 18y ago
You are right - I thought the article is useless and the discussion is also useless as there is nothing to discuss but techcrunch/news.yc going downhill etc etc. As there was nothing to dilute I posted the "No" comment just to see what will
822.
by ntoshev 18y ago
No.
823.
by ntoshev 18y ago
How can you teach concurrent programming without teaching sequental programming first?
824.
by ntoshev 18y ago
You can get the same input via the auditory channel. All you need is webcam and software, or a Series 60 mobile phone: http://news.bbc.co.uk/2/hi/science/nature/3171226.stm http://www.seeingwithsound.com/
825.
by ntoshev 18y ago
Can you elaborate why you'd feel better if they assured you your data are hosted separately and maybe give you ssh to the machine? They would still be able to access/control all your data. I really wish I understood how business people perc
826.
by ntoshev 18y ago
The notation you'd use to represent the program/proof is very important.
827.
by ntoshev 18y ago
That's cool. What if I take an e-mail from Mike Arrington's iPhone and just change the contents (well, I would also change stuff like headers with date/time)? I can't try it with Arrington because I have never received email from him, but I
828.
Discussing the Global Brain and the Future of the Web with Peter Norvig et al (Video)(blogoscoped.com)
5 points by ntoshev 18y ago|0 comments
829.
by ntoshev 18y ago
principle
830.
by ntoshev 18y ago
How does Posterious authenticate you? Anyone could forge the email sender.
831.
by ntoshev 18y ago
Monitoring social trends this way is very cool. I wonder if they could help predict things like consumer confidence index similarly.
832.
by ntoshev 18y ago
I always wondered what a government would look like if everyone could decide for theirselves where to go, say, 50% of their taxes.
833.
by ntoshev 18y ago
I would rather blame their protectionist policies, not the success in exports.
834.
by ntoshev 18y ago
Everyone who follows Hacker News knows English and has seen the original, so this post is clearly redundant.
835.
by ntoshev 18y ago
The hard part is molding the social behavior into something that adds a lot of value to the users and eventually to Facebook shareholders. How hard is that? I remember a thread here in HN asking how much time would you need to build a site
836.
by ntoshev 18y ago
Their real problem is not the burn rate but that they can't bring in reasonable revenue (for the traffic they have). Here is an interesting article on how other social networks are doing: http://www.usatoday.com/tech/techinvestor/industry/
837.
by ntoshev 18y ago
Maybe it's a really hard problem that we don't fully understand. EDIT: deleted unnecessary nagging
838.
by ntoshev 18y ago
I'm looking at it, and it seems it makes sense when your team is distributed, but not if the two developers can actually make the code review on the same computer.
839.
by ntoshev 18y ago
Tim O'Reilly has explained this for me: http://news.ycombinator.com/item?id=345217
840.
The Young Man's Business Model(paultyma.blogspot.com)
65 points by ntoshev 18y ago|14 comments

More ›