As mentioned in Graham's article, the question isn't which companies have headquarters there or offices there --- the question is what companies were started there. I don't know Atlanta's track record, but I suspect it doesn't look quite as good by that measure.
This being Slashdot, I figured someone would say "this is just a survival trait; natural selection in action." But from the original article is an interesting bit:
For the study, Fisher developed a questionnaire about passionate love, including such questions as "Would you die for your partner?" She said she was shocked by the answers to that query: All of the subjects said they would.
What especially surprised her was the casual way in which they responded.
Note she didn't phrase it: "for your offspring", "for your family", or "for your partner after you'd reproduced". It becomes far less obvious that this is clearly just a survival trait. Yes, evolution works in mysterious ways, and that's about all you can say about evolution's role here.
(I'm all for a well-researched evolutionary explanation, but off-the-cuff natural selection claims like this are really no different from saying "It's the will of God".)
I've studied four foreign languages, though I wouldn't say I "know" (am proficient in) any of them. But I have a few comments.
Immersion is the best, if you can afford the time/stress. Learning really sinks in when you need phrases to find food or a place to sleep. To this day I remember what a Vesperkarte is.
For a first foreign language, taking a class is probably next best. I think the quality of classes varies a lot so ask around. In my experience, one class (one hour) per week with no reinforcement in between is not enough; you really need to spend time on it every few days if not every day.
People recommend the Pimsleur method. I've used this for Spanish (levels I and II) and French I. It's very good for spoken dialog, and you'll probably retain a fair amount, but I found that when I wanted to dig into grammar, spelling, vocabulary, etc., I just got frustrated with it. It does come with written materials but they're not very substantial.
Finally, as meta-advice, I'd recommend the book How to Learn Any Language by Barry Farber. His advice and stories are very good, and they'll get you motivated if nothing else.
> A bunch of people looking out for their own interests, makes the world a better place.
I think a more accurate generalization would be: A bunch of people looking out for their own interests, under some conditions, at some tasks, can make the world a better place.
Art by committee is almost always a failure, as this "hive mind" experiment seems to demonstrate. Literature by committee too. Software by committee has a lot of successes, some failures.
What would be more interesting would be insightful observations about which tasks/projects are better undertaken alone and which can be helped by collaboration.
As for the Carlin quote, I prefer Scott Adams' observation that everyone is stupid, sometimes, at some things. Welcome to the human race.
In the end the consumer will always pay no matter what happens. If they exclusively make financial institiutions responsible for phishing then that just means they will charge us more for their services.
Maybe so, but Bruce Schneier's point (see his last paragraph) is that the bank is better qualified to handle the risk, and so should be responsible for it. Economies of scale and all that.
it will be interesting to see the brightest minds that money can buy trying to solve what decades of diplomats have unsuccessfully wrestled with
/* Tell them what they want to hear */ if (country_code(referrer_IP())) eq "cn") {
country_label = "Taiwan, a Province of China"; } else {
country_label = "Taiwan, an independent country"; }
I think you're missing his point, because continual testing is doesn't address it. The machine learning regimen you describe (M by N-fold CV) only works for static, non-drifting distributions. When given a drifting distribution like spam, there will be some "temporal data leakage" from the training sets into the test sets, ie the classifier will be trained on data that a real classifier (in the field) would not have been able to see yet. The result will be overly optimistic error rates; specifically, the false negative error rates will very likely be lower than they will be in the field.
Simply repeating the tests or increasing the data update frequency will not remove this effect.
If you read the story, it was not a physical recycle bin, but the "Recycle Bin" on one of Lee's computers. Which makes it even more stupid that he would have such a document. If you're going to negotiate employment with a competitor, especially a potentially hostile/actionable move like this, for god's sake don't use your employer's computers to do it. Had Lee never heard of backup tapes or email scanning/archival? Amazing.
OK, I love AUCtex. I wrote my thesis in emacs using AUCtex. But it's not WYSIWYG. It's not even close. {\em This} is not {\large WYSIWYG}. Neither is this equation: $y=\int_0^T x(t) dt$. Sorry.
OK, fair enough. And AUCtex with the preview-latex package isn't really WYSIWYG either, but I'll claim it gets you pretty close. It calls out to latex to typeset important stuff (equations, tables, figures, images) and inserts the results into the buffer. I claim those are the most important elements to see. All three of the elements you typed above are displayed in the xemacs buffer more or less as they'd appear.
(And to the original poster: For god's sake don't use Word to write your thesis. That's a world of hurt.)
I have been there and done that. Here are my recommendations:
Use the TeTex distribution of Latex, available for just about every distribution (and unix-like platform).
For editing LaTex code I recommend AUCtex under emacs/xemacs. If you're not a *emacs fan you may balk at this, in which case I'm not sure what to recommend. AUCtex mode under *emacs is a first-rate method of editing and almost-WYSIWYG text processing.
For managing Bibtex bibliographies there are numerous graphical editors I've tried, but I've always come back to bibtex mode under *emacs. You're editing the raw text, but the commands for navigation, manipulation and clean-up are powerful enough that you won't miss the fancier graphical apps. Also, get reftex, which is like a bridge between bibtex and AUCtex. I have bibtex files with thousands of entries and I've found bibtex/reftex good enough to manage them.
Actually, vigilantism isn't illegal. For instance, the Minuteman Project, or more basically a neighborhood watch program, is vigilantism. It's only illegal if you attempt to enforce the law yourself, bypassing the police.
> How long till we have real time crime data showing up on Google's map?
and then the next logical step...
Dear Google Inc.:
I was pleased to hear that Google's map data had finally been merged with real-time crime data. To celebrate, I knocked over two liquor stores on the 800 block of Harrison, then mugged a guy over on Grant and committed some minor vandalism around Eastwood. Then I headed on back to my apartment to see my efforts rewarded on your site.
Imagine my surprise when I got back to my browser and discovered... NOTHING! I kept reloading the damn window every 15 minutes, but not a blip showed up. I cannot express my disgust.
It used to be the Google name meant something, but ever since your stock price hit $240 you've just been slacking. It's like nobody cares any more. It's enough to make a petty criminal like me lose his faith in humanity.
Good point. But this leads to more general issues: - how long will it be until the worm writers start to care about the honeymonkeys? - how long will it be until the worm writers figure out how to differentiate between a honeymonkey and a normal host? - how long will it be until they figure out how to respond to the two differently so as not to set off honeymonkey alarms?
As with many things on the net, this can turn into an arms race, and that would be the first cycle of it. (Personally, I think there are so many unprotected/unpatched hosts that the honeymonkey effort won't make a dent in worm propagation, but I'm willing to be proven wrong.)
In other news, the CEO of Intel declared that he is looking for a copy of the April 19, 1965 issue of Electronics containing Moore's original article predicting 'Moore's Law.' "If I can find one", he declared, "I will personally swim from Santa Clara, California to Phoenix, Arizona!" After aides pointed out to him that there is technically no waterway connecting the two cities, he announced that his decision was not yet final and "some details still had to be worked out". Speaking on condition of anonymity, several top PR flacks worldwide expressed concern that this whole thing was getting out of hand.
Gee... you mean The Gap doesn't publish catalogs???
What do catalogs have to do with store display layouts?
Government agencies may not have a right to watch you, but owners of private property have the right to do anything they want... including monitor you in the restroom.
Actually, they don't. The mall may be privately owned but it is a public place (eg, you can't expose yourself in a mall just because it's private property). In a restroom you have a "reasonable expectation of privacy" and the owners can't violate that without consequences.
Nationalist claims aside, this quote just seems bizarre.
Mathematics is one skill/talent.
Writing code is another.
Hacking/cracking into systems is another.
The three skills are very different and have little to do with each other. In fact, I'd say they're negatively correlated. The mathematicians I've know were great with equations but they produced awful code, and probably couldn't have cracked their way into a paper bag. And I'd suspect most crackers wouldn't be comfortable in the formal, logical, precise world that mathematicians inhabit.
These are good questions. Everyone here seems to assume that grass is everywhere and free, so it must be renewable -- forgetting that it does take water and fertilizer and non-negligible amount of effort to harvest. That's one of the reasons alcohol from corn (or any other crop) won't solve all of our fuel problems: depletion of the topsoil is a real problem, it's not a closed cycle. And there's no such thing as a free lunch.
But Cherney seems to be a reputable guy from Cornell, so I assume he has an answer to these things. I wish the article discussed them.
The professionals in the field want to publish in prestigous journals for their reputations, journals become prestigous in part through extensive peer-review processes and widespread publication, and all that takes time/staff/money.
It takes time but not money. In my field (CS/AI) the reviewers, editors and authors aren't paid for their work. And they do wonder where all the money goes that publishers collect.
As for adoption, it's certainly happened. Two examples: Journal of AI Research (www.jair.org) and Journal of Machine Learning Research (www.jmlr.org) are both prestigious web-published journals, with citation statistics at the top of the field.
Being published in a web journal is not the same as throwing a paper up on your web site. Papers still go through an extensive review and editing process.
In the end, it's the reviewers and editors who determine the quality of a journal, not the publisher.
Slashdot Mirror
As mentioned in Graham's article, the question isn't which companies have headquarters there or offices there --- the question is what companies were started there. I don't know Atlanta's track record, but I suspect it doesn't look quite as good by that measure.
For those who want to read the original article, the final published version is here (for a fee, looks like):v let?prog=normal&id=PRLTAO000096000008088702000001& idtype=cvips&gifs=Yes
http://scitation.aip.org/getabs/servlet/GetabsSer
It's available for free (possibly a draft version) from the arXiv network:
http://arxiv.org/abs/physics/0602091
Bon apetit.
Note she didn't phrase it: "for your offspring", "for your family", or "for your partner after you'd reproduced". It becomes far less obvious that this is clearly just a survival trait. Yes, evolution works in mysterious ways, and that's about all you can say about evolution's role here.
(I'm all for a well-researched evolutionary explanation, but off-the-cuff natural selection claims like this are really no different from saying "It's the will of God".)
Indeed he did. Read about Tennyson's life sometime, and the experiences which led him to write it...
From Alfred Tennyson's poem 'In Memoriam:27', 1850.
I hold it true, whate'er befall;
I feel it, when I sorrow most;
'Tis better to have loved and lost
Than never to have loved at all.
I've studied four foreign languages, though I wouldn't say I "know" (am proficient in) any of them. But I have a few comments.
Immersion is the best, if you can afford the time/stress. Learning really sinks in when you need phrases to find food or a place to sleep. To this day I remember what a Vesperkarte is.
For a first foreign language, taking a class is probably next best. I think the quality of classes varies a lot so ask around. In my experience, one class (one hour) per week with no reinforcement in between is not enough; you really need to spend time on it every few days if not every day.
People recommend the Pimsleur method. I've used this for Spanish (levels I and II) and French I. It's very good for spoken dialog, and you'll probably retain a fair amount, but I found that when I wanted to dig into grammar, spelling, vocabulary, etc., I just got frustrated with it. It does come with written materials but they're not very substantial.
Finally, as meta-advice, I'd recommend the book How to Learn Any Language by Barry Farber. His advice and stories are very good, and they'll get you motivated if nothing else.
Yahoo! Buys del.icio.us
Wonderful headline.
I'll take Why company names shouldn't have punctuation in them for $100, Alex.
What is even more cool is that their facilities include [...] programmable high temperature ovens (think kilns)
What's a think kiln? Is that where crackpots are hardened?
> A bunch of people looking out for their own interests, makes the world a better place.
I think a more accurate generalization would be: A bunch of people looking out for their own interests, under some conditions, at some tasks, can make the world a better place.
Art by committee is almost always a failure, as this "hive mind" experiment seems to demonstrate. Literature by committee too. Software by committee has a lot of successes, some failures.
What would be more interesting would be insightful observations about which tasks/projects are better undertaken alone and which can be helped by collaboration.
As for the Carlin quote, I prefer Scott Adams' observation that everyone is stupid, sometimes, at some things. Welcome to the human race.
In the end the consumer will always pay no matter what happens. If they exclusively make financial institiutions responsible for phishing then that just means they will charge us more for their services.
Maybe so, but Bruce Schneier's point (see his last paragraph) is that the bank is better qualified to handle the risk, and so should be responsible for it. Economies of scale and all that.
it will be interesting to see the brightest minds that money can buy trying to solve what decades of diplomats have unsuccessfully wrestled with
/* Tell them what they want to hear */
if (country_code(referrer_IP())) eq "cn") {
country_label = "Taiwan, a Province of China";
} else {
country_label = "Taiwan, an independent country";
}
TOP 5 NAMES SUGGESTED BY RAYTHEON ENGINEERS BEFORE THEIR PR DEPT TOOK OVER:
I think you're missing his point, because continual testing is doesn't address it. The machine learning regimen you describe (M by N-fold CV) only works for static, non-drifting distributions. When given a drifting distribution like spam, there will be some "temporal data leakage" from the training sets into the test sets, ie the classifier will be trained on data that a real classifier (in the field) would not have been able to see yet. The result will be overly optimistic error rates; specifically, the false negative error rates will very likely be lower than they will be in the field.
Simply repeating the tests or increasing the data update frequency will not remove this effect.
If you read the story, it was not a physical recycle bin, but the "Recycle Bin" on one of Lee's computers. Which makes it even more stupid that he would have such a document. If you're going to negotiate employment with a competitor, especially a potentially hostile/actionable move like this, for god's sake don't use your employer's computers to do it. Had Lee never heard of backup tapes or email scanning/archival? Amazing.
OK, I love AUCtex. I wrote my thesis in emacs using AUCtex. But it's not WYSIWYG. It's not even close.
{\em This} is not {\large WYSIWYG}. Neither is this equation: $y=\int_0^T x(t) dt$. Sorry.
OK, fair enough. And AUCtex with the preview-latex package isn't really WYSIWYG either, but I'll claim it gets you pretty close. It calls out to latex to typeset important stuff (equations, tables, figures, images) and inserts the results into the buffer. I claim those are the most important elements to see. All three of the elements you typed above are displayed in the xemacs buffer more or less as they'd appear.
(And to the original poster: For god's sake don't use Word to write your thesis. That's a world of hurt.)
I have been there and done that. Here are my recommendations:
Use the TeTex distribution of Latex, available for just about every distribution (and unix-like platform).
For editing LaTex code I recommend AUCtex under emacs/xemacs. If you're not a *emacs fan you may balk at this, in which case I'm not sure what to recommend. AUCtex mode under *emacs is a first-rate method of editing and almost-WYSIWYG text processing.
For managing Bibtex bibliographies there are numerous graphical editors I've tried, but I've always come back to bibtex mode under *emacs. You're editing the raw text, but the commands for navigation, manipulation and clean-up are powerful enough that you won't miss the fancier graphical apps. Also, get reftex, which is like a bridge between bibtex and AUCtex. I have bibtex files with thousands of entries and I've found bibtex/reftex good enough to manage them.
Best of luck on your thesis...
Actually, vigilantism isn't illegal. For instance, the Minuteman Project, or more basically a neighborhood watch program, is vigilantism.
It's only illegal if you attempt to enforce the law yourself, bypassing the police.
That's the very definition of vigilante
Remember, vigilante comes from vigilance.
Comes from, originally. Not synonymous with.
> How long till we have real time crime data showing up on Google's map?
and then the next logical step...
Dear Google Inc.:
I was pleased to hear that Google's map data had finally been merged with real-time crime data. To celebrate, I knocked over two liquor stores on the 800 block of Harrison, then mugged a guy over on Grant and committed some minor vandalism around Eastwood. Then I headed on back to my apartment to see my efforts rewarded on your site.
Imagine my surprise when I got back to my browser and discovered... NOTHING! I kept reloading the damn window every 15 minutes, but not a blip showed up. I cannot express my disgust.
It used to be the Google name meant something, but ever since your stock price hit $240 you've just been slacking. It's like nobody cares any more. It's enough to make a petty criminal like me lose his faith in humanity.
Signed,
-Disgusted in Chicago
Good point. But this leads to more general issues:
- how long will it be until the worm writers start to care about the honeymonkeys?
- how long will it be until the worm writers figure out how to differentiate between a honeymonkey and a normal host?
- how long will it be until they figure out how to respond to the two differently so as not to set off honeymonkey alarms?
As with many things on the net, this can turn into an arms race, and that would be the first cycle of it.
(Personally, I think there are so many unprotected/unpatched hosts that the honeymonkey effort won't make a dent in worm propagation, but I'm willing to be proven wrong.)
It's my God given right as an American to be able to sit at home in my underwear and kill shit.
Ah, but you didn't read the last paragraph of the article, which says:
Supporters have suggested the remote hunting could be beneficial for hunters with disabilities
Apparently it's some God given right to be able to sit at home in a wheelchair and kill shit.
In other news, the CEO of Intel declared that he is looking for a copy of the April 19, 1965 issue of Electronics containing Moore's original article predicting 'Moore's Law.' "If I can find one", he declared, "I will personally swim from Santa Clara, California to Phoenix, Arizona!" After aides pointed out to him that there is technically no waterway connecting the two cities, he announced that his decision was not yet final and "some details still had to be worked out". Speaking on condition of anonymity, several top PR flacks worldwide expressed concern that this whole thing was getting out of hand.
Gee... you mean The Gap doesn't publish catalogs???
What do catalogs have to do with store display layouts?
Government agencies may not have a right to watch you, but owners of private property have the right to do anything they want... including monitor you in the restroom.
Actually, they don't. The mall may be privately owned but it is a public place (eg, you can't expose yourself in a mall just because it's private property). In a restroom you have a "reasonable expectation of privacy" and the owners can't violate that without consequences.
Newsflash: Dorkwads Prank Dickwads in Famous Wad Rivalry!
Nationalist claims aside, this quote just seems bizarre. Mathematics is one skill/talent. Writing code is another. Hacking/cracking into systems is another. The three skills are very different and have little to do with each other. In fact, I'd say they're negatively correlated. The mathematicians I've know were great with equations but they produced awful code, and probably couldn't have cracked their way into a paper bag. And I'd suspect most crackers wouldn't be comfortable in the formal, logical, precise world that mathematicians inhabit.
These are good questions. Everyone here seems to assume that grass is everywhere and free, so it must be renewable -- forgetting that it does take water and fertilizer and non-negligible amount of effort to harvest. That's one of the reasons alcohol from corn (or any other crop) won't solve all of our fuel problems: depletion of the topsoil is a real problem, it's not a closed cycle. And there's no such thing as a free lunch.
But Cherney seems to be a reputable guy from Cornell, so I assume he has an answer to these things. I wish the article discussed them.
The professionals in the field want to publish in prestigous journals for their reputations, journals become prestigous in part through extensive peer-review processes and widespread publication, and all that takes time/staff/money.
It takes time but not money. In my field (CS/AI) the reviewers, editors and authors aren't paid for their work. And they do wonder where all the money goes that publishers collect.
As for adoption, it's certainly happened. Two examples: Journal of AI Research (www.jair.org) and Journal of Machine Learning Research (www.jmlr.org) are both prestigious web-published journals, with citation statistics at the top of the field.
Being published in a web journal is not the same as throwing a paper up on your web site. Papers still go through an extensive review and editing process.
In the end, it's the reviewers and editors who determine the quality of a journal, not the publisher.