Slashdot Mirror

Forget about silly functions like stocking grocery shelves, cleaning, etc. A friend of mine has invented a system that allows AI to do the single most important human activity:

Watching reality TV.

That's right. When the new visually acute robots put you out of a job, and you take your severance check and slink home to watch "Cops," you'll find a robot already hogging the La-Z-Boy, remote control in hand. Not only are we obsolete--our obsolescenece is obsolete, too.

One must also ask, and this is something I rarely see in the general debate : "What about all the nitrogen?"

Even back in 1994 the global warming potential of fertilizers were known :

"In wet soils, denitrifying bacteria convert nitrate to nitrous oxide and gaseous nitrogen. The former is a greenhouse gas that has an energy reflectivity per mole 180-fold higher than that of carbon dioxide."

I came across the notion in an MIT courseware video lecture (16 or 17 I think)

On a slightly different tack nitrogen's role in reducing carbon fixing was documented in 1996
  and thus warning against adding nitrogen to the ecosystem because it reduces the ability to fix the dreaded carbon, ignoring N's own contribution.

Yet here we have Nasa saying that carbon fixing is nitrogen limited and we should add more nitrogen to the system.

Not that all modern thinking is pro-nitrogen.

Add into the mix the world's estimated 1,300,000,000 cattle belching out 400 litres of methane each per day : 520,000,000,000 litres

Here's more on methane

Methane is responsible for nearly as much global warming as all other non-CO2 greenhouse gases put together. Methane is 21 times more powerful a greenhouse gas than CO2. While atmospheric concentrations of CO2 have risen by about 31% since pre-industrial times, methane concentrations have more than doubled. Whereas human sources of CO2 amount to just 3% of natural emissions, human sources produce one and a half times as much methane as all natural sources. In fact, the effect of our methane emissions may be compounded as methane-induced warming in turn stimulates microbial decay of organic matter in wetlands--the primary natural source of methane.

and more

What conclusions?
My conclusion is that reducing one's carbon footprint will not suffice. The way to fix more nitrogen is to grow more pulses and legumes which is good because you're going to need something to replace the cows you're eating now. Stop pouring nitrogen on to the fields and start eating more organic produce.

As we've been saying for a while : "think globally, act locally"

The paper claims the source code is (or will be) here. Next time, ask the paper.

and seeing the spam for what it is

oh and here is the PDF
http://cbcl.mit.edu/projects/cbcl/publications/ps/ serre-wolf-poggio-PAMI-07.pdf

not that Roland would even understand what it says, he just reads press releases via RSS, copies the summary and hits submit
We appreciate that the Editor removed his spammy link to ZDNet (no wonder they are losing cash)
but is Slashdot that short of good stories that they have to choose a known plagiarisers articles and actively edit them over the hundreds of original submissions they get daily ?

i would of chosen to read Digg instead but that is even worse, full of credit card scams, made for adsense blogs and millions of MLM bloggers all hawking their refferal links and real estate blogs hoping people will click on their crappy asbestos and insurance links

sheesh can't a geek get some decent news for a change ? obviously not, Internet 2 anybody

Using queues in parallel languages is not new. Your project also contains, like, 5 files and was started on January 30. Apparently you're not very far along yet.

I think this quote, by the Attorney General of the State of Massachusetts, Martha Coakley, sums up the overreaction and the unwillingness to look at the situation rationally:

"For those who responded to it, professionals, it had a very sinister appearance," Coakley said. "It had a battery behind it and wires."

(My source for that quote is a Boston Globe article.)

Oooooooh! Batteries and wires!! Run away!

My feeling is this: if I lived in the state, I'd damn well make sure I stayed away from Radio Shack, because I'm likely to get caught in the crossfire when someone buys a few electronics components and the SWAT team comes in to take out the "terrorist" with a storm of bullets. Have these people never, ever seen a homemade electronics project before!? For God's sake, MIT is located in their state!

Dude, it's MIT. It's the land of mental masturbation. When was the last time anything significant came out of that freakshow? Everything they ever announce always has a huge drawback that makes it totally useless. This is no different.

Whatever

But why not convert the heat directly into electricity without an intermediate mechanical generator? http://web.mit.edu/newsoffice/2001/electricity-120 5.html

So, they gave each base-pair a color? What on earth is the point? 98% of that sequence doesn't do anything. And why is a virtually random sequence of pixels of 4 different colors "beautiful"?

I can understand if they took two different genomes from the same species and did some kind of comparison: different colors for matches, indels, translocations, silent/synonymous/non-synonymous SNPs, etc. Or translated the sequence and colored by hydrophobicity/charge/polarity/whatever. Or showed haplotype conservation between species.

At least that would tell you something, this is just a bunch of pixels with no meaning. A vaguely similar thing I've done was to plot plot SNP density (as color intensity) over the genome - but that was for a specific project, I didn't realize such things are "new visions".

There are definitely prettier visualizations out there too: http://acg.media.mit.edu/people/fry/genomevalence

Even this is a lot more informative (I think www.visualcomplexity.com was mentioned on slashdot a couple of years ago).

Perhaps you should do some more reading on the subject then, like the article Soft Surveillance: Mandatory Voluntarism and the Collection of Personal Data by Gary T. Marx. Here's a good quote:

The first task of a society that would have liberty and privacy is to guard against the misuse of physical coercion by the state and private parties. The second task is to guard against the softer forms of secret and manipulative control. Because these are often subtle, indirect, invisible, diffuse, deceptive, and shrouded in benign justifications, this is clearly the more difficult task.

Two decades later the hot-button cultural themes of threat, civil order, and security that Lewis emphasized are in greater ascendance and have been joined by the siren calls of consumption. If our traditional notions of liberty disappear, it will not be because of a sudden coup d'état. Nor will the iron technologies of industrialization be the central means. Rather, it will occur slowly, with an appeal to traditional American values in a Teflon- and sugar-coated technological context of low visibility, fear, and convenience.

Beowulf Cluster Computing with Windows

Yeah, I mean Harvard and MIT have never contributed anything to phys... oh, never mind.

...is provably collision-resistant.

http://senderek.de/SDLH/

'Proof' by Ron 'RSA' Rivest...

http://diswww.mit.edu/bloom-picayune/crypto/13190

SDLH is simple and secure to any number of bits of security desired once set up properly.

Factoring the modulus in SDLH is the only way to break it.

For that you need a state of the art number factoring algorithm (currently General Number Field Sieve or Shor's Algorithm).

Case closed.

Increase your odds for the suit -- its inventor upgraded it recently.

Seriously, DARPA has been working with MIT through the Institute for Soldier Nanotechnologies to develop advanced armor, apparently including powered armor.

For the average Internet user, this news is not a cause for panic. No one is going to be breaking digital signatures or reading encrypted messages anytime soon. The electronic world is no less secure after these announcements than it was before.

A short note about the attack has been available for a couple of years as well. The note shows collisions for two different reduced versions of SHA-1.

Though it's not absolutely certain, my guess is that the reality behind the new announcement is that they've actually found a collision for the full version of SHA-1, and possibly for MD-5 as well. OTOH, maybe the mention of MD-5 is just a journalist's hashed (no pun intended) version of the fact that SHA-1 is based closely enough on MD-5 that an algorithm that's successful against SHA-1 will probably be effective with respect to MD-5 as well.

Dilbert on banning camera phones

Clearly the voice of someone who's never seen Random Hall's laundry server :)

MIT has deployed wireless all over the place. I go there once a year for the MIT Mystery Hunt. I was amazed last year as part of the hunt took us into a boiler room deep inside the basement of some building. The boiler room was a maze of pipes and of questionable safety, but screwed to the wall was a WAP; in the boiler room.

MIT has deployed wireless all over the place. I go there once a year for the MIT Mystery Hunt. I was amazed last year as part of the hunt took us into a boiler room deep inside the basement of some building. The boiler room was a maze of pipes and of questionable safety, but screwed to the wall was a WAP; in the boiler room.

IST? Are you sure you don't mean ITS?

As far as DHCP needing registered MACs, you're half right. If you need your computer to only use one IP address, then yes, you would need to "lock-in" a MAC address to that IP in the DHCP configuration. If you don't mind who gets what IP, you can just set up a pool of addresses and let DHCP assign leases to those addresses for a period of time -- not good for something such as a web server. The latter is what most home routers do for wired connections.

The parent poster is a student at MIT. You're obviously not.

He doesn't mean ITS. He means IST, the on-campus group that keeps the networks running. They have little cars that they use to run around campus and fix stuff that breaks.

As for needing to register MACs, he's talking about the MIT network specifically, not DHCP in general, so he's completely right, not half right. The MIT wireless will refuse to hand you an IP address until you register your MAC and provide credentials (either by logging in, or by identifying yourself as a visitor). Students have to clone the MAC addresses of DS's and Wiis temporarily so that they can register those MACs in order to get them on the network.

Estimated time to colonize the galaxy http://stuff.mit.edu/people/etekle/Articles/0700cr awfordbox3.html by Ian Crawford, an astronomer in the department of physics and astronomy at University College London. Taken from Scientifc American July 2000 issue http://stuff.mit.edu/people/etekle/Articles/aliens .html

Estimated time to colonize the galaxy http://stuff.mit.edu/people/etekle/Articles/0700cr awfordbox3.html by Ian Crawford, an astronomer in the department of physics and astronomy at University College London. Taken from Scientifc American July 2000 issue http://stuff.mit.edu/people/etekle/Articles/aliens .html

No, I meant more like these-

http://news.bbc.co.uk/1/hi/world/americas/4097267. stm - Missle defence shield test fails
http://www.cbsnews.com/stories/2005/01/12/national /main666433.shtml - Missle defense fails again
http://www.counterpunch.org/stclair09232006.html - Star Wars Goes Online...Crashes
http://news.bbc.co.uk/2/hi/americas/824828.stm - Test failur fuels skepticism
http://www.newscientist.com/article.ns?id=dn2924 - US missile defence test dodges decoys
http://www-tech.mit.edu/V125/N5/long4_5.5w.html - Missile Defense System Test Fails
http://en.wikinews.org/wiki/Second_U.S._anti-missi le_defence_system_test_fails -Second U.S. anti-missile defence system test fails

And if you look worldwide, rather than at just the Atlantic, they were, this last season.

The Atlantic didn't have many hurricanes, which is usual in an El Nino year.

The most recent Windows Video, RealVideo, and Sorenson video (Used by QuickTime prior to H.264) are what I might call proprietary in that there is no public standard. They are likely covered by patents as well. MPEG-4's AVC or H.264 may be publically documented but is covered by patents which put free software implementations in the same legal mud as any proprietary codec. On the audio side, MP3 and AAC are in similar situations: publically available but patented.

For publically available and unpatented codecs, as far as I know we've got MPEG-1 video and MPEG-1 Layer II audio (MP2), and this only because they are so old that patents have expired. Vorbis is a modern unpatented audio codec, and Theora is covered by patents which have been freely licensed to the public making it effectively patent free.

MPEG-1 suffers because it simply isn't as good in a quality vs. bitrate sense as the modern video codecs. It's possible that bandwidth and disk size increases could this help somewhat. MPEG-1 is also only good at resolutions near or below 352x240, even from higher resolution source material. However, MPEG-1 remains the most likely to be playable on a given computer. The one additional advantage it has is that decoding may be less CPU intensive than, say, H.264.

Ogg Theora with Vorbis audio is nice, and competitive with though inferior to other modern codecs in terms of quality vs. bitrate. Like all the Xiph codecs, Theora suffers from obscurity. Version 1 is not officially released, but there are limited tools for playback but little for encoding Theora. Xiph's QuickTime Components recently added Theora playback to the QuickTime system (Mac and Windows). Some Theora filters for DirectShow (Windows Media) support playback and, for the intrepid, encoding.Due to differences between the Ogg container format and other established containers, it has had some trouble properly interfacing with DirectShow for example, but playback does work fairly well. Real's Helix system (Helix Player for Linux, RealPlayer 10 on Windows) also has plugins for Theora playback and encoding.

As an occasional dabbler in filmmaking, I release my videos in Ogg Theora+Vorbis and MPEG-1 Video+MP2 audio. I have released versions using MPEG-4 ASP video (DivX, XviD, ffmpeg's mpeg4) and MP3 audio, yet I lack the necessary patent licenses for that and am somewhat reluctant to do so. Commercial software such as DivX typically extends its patent license to its customers for various uses (I believe Microsoft does not allow commercial use of Windows Media Video by default).

Wheras your extensive research, which you will no doubt supply in a follow-up to this comment, shows clearly that placing no limits on copying is better.

Wheras your extensive research, which you will no doubt supply in a follow-up to this comment, shows clearly that placing no limits on copying is better.

Wheras your extensive research, which you will no doubt supply in a follow-up to this comment, shows clearly that placing no limits on copying is better.

Next thing you know, they'll be mandating a terrorism futures exchange

Oh, wait...

Only partially true -- MIT did want to create a medical school, partnering with nearby Mass. Gen Hospital several decades ago. The rumor goes that Harvard did not want to give up one of its premier hospitals, and the HST joint program was formed. I (and other attendees) graduate with an MD degree, and get two degrees from both Harvard and MIT.

http://hst.mit.edu/

Ironic what you say about MIT and biology, considering that MIT has one of the top 3 genome sequencing centers in the country, the Broad Institute.

The SAT is a proxy IQ test. It's good enough that most high IQ societies will accept sufficient SAT scores in lieu of an IQ test. The only place IQ has been discredited is in the popular mind. The military and education system are still firm practitioners, simply because the concept works.

The SAT is taken in high school, way before any of these colleges can "work their magic".

Caltech has the pick of the high IQ (but smaller numbers of students), MIT follows, and then come the other Ivy League schools not far behind. See the attached link.

If you notice that IQ is roughly normally distributed (especially in a genetically similar population), look at the population of high IQ college age kids in the USA, and then compare to the populations of the elite US schools, you will see that they are very similar. It did not happen that way by accident.

Hell, put the student population of Caltech in your local community college and you'd find all sorts of revolutionary science suddenly springing from there too.

The US government prevents the corporate world nabbing the A-list by banning IQ tests in job interviews. Thus corporations use the proxy of school (or in the case of companies like M$, they ask questions that serve as a proxy IQ test). In the popular mind, the cause and effect gets confused between the brand (MIT/Harvard/Yale etc.) and the student body (high IQ/SAT scoring individuals). Universities don't exactly have a huge financial incentive to dissuade people either.

http://www-tech.mit.edu/V111/N41/usnews.41n.html

On the other hand, if you want to design a cannon that will destroy the moon, go to Caltech.

Or just go to MIT and steal it from Caltech
http://hacks.mit.edu/Hacks/by_year/2006/mitcannon/

The Sloan School at MIT is ranked as one of the best business schools in the country: http://entrepreneurship.mit.edu/national_rankings. php and offers both undergraduate and graduate degrees. Mind you, as an engineering alum, I don't give it much credit. chiefthe

education is one of the hardest of services to deliver

internet access + http://ocw.mit.edu/ = better education

The "pull yourself up by the bootstraps" argument doesn't hold a lot of water either.

Isn't that how the western world got to this point of development? Sure the issues you've raised are real and quite difficult, but our forefathers went through that stuff too. Hopefully it will take the poorest countries less time if we tell them the lessons we learned along the way.

There was an article on Slashdot the other day mentioning MIT's Open Course Ware. I've known about this for two years, but there is a GREAT class (w/ all videos) called 6.002 Circuits and Electronics, Fall 2000. It has all of video lectures! http://ocw.mit.edu/OcwWeb/Electrical-Engineering-a nd-Computer-Science/6-002Circuits-and-ElectronicsF all2000/CourseHome/index.htm

MIT Open Courseware

The name of it escapes me right now, but I did take a class where we reviewed certain classic software failures. (A good class for me since I'd already read about them :).

If you'd like to read a few, check out:
Therac-25 (Race conditions, software lockouts in lieu of hardware)
London Ambulance Service (Poor software design and design process)
Ariane 5 (Cutbacks on testing procedures, inappropriate software re-use, variable overflows, flight hardware allowed to generate error output)

then there's the Denver airport baggage system, the Mars Climate Orbiter, etc.

In general, you may want to read the Risks Digest, where stuff like this happens every month!

Since 'Mutually Assured Destruction' doesn't work when defending against a patent troll (they don't produce anything that can infringe your patents), maybe it would be a start to limit the maximum amount of awarded damages to $5M (for example.) If a patent of a legitimate inventor is ever infringed by some big business, the settlement is more than enough for them to retire and continue inventing if they so choose. Wouldn't it however remove the incentive for trolls, spending easily as much as $5M in the hopes of getting awarded $200M?

Then as an encore someone might be able to convince USPTO there isn't such a thing as software patents. Good reading materials: An industry at risk, So Small a Town, So Many Patent Suits

One of the courses had this comment in the online syllabus material:

"You should attend all the classes, no matter what. Nothing you could read will replace what goes on in class."

(link)

While it does mean spending money, there's nothing stopping you from buying the books, though the prices on them can be... significant. For Single Variable Calculus (the first math course listed), the book used is Calculus with Analytic Geometry , which seems to go for about $150 new no matter where you look. The book is also used in Multivariable Calculus and Principles of Aeronautic Control, so at least it can be spread out a bit.

Ow.

While it does mean spending money, there's nothing stopping you from buying the books, though the prices on them can be... significant. For Single Variable Calculus (the first math course listed), the book used is Calculus with Analytic Geometry , which seems to go for about $150 new no matter where you look. The book is also used in Multivariable Calculus and Principles of Aeronautic Control, so at least it can be spread out a bit.

Ow.

While it does mean spending money, there's nothing stopping you from buying the books, though the prices on them can be... significant. For Single Variable Calculus (the first math course listed), the book used is Calculus with Analytic Geometry , which seems to go for about $150 new no matter where you look. The book is also used in Multivariable Calculus and Principles of Aeronautic Control, so at least it can be spread out a bit.

Ow.

Don't get me wrong: Having the material available for free is great, even though a large part of the courses are incomplete in that they refer you to the standard literature for reference like most regular university courses will. But this is basically a logistic solution, a lot of knowledge is available today to anybody who can get hold of a library card at the local university and a lot of basic knowledge is no further away than the wikipedia.

But you will find that the number of people studying advanced calculus or Sino-Tibetian languages outside of university courses is small, even though a lot of material is available for free. Learning complex subjects is a process, not just a question of getting the information, and the process (with tutorials and working with other students and asking questions and assignments and so on) is what MIT is still selling, the content of OCW is only a small part of that.

Fortunately OCW is not simply free, but (at least partly) licensed under a Creative Commons license allowing non commercial sharing and remixing (Attribution-NonCommercial-ShareAlike 2.5). While you may not be able to replicate the experience of studying at MIT, someone may take the content and add e.g. a technical communications layer.

You are into advanced web 3.0 elearning platform development, but have no way to create the content? Take OCW, reuse what they have and give the world a new learning experience? You always wanted to write a shoot-'em up game based on and explaining the principles on quantum physics? You solve the DirectX/OpenGL/game engine magic and compensate your lack of talent as a physics tutor by using parts of 8.04 Quantum Physics I, Spring 2006.

These are primitive ideas, but I think about OCW more as a basis on which people can experiment than a library. Libraries have been around for a long time, unfortunately the majority of people don't use them. To reach the masses, you have to somehow turn the content of OCW into something compatible to a game console. Give it a shot!

O'Caml has side effects and is not a pure functional language.

Also, see the work done on pH (Parallel Haskell): http://www.csg.lcs.mit.edu/projects/languages/ph.s html.

There are even better ways to undo a smudge or motion blur. The linked process takes about an hour to compute, but gives impressive results... I give it three years til it shows up in Photoshop.

The mosaic effect cannot be directly deconvolved the same way; the approach in the article is based on a priori knowledge of the original image, which greatly constrains the problem and makes it well-posed. The basic reason it works is that different numerals use different amounts of ink, a property which is not completely destroyed by blurring or mosaicing. An approach such as this one (scroll down to Figure 8) would solve this problem, and also make documents less readable by someone looking at them from further away.

There are even better ways to undo a smudge or motion blur. The linked process takes about an hour to compute, but gives impressive results... I give it three years til it shows up in Photoshop.

The mosaic effect cannot be directly deconvolved the same way; the approach in the article is based on a priori knowledge of the original image, which greatly constrains the problem and makes it well-posed. The basic reason it works is that different numerals use different amounts of ink, a property which is not completely destroyed by blurring or mosaicing. An approach such as this one (scroll down to Figure 8) would solve this problem, and also make documents less readable by someone looking at them from further away.

Just so that everyone is aware, MIT's OCW program has been offering the lectures, course materials, etc. for almost all of their undergrad courses. They only provide the bibliographic info for copyrighted reading materials though (not links to the actual text) - you'll have to find that yourself.

The course was taught by Keith Winstein, the same guy who as senior editor of The Tech interviewed Jack Valenti and showed him his DeCSS Perl script.