Still better than talking to a hand. Google doesnt listen to the users, they dont even have a place to post any concerns/complaints (probably because there was no way to automate it).
Define dropout. I signed up for >30 courses in total. I have 7 certs of accomplishment, all with 95-100% score. The rest I didnt bother doing quizzes/assignments, I downloaded all the videos and learn at my own pace for myself, I dont need any more diplomas. I only failed in one class I was doing quizzes/assignments, by 3% (at first they send me cert and said i passed, but later revoked it after recount/quiz fixes or something).
Im sure I am destroying their stats taking all those courses and not doing any "work".
Hopefully app acl list will be USER defined instead of app defined. How retarded is it for an app to blackmail you for Contact list/messages and phone calls just so you can play a game? How come its the app that decides and has the upper hand instead of the user? How come App can blackmail me and I cant simply REFUSE to give it data? I want to be able to define ACLs per application. I dont give a shit what app wants, it can eat a duck for all I care. Phone status? sure, feed it fake USER DEFINED status. I dont need a game to know that I have GPS enabled. You want GPS data? Sure, let me feed N38 53.86205 W77 2.19162. Its my phone damnit!
I dont know the math behind you algorithm, but it might be even possible to directly use PDM stream (http://en.wikipedia.org/wiki/Pulse-density_modulation) from those microphones without converting it to PCM first. At 3MHz per micropfone you can cram 32 mics into 12MB/s stream in a small USB dongle that costs ~$40 to make (as low as $30 in bulk if you mfg in China).
Tablet is an ideal UI for something like this. If you could modify your code to visualize source of sound like in this/. article it would sell like hotcakes on kickstarter at $200.
I'm not an EE so I won't pretend to fully understand this particular case but I like it when tech companies reach a bit and try something hard. This may or may not be a good idea but I'm still excited about it.
One from the article solves sync problems by building an array of cheap ($1 for Nokia ones) cellphone digital microphone modules. This is quite brilliant and reduces cost tremendously. All you need is one USB I/O chip with 20-30 IO pins.
How fast is your algorithm? Could it run on average (2x 1GHz A8) ARM tablet? Is it expendable to more microphones?
This smells like a perfect kickstarter project - app + USB dongle you connect to android tablet/phone.
I admit I was misleading. GPU is faster, CPU is slower. Rockchip on the other hand wins both cpu and gpu comparison (with crusty overclocked quad mali400) against fastest Tegra3.
My point was cheap Chinese SOCs aren't about bottom dollar_and_performance anymore. They are still bottom dollar, but performance starts to catch up.
GPU is not a problem, video decoder (cedar) is, but there is a working open source driver now for A10. There is also a working open source driver for MALI400. Both experimental, but nightly builds work ( as good as experimental stuff would)
Data in article was straight from packets, your conjecture is just an ass_umption you pulled out of your ass. People pirate VMWare, macs are randomly generated.
Half of Hintons course was about history and what didnt work in AI. Its great to know those things if you have interest in the field, but its not something you should start with (snorefest).
I think they are usually blasting up using sat uplink (military drones). Civilian ones use ordinary RC radios - good luck distinguishing between cordless phone, bluetooth and Fatuba working on 2.4GHz.
Slashdot Mirror
cutting silicon with laser to cripple it does NOT lower the cost.
Still better than talking to a hand. Google doesnt listen to the users, they dont even have a place to post any concerns/complaints (probably because there was no way to automate it).
I think what she meant was "You can't win the lottery if you don't buy a ticket".
Actually you win lottery EVERY TIME you dont buy a ticket, but only 1 in few hundred millions if you do buy it.
Define dropout.
I signed up for >30 courses in total. I have 7 certs of accomplishment, all with 95-100% score. The rest I didnt bother doing quizzes/assignments, I downloaded all the videos and learn at my own pace for myself, I dont need any more diplomas.
I only failed in one class I was doing quizzes/assignments, by 3% (at first they send me cert and said i passed, but later revoked it after recount/quiz fixes or something).
Im sure I am destroying their stats taking all those courses and not doing any "work".
Hopefully app acl list will be USER defined instead of app defined.
How retarded is it for an app to blackmail you for Contact list/messages and phone calls just so you can play a game? How come its the app that decides and has the upper hand instead of the user? How come App can blackmail me and I cant simply REFUSE to give it data?
I want to be able to define ACLs per application. I dont give a shit what app wants, it can eat a duck for all I care. Phone status? sure, feed it fake USER DEFINED status. I dont need a game to know that I have GPS enabled. You want GPS data? Sure, let me feed N38 53.86205 W77 2.19162.
Its my phone damnit!
You had to buy specialized routers/expansion cards for decades to do certain things. Now you reconfigure those things on the fly.
He does this because malaria was killing his cheap OIL RIG labor in Africa.
It can be owned by Elon himself.
Yes, but using Mems microphones like those
http://www.ebay.com/itm/New-Wholesale-20pcs-Mic-Speaker-For-Nokia-N8-N9-5230-5800-E63-E71-E52-E5-C5-X7-/271133891559?pt=US_Cell_Phone_Replacement_Parts_Tools&hash=item3f20d6dbe7
20 for $12.
I dont know the math behind you algorithm, but it might be even possible to directly use PDM stream (http://en.wikipedia.org/wiki/Pulse-density_modulation) from those microphones without converting it to PCM first. At 3MHz per micropfone you can cram 32 mics into 12MB/s stream in a small USB dongle that costs ~$40 to make (as low as $30 in bulk if you mfg in China).
Tablet is an ideal UI for something like this. If you could modify your code to visualize source of sound like in this /. article it would sell like hotcakes on kickstarter at $200.
I'm not an EE so I won't pretend to fully understand this particular case but I like it when tech companies reach a bit and try something hard. This may or may not be a good idea but I'm still excited about it.
Thats what she said.
its commonly called
http://en.wikipedia.org/wiki/Superresolution
Satellites have been using this technique since forever (at least the ones facing outwards from our planet).
Aperture synthesis, take 5 pictures in succession and you have your plate
Nice.
One from the article solves sync problems by building an array of cheap ($1 for Nokia ones) cellphone digital microphone modules.
This is quite brilliant and reduces cost tremendously. All you need is one USB I/O chip with 20-30 IO pins.
How fast is your algorithm? Could it run on average (2x 1GHz A8) ARM tablet? Is it expendable to more microphones?
This smells like a perfect kickstarter project - app + USB dongle you connect to android tablet/phone.
Plenty of A31 around, still If I was shopping for a cheap tabled I would look at RK3188 instead.
I admit I was misleading. GPU is faster, CPU is slower.
Rockchip on the other hand wins both cpu and gpu comparison (with crusty overclocked quad mali400) against fastest Tegra3.
My point was cheap Chinese SOCs aren't about bottom dollar_and_performance anymore. They are still bottom dollar, but performance starts to catch up.
GPU is not a problem, video decoder (cedar) is, but there is a working open source driver now for A10. There is also a working open source driver for MALI400.
Both experimental, but nightly builds work ( as good as experimental stuff would)
Allwinters most popular chip at the moment is quad Cortex-A7.
Rockchip is another Chinese manufacturer making quad A8 (RK3188)
Both are faster than fastest Tegra3.
You are right, I'm sorry. I get really agitated when someone commits fallacy of the converse.
Data in article was straight from packets, your conjecture is just an ass_umption you pulled out of your ass.
People pirate VMWare, macs are randomly generated.
If you really want to learn about _working_ AI and not "when I was a boy we did it in the snow, both ways, uphill" then do
https://class.coursera.org/ml/class
Machine Learning by Andrew Ng.
After that you can do
http://work.caltech.edu/telecourse.html
Learning from data by Yaser Abu-Mostafa
Half of Hintons course was about history and what didnt work in AI. Its great to know those things if you have interest in the field, but its not something you should start with (snorefest).
and I dont recall ads embedded in video on Youtube
Computer hehe, thats cute.
One oscilloscope can cost over $50K. Computers are practically free when it comes to hardware engineering.
No, Americans just want another way of saying their dick is bigger.
I think they are usually blasting up using sat uplink (military drones). Civilian ones use ordinary RC radios - good luck distinguishing between cordless phone, bluetooth and Fatuba working on 2.4GHz.