according to that article, they used LIGO laser to generate gravitational waves, but they state "it's so feeble you can't detect it directly". Spinning a large mass you say? Wouldn't spinning 2 dense masses around each other be more effective? Like the moon around the earth? Or a binary star? Spinning a large mass isn't as important as the density of the mass. At least that's what I remember in my college kinematics class. I'm not a physicist.
IBM Watson is for all purposes dead. IBM has pumped billions into the group and tried to make a profit. They've only had a few high profile "dog and pony" shows, but no profit yet. The CEO is correct ML is the future, but just not with IBM. They will do what they usually do, they'll buy someone to get a foot hold.
If you objectively ask people about genetic engineering, I'm going to guess they are more open to it. The issue with GMO isn't science, but the companies rushing products out and then screwing up farmers. For example, what Monsanto does to farmers like making seeds that can't reproduce (ie you have to buy seeds every year) and going after organic farmers. This is all well documented. The issue is the scientists that work for these evil corporations stay silent and are complicit with immoral practices. Not all corporations are evil, but the big multi-national corps rubs the public the wrong way.
People don't have a problem with science being used for good, but when Monsanto releases a new pesticide that destroys crops of other farmers, they have every right to scream. Clearly the author of the article is s shill for big agriculture and isn't actually reporting.
this freakin stupid fad of thinner and lighter laptops passed the point of diminishing returns several years back. No, I want a laptop that lasts longer, not one that gets less batter life just so it can be.02mm thinner. All the stupid people going stupid over "it's thinner" are partly to blame.
They are kind of vague, but I'm gonna guess they mean they trained a CNN segmentation model on a billion items and gets a good F1 score that's above 95%. It is impressive, but that's really just a matter of applying the latest segmentation models on a dataset of 1 billion labeled images. I would be more impressed if google made their dataset available for ML researchers to use. The open datasets like Coco and Imagenet aren't nearly as big.
uhh, are you serious? The smartest and best code. Clearly you haven't actually read much open source code or worked with open sores people who think they are the smartest guy in the room. There's lots of good open source programmers, but lets be totally blunt. 99% of the code in open source world is ugly and sucks. Sucky code still works, but it's not clean, elegant, designed properly or well documented. Go look at Google's Tensorflow code base. The 1.9 is kind of mess and Google has finally admitted it. That's why 2.0 is going to clean out a bunch of stuff and make it easier. Look at linux from 1998 and compare it to 2018. Just about every part of the Kernel has gone through significant rewrite and clean up. I've been contributing to open source since 2002 and I've literally read more than 1 million lines of open source code in different languages for a wide variety of projects. The biggest problem in OSS isn't CoC. It's jerky assholes who think they write perfect code, behave like assholes and feel coding gives them a right to be a total dick.
I've participated in my fair share of flame wars and it's not pretty. I won't bore you and cite 101 examples of shitty code written by good programmers. I get lots of guys feel defensive because they know their behavior has been bad in the past. Get over it and start acting like how you want to be treated.
As others have pointed out, it was a photoshop and not a real photo. By law, they're supposed to turn that evidence over to the FBI. Since the FBI already said that it didn't happen and they have no evidence, I would say Bloomberg isn't reliable.
This is the same Bloomberg that runs news story suggested by Wall Street elite to pump and dump stocks. This is the same Bloomberg that is the unofficial marketing arm of Wall Street. This is the same Bloomberg that has been saying regulations aren't needed anymore because the market can regulate itself, except that they know they can't. So what good evidence do you have that Bloomberg isn't more than just a wall street marketing machine?
It's pretty clear a big percent of the news papers based in NYC do shoddy reporting and in some cases, it's out right propaganda for Wall Street. If Bloomberg had proof, they should have produced it and given it to the FBI for verification.
The biggest red flag on the Bloomberg report is it's a sorry hack attempt. To put an additional chip on a board that can easily be caught with automated visual scans (ie computer vision) is just sloppy and stupid. There are so many other ways to compromise a MB without leaving a visual trace. Plus given the Intel CPU bugs, why go through the hassle? You can more easily root an OS without leaving physical evidence. A good attacker knows how to erase their digital foot print. If it really was the chinese government, I seriously doubt they'd be that stupid. From a spying perspective, none of the big countries with the resources would choose such a stupid attack.
did you read the article? I don't like the article, but that was not the message. The author is making the point that all A isn't beneficial and people shouldn't obsess over it. Getting A is not a measurement of understanding the material. Even in engineering majors, plenty of people graduate with 4.0 and end up being complete crap in the work place.
People should stop giving a crap to Forbes top X schools, because it's just BS non-sense. There are lots of good colleges that are affordable. For undergrad you basically don't get shit from the top universities like Harvard, Yale, Stanford, Berkeley, etc. Good luck seeing your teachers during their office hours at those universities. Go to an affordable school and kick ass, then apply to a top research university. The only good reason to go to stanford, MIT or CMU is so google will hire you, since they are degree bigots. The only good reason to go to yale and harvard is to meet rich spoiled kids and build your social network. People who interview applicants for those schools are told answer this question "will this applicant hold up our prestige?" In other words, it's not about you, it's about them and maintaining their exclusive elite status.
Anyone that's actually tried using AWS will know first hand how shitty their docs are. Many of the web console docs are several releases out of date and the screen captures look nothing like the current version. Then there's commonly used stuff that has crap docs. To the point where it's basically unusable docs. If their lessons are like their docs, it will suck ass and be worthless. There are plenty of good free resources that teach various tools like tensorflow, pytorch, cafe2, keras, deeplearning4java and mxnet.
Honestly, most open source programmers I know suck at UX. Even if they were to try, it would still suck. It would be less shitty. I've been contributing to open source for over 15 years and I've done a bit of rich UI, but honestly I couldn't build a great UX to save my life. Someone that is truly great at UX thinks differently than some one that codes. There might be a rare bird out there that is a badass programmer and great UX designer, but finding one with time is basically impossible. I've known a few good UX designers and they are in high demand. So much so that they demand high pay and often are over booked. You have a better chance of finding someone ok at coding and great at UX.
last time I tried Pypy it wasn't 100% compatible with standard python and it was only slightly faster for doing computer vision. In the end, the hassle of getting Pypy setup in Raspbian just wasn't worth it. The worse parts of Python is the VM lacks a rich set of debugging API like JVM and CLR, so people are still forced to use command pdb, the documentation kinda sucks and the multi-threading still sucks after all these years. As a functional language, the syntax is a little weird and quirky. Smalltalk and LISP have a more consistent and well thought-out design. Python grew organically, and still doesn't have an official specification.
Yup, I'm happy to be an idiot that's spent tens of thousands of hours contributing to various open source projects. I'm glad there are thousands of other people that feel it's more important to contribute to society than acting like a greedy asshole. I will remind you that linus did work for free for many years before linux foundation was created to support linus. I don't agree with how linus treats people or his poor communication skills, but he earned his position. Very few programmers have made such a big contribution. Even though I hate GIT and curse it daily, the work he's done since the early 90's is why he deserves that salary. I remember using slackware and was lucky enough to see linux grow. Compare linus to say steve jobs, Jobs was a bigger asshole and couldn't code himself out of a paper bag!
If you read the book, you'll see that scientists have cured cancer in mice for more 2 decades, but it never works in humans. Anyone foolish enough to believe the hype isn't thinking clearly and has a sad understanding of science.
I guess you've never heard of bios or boot sector virus/trojan. This is well documented over the last 3 decades. There are trojans that can infect drivers or system services, which in many cases can't be automatically removed. In those cases, the best bet is to wipe the system and do a fresh install. Back in the 90's there was a particularly bad boot sector virus that bricked thousands of systems. That was before bios had any virus protection. These days most MB have bios virus protection, so bricking a MB rarely happens. On linux, attackers used to break in, upload a root kit and recompile the kernel. In that case, your only choice was to wipe the HD.
the most obvious reason is binocular vision for more AR stuff. The current setup of 2 different cameras aren't idea for binocular vision. Not that it can't be done, but typically binocular vision is done with two identical cameras. The problem though is the cameras have to be oriented horizontally and have enough distance between them to get good results. That would mean a dramatic change in design, which is highly doubtful.
how about stop giving money to the fat dumb asses that own the companies and train the employees in new jobs. Ones that are in demand and will be in demand for decades.
NVidia already announced the DGX-2 appliance a few weeks back. Now they announce HGX-2, which apparently is the design for companies that want to build their own AI cloud. But I have to ask, given the recent shortage of chips, how would any third party build their own DGX-2 from the reference design if they can't get the chips? Given the memory shortage, it's not like building your own DGX-2 for a AI cloud is going to be more efficient than just buying DGX-2 from Nvidia.
you're kidding right. A 5 year old knows why roosters crow in the morning. Other things like linear algebra can take decades or even a life time for some people, but some people have a natural ability for math. Some people have a natural physical intelligence and end up in professional sports, dance or performance. Humans generally need much less data than current approach with CNN and RNN. Now if you compara Neurala's approach to CNN from 2 years ago, it doesn't need billions of images to train. So I think you'd better read up on the current state of Art in AI and look at how humans actually learn. Over generalizing for the sake of an argument isn't working and is just plain wrong.
clearly you've never lived in LA. The streets in southern CA are quite wide and the 405 freeway is 4 lines in each direction from Orange county to Sherman oaks. There are sections where it's 5 lanes each side. The problem isn't the roads aren't wide enough. The problem is the car culture of southern CA and the city planning. For anyone that knows the history of CA, you'd know the auto industry played a big hand in the design and zoning. Basically, the design forces people to drive 1-2 miles to do basic things like buy groceries. Compare that to European cities where you can walk from your house to buy groceries and run basic errands.
If you take a minute to look at southern CA on google maps, you'll see there's a freeway every 4-5 miles running north-to-south and east-to-west. Many streets are 3 lanes each side like Hawthrone and PCH. Don't take my word for it, look on google maps. I grew up in LA area and it's gotten worse since the 80's.
Slashdot Mirror
according to that article, they used LIGO laser to generate gravitational waves, but they state "it's so feeble you can't detect it directly". Spinning a large mass you say? Wouldn't spinning 2 dense masses around each other be more effective? Like the moon around the earth? Or a binary star? Spinning a large mass isn't as important as the density of the mass. At least that's what I remember in my college kinematics class. I'm not a physicist.
IBM Watson is for all purposes dead. IBM has pumped billions into the group and tried to make a profit. They've only had a few high profile "dog and pony" shows, but no profit yet. The CEO is correct ML is the future, but just not with IBM. They will do what they usually do, they'll buy someone to get a foot hold.
People don't have a problem with science being used for good, but when Monsanto releases a new pesticide that destroys crops of other farmers, they have every right to scream. Clearly the author of the article is s shill for big agriculture and isn't actually reporting.
this freakin stupid fad of thinner and lighter laptops passed the point of diminishing returns several years back. No, I want a laptop that lasts longer, not one that gets less batter life just so it can be .02mm thinner. All the stupid people going stupid over "it's thinner" are partly to blame.
They are kind of vague, but I'm gonna guess they mean they trained a CNN segmentation model on a billion items and gets a good F1 score that's above 95%. It is impressive, but that's really just a matter of applying the latest segmentation models on a dataset of 1 billion labeled images. I would be more impressed if google made their dataset available for ML researchers to use. The open datasets like Coco and Imagenet aren't nearly as big.
I've participated in my fair share of flame wars and it's not pretty. I won't bore you and cite 101 examples of shitty code written by good programmers. I get lots of guys feel defensive because they know their behavior has been bad in the past. Get over it and start acting like how you want to be treated.
This is the same Bloomberg that runs news story suggested by Wall Street elite to pump and dump stocks. This is the same Bloomberg that is the unofficial marketing arm of Wall Street. This is the same Bloomberg that has been saying regulations aren't needed anymore because the market can regulate itself, except that they know they can't. So what good evidence do you have that Bloomberg isn't more than just a wall street marketing machine?
The biggest red flag on the Bloomberg report is it's a sorry hack attempt. To put an additional chip on a board that can easily be caught with automated visual scans (ie computer vision) is just sloppy and stupid. There are so many other ways to compromise a MB without leaving a visual trace. Plus given the Intel CPU bugs, why go through the hassle? You can more easily root an OS without leaving physical evidence. A good attacker knows how to erase their digital foot print. If it really was the chinese government, I seriously doubt they'd be that stupid. From a spying perspective, none of the big countries with the resources would choose such a stupid attack.
did you read the article? I don't like the article, but that was not the message. The author is making the point that all A isn't beneficial and people shouldn't obsess over it. Getting A is not a measurement of understanding the material. Even in engineering majors, plenty of people graduate with 4.0 and end up being complete crap in the work place.
People should stop giving a crap to Forbes top X schools, because it's just BS non-sense. There are lots of good colleges that are affordable. For undergrad you basically don't get shit from the top universities like Harvard, Yale, Stanford, Berkeley, etc. Good luck seeing your teachers during their office hours at those universities. Go to an affordable school and kick ass, then apply to a top research university. The only good reason to go to stanford, MIT or CMU is so google will hire you, since they are degree bigots. The only good reason to go to yale and harvard is to meet rich spoiled kids and build your social network. People who interview applicants for those schools are told answer this question "will this applicant hold up our prestige?" In other words, it's not about you, it's about them and maintaining their exclusive elite status.
Brandon Rohrer
Aurélien Géron
then there's articles on Medium and various discussion channels for Keras, pytorch, tensorflow, cafe2, DL4J and various Github projects.
Anyone that's actually tried using AWS will know first hand how shitty their docs are. Many of the web console docs are several releases out of date and the screen captures look nothing like the current version. Then there's commonly used stuff that has crap docs. To the point where it's basically unusable docs. If their lessons are like their docs, it will suck ass and be worthless. There are plenty of good free resources that teach various tools like tensorflow, pytorch, cafe2, keras, deeplearning4java and mxnet.
I hate both parties. Both parties suffer from personality worship disorder. Some one needs to invent slime-antivirus to clean out DC.
Honestly, most open source programmers I know suck at UX. Even if they were to try, it would still suck. It would be less shitty. I've been contributing to open source for over 15 years and I've done a bit of rich UI, but honestly I couldn't build a great UX to save my life. Someone that is truly great at UX thinks differently than some one that codes. There might be a rare bird out there that is a badass programmer and great UX designer, but finding one with time is basically impossible. I've known a few good UX designers and they are in high demand. So much so that they demand high pay and often are over booked. You have a better chance of finding someone ok at coding and great at UX.
Google does make regular donations to wikipedia, google it and you'll see for yourself.
last time I tried Pypy it wasn't 100% compatible with standard python and it was only slightly faster for doing computer vision. In the end, the hassle of getting Pypy setup in Raspbian just wasn't worth it. The worse parts of Python is the VM lacks a rich set of debugging API like JVM and CLR, so people are still forced to use command pdb, the documentation kinda sucks and the multi-threading still sucks after all these years. As a functional language, the syntax is a little weird and quirky. Smalltalk and LISP have a more consistent and well thought-out design. Python grew organically, and still doesn't have an official specification.
Yup, I'm happy to be an idiot that's spent tens of thousands of hours contributing to various open source projects. I'm glad there are thousands of other people that feel it's more important to contribute to society than acting like a greedy asshole. I will remind you that linus did work for free for many years before linux foundation was created to support linus. I don't agree with how linus treats people or his poor communication skills, but he earned his position. Very few programmers have made such a big contribution. Even though I hate GIT and curse it daily, the work he's done since the early 90's is why he deserves that salary. I remember using slackware and was lucky enough to see linux grow. Compare linus to say steve jobs, Jobs was a bigger asshole and couldn't code himself out of a paper bag!
If you read the book, you'll see that scientists have cured cancer in mice for more 2 decades, but it never works in humans. Anyone foolish enough to believe the hype isn't thinking clearly and has a sad understanding of science.
Many google tech is designed to appease their corporate partners, so it's a feature, not a bug.
I guess you've never heard of bios or boot sector virus/trojan. This is well documented over the last 3 decades. There are trojans that can infect drivers or system services, which in many cases can't be automatically removed. In those cases, the best bet is to wipe the system and do a fresh install. Back in the 90's there was a particularly bad boot sector virus that bricked thousands of systems. That was before bios had any virus protection. These days most MB have bios virus protection, so bricking a MB rarely happens. On linux, attackers used to break in, upload a root kit and recompile the kernel. In that case, your only choice was to wipe the HD.
the most obvious reason is binocular vision for more AR stuff. The current setup of 2 different cameras aren't idea for binocular vision. Not that it can't be done, but typically binocular vision is done with two identical cameras. The problem though is the cameras have to be oriented horizontally and have enough distance between them to get good results. That would mean a dramatic change in design, which is highly doubtful.
how about stop giving money to the fat dumb asses that own the companies and train the employees in new jobs. Ones that are in demand and will be in demand for decades.
NVidia already announced the DGX-2 appliance a few weeks back. Now they announce HGX-2, which apparently is the design for companies that want to build their own AI cloud. But I have to ask, given the recent shortage of chips, how would any third party build their own DGX-2 from the reference design if they can't get the chips? Given the memory shortage, it's not like building your own DGX-2 for a AI cloud is going to be more efficient than just buying DGX-2 from Nvidia.
you're kidding right. A 5 year old knows why roosters crow in the morning. Other things like linear algebra can take decades or even a life time for some people, but some people have a natural ability for math. Some people have a natural physical intelligence and end up in professional sports, dance or performance. Humans generally need much less data than current approach with CNN and RNN. Now if you compara Neurala's approach to CNN from 2 years ago, it doesn't need billions of images to train. So I think you'd better read up on the current state of Art in AI and look at how humans actually learn. Over generalizing for the sake of an argument isn't working and is just plain wrong.
If you take a minute to look at southern CA on google maps, you'll see there's a freeway every 4-5 miles running north-to-south and east-to-west. Many streets are 3 lanes each side like Hawthrone and PCH. Don't take my word for it, look on google maps. I grew up in LA area and it's gotten worse since the 80's.