I already sit in front of a computer almost all day long and have instant access to a calculator. Anything I want to remember, I just copy and paste into my computer's memory. If we have brain implants, we need something more different han a math or memory coprocessor.
I blame large corporations. Large corporations look for incremental improvements. They worry about the next quarter, not about the next decade. Innovation is risky and they avoid risks.
Imagine if Thomas Edison were alive today. He would have joined a company like IBM, and his innovations would have been lost in the maze of bureaucracy.
Large companies do a great service. They take innovations and make them into reality. The problem today is there are too few innovations to be turned into reality.
What we need are more small groups of people willing to look beyond the next quarter. Google has a great example of innovation. However, I doubt Google will still be so innovative after that grow to the size of companies like Cisco.
The problem with spreadsheets is that there is no visibility. Understanding a spreadsheet that someone else wrote is very tough. It is like trying to read a C program with only one function. And all of the C variables names are 'a14' and 'b59'.
I would much rather see the code used to produce the math. A program like Matlab has any easy to understand scripting language. It is much easier to see how the calculations were made.
Excel only works when you have simple calculations. For complex math, spreadsheets are a huge mistake.
It's not only the processor speed, but also the memory bus speed, and graphics coprocessor speeds. There is a big difference between a computer today and a computer of 3 years ago.
Some examples, Google maps, PDF's of scientific papers, any web page that has complex rendering.
Computer techs and Human resources have a lot in common. Neither get their due respect. Both are essential job functions that the CEO's view as a nuisense. Both are their own "islands" inside a company. HR people mainly hang out with HR people only and the same with computer techs. Also, both are dominated by a single sex.
The same could be said for all of the other professionals that aren't in the companies core business. For example, lawyers, accountants, and technical writers.
The reason why computer techs aren't respected, is because they don't work at a company whose sole purpose is computer support. If you don't work in a company's core business, you are overhead.
Have you ever been in a code review that was a comment review? Where the reviewers, unwilling or unable to find problems in the code, nitpicked about how the comments looked like? Too many managers and second tier programmers spend their time worrying about the details of the comments and coding style because that is all they understand.
What good are great comments if the code doesn't work? Sure, comments matter, but all of the comments in the world aren't going to make buggy software run better. It will just make the next programmer understand how screwed up the original programmer was.
If given perfect information about the relationships between concepts, you could derive a very intellegent machine. TAke a human for example..
A baby hears the word mom spoken by his mom. Gradually, the baby knows there is a relationship between that sound and a smily face.
The child, growing up, starts to see relationships. Intense pain, which is rare, when correlated with a hot stove, has strong meaning in his mind.
Everything is learned initially through correlations. The advantage of human beings is that there are many more data points for correlation. Google's correlations are weak and don't give nearly as much information.
My point was that "Atlas Shrugged" explains why communism will never work. If you can ignore her philosphy, you can realize that she explains very well the problem with communism. She even lived at one point in a communist block country so she has experienced all of these problems. How many of us can say that??
The problem is that communism is a contradiction. It requires that there are no rights to keep what you produce. Most people suddenly lose any desire to produce and the system can not support itself.
She wrote this is 1955. Yet, this is what later happened to China (which was forced to switch to capatalism light), and to the Soviet Union.
I sometimes wonder why we don't do something similar at our land borders? We are protecting our sea borders by detecting all ships. Why not create land buoys to detect illegal crossings.
In a city like Tokyo, this would make the traffic worse. When the cars move quickly, the road will support a very high auto density. When there is a traffic jam, the road will support a much lower density.
It used to be, that some of the side roads, due to not being known, would have heavy traffic, but not enough to cause a traffic jam. These side roads were near there maximum capacity. As soon as the side road ussge increeased, the road turns into a traffic jam, and the maximum capacity of the side roads decrease
Now, the total delays for the entire system are increased. Instead of helping, these navagition systems just create traffic jams on side roads that were never had a problem before.
ASICS are the key.. At it's core, artificial intellegence is a very very complex signal processing problem. High speed signal processing algorithms are typically simulated on computers and implemented in ASICs. I would guess we could probably get 10-100x the power by implementing the AI algorithms in ASICS. The key is to get the algorithms to work first.
It seems like there is only one good solution. That is for only _COMMERCIAL_ VOIP should be regulated, and even those regulations should be kept to a minimum. Free VOIP programs should have no regulations.
The problem with not regulating commercial VOIP services is that most of the telephone companies are switching over to VOIP. There needs to be some sort of regulation to support a national telephone directory, emergency calling, and other services.
On the other hand, free services should be able to do whatever they want. If someone wants to use an open source program to call a friend in another state, that should be their own business.
Lindbergh was NOT the first person to fly across the Atlantic. The first person to fly in an airplane was Albert Read when Lindbergh was only 17. In fact, several other pilots flew across the Atlantic before Lindbergh
The Wright brothers were not the first to fly an airplane. They were just the first to be photographed while flying an airplane.
Name almost any successful product that has been around for 10 years. That product was probably developed by someone else first. That's true for operating systems, search engines, cars..
Being first doesn't matter.
I finally downloaded firefox to see which I would prefer. I really like the security and some of the other features.
However, I won't be switching unless the following 2 problems are fixed..
1- You can't read the text easily when scrolling. I'm not sure why internet explorer's scrolling works so much better.
2- The favorites panel. I really like to use the favorites panel on the left of the screen in internet explorer. The favorites panel provided by firefox is very clunky.
The problem with NASA is that it's primary mission is to increase it's funding. Why else are they talking about sending a man to Mars? It appeals to the taxpayers hearts more than a space probe ever will. These days, NASA only has a few programs that are expanding the field of science. Hubble and their space probes.
I would like to see NASA gutted except for non-manned space craft. Then I would like to see the 100 billion dollars or so left over given to private groups headed by people like Burt Rutan. Think what Burt Rutan could do with $10 billion in funding.
I think this will be a great tool for researchers at universities and companies. No longer will you have to contact the hardware group, beg and plead, and wait forever just to make a simple prototype. Instead, just send a simple an online order, and they will deliver the prototype to your doorstep.
Look at your own address zippy@cs.brandeis.edu. Isn't this a nonsense statement?
"zee,eye,pee,pee,why,at,sea,s,dot,bee,our,a,en,de, e,eye,es,dot,e,de,you
I think it's impossible for voice recognition to be good enough. Let's say your name is bob Barnes. You choose an address bobb@yahoo. You say the word 'bobb'. This can either be bobby, bobbee, bobe, boby, baahbe, bahbe, etc... There isn't a one to one match between the word and the enunciation.
The problem is the form factor of the cell phone. Cell phones typically have a numeric keypad to keep the size reasonable. The easiest way to input a phone number is to enter numbers.
The biggest current trend is that everyone is switching to wireless phones. Most people don't want to carry around a phone large enough to contain a keyboard. Voice recognition works well only for words that are commonly used. For weird IP addresses, you would have to say each letter one at a time.
Imagine you meet somebody. You want to store his/her phone number your phone book on your cell phone. Which is easier? Typing 820-833-5214 or typing a 16 letter word into your 10 button keypad?
In order for Kurzweil's AI to work, it isnt enough to have 100Tflop. Kurzweil based the 100Tflop number on the idea that we would have algorithms available to replace the brain functions such as vision, pattern recognition, etc.. Just using a simple neural simulator would require many orders magnitude more of power to simulate a human brain. We are a long ways from having the needed algorithms (probably 30 years at a mininum).
Slashdot Mirror
I already sit in front of a computer almost all day long and have instant access to a calculator. Anything I want to remember, I just copy and paste into my computer's memory. If we have brain implants, we need something more different han a math or memory coprocessor.
I blame large corporations. Large corporations look for incremental improvements. They worry about the next quarter, not about the next decade. Innovation is risky and they avoid risks.
Imagine if Thomas Edison were alive today. He would have joined a company like IBM, and his innovations would have been lost in the maze of bureaucracy.
Large companies do a great service. They take innovations and make them into reality. The problem today is there are too few innovations to be turned into reality.
What we need are more small groups of people willing to look beyond the next quarter. Google has a great example of innovation. However, I doubt Google will still be so innovative after that grow to the size of companies like Cisco.
Steve to Bill: We finally did it. We included every possible feature in word, excel in power point. What else can we do?
Bill to Steve: Add more auto correcting. Users don't know what they mean to write. Let the software write the entire document for them.
Steve to Bill: Users might actually figure out that auto-correcting is worthless. They might start buying _less_ copies.
Bill to Steve: Lets try a different file standard. Then everyone else will have to upgrade if one person upgrades.
Steve to Bill: We've already upgraded to XML and changed file formats every time. Customers are going to freak out.
Bill to Steve: Lets make a new version of XML. No one will complain and everyone will upgrade.
The problem with spreadsheets is that there is no visibility. Understanding a spreadsheet that someone else wrote is very tough. It is like trying to read a C program with only one function. And all of the C variables names are 'a14' and 'b59'. I would much rather see the code used to produce the math. A program like Matlab has any easy to understand scripting language. It is much easier to see how the calculations were made. Excel only works when you have simple calculations. For complex math, spreadsheets are a huge mistake.
It's not only the processor speed, but also the memory bus speed, and graphics coprocessor speeds. There is a big difference between a computer today and a computer of 3 years ago. Some examples, Google maps, PDF's of scientific papers, any web page that has complex rendering.
Computer techs and Human resources have a lot in common. Neither get their due respect. Both are essential job functions that the CEO's view as a nuisense. Both are their own "islands" inside a company. HR people mainly hang out with HR people only and the same with computer techs. Also, both are dominated by a single sex.
The same could be said for all of the other professionals that aren't in the companies core business. For example, lawyers, accountants, and technical writers.
The reason why computer techs aren't respected, is because they don't work at a company whose sole purpose is computer support. If you don't work in a company's core business, you are overhead.
Have you ever been in a code review that was a comment review? Where the reviewers, unwilling or unable to find problems in the code, nitpicked about how the comments looked like? Too many managers and second tier programmers spend their time worrying about the details of the comments and coding style because that is all they understand.
What good are great comments if the code doesn't work? Sure, comments matter, but all of the comments in the world aren't going to make buggy software run better. It will just make the next programmer understand how screwed up the original programmer was.
You mentioned you use Yahoo for maps. Try maps.google.com and you will never go back.
If given perfect information about the relationships between concepts, you could derive a very intellegent machine. TAke a human for example..
A baby hears the word mom spoken by his mom. Gradually, the baby knows there is a relationship between that sound and a smily face.
The child, growing up, starts to see relationships. Intense pain, which is rare, when correlated with a hot stove, has strong meaning in his mind.
Everything is learned initially through correlations. The advantage of human beings is that there are many more data points for correlation. Google's correlations are weak and don't give nearly as much information.
My point was that "Atlas Shrugged" explains why communism will never work. If you can ignore her philosphy, you can realize that she explains very well the problem with communism. She even lived at one point in a communist block country so she has experienced all of these problems. How many of us can say that??
The problem is that communism is a contradiction. It requires that there are no rights to keep what you produce. Most people suddenly lose any desire to produce and the system can not support itself.
She wrote this is 1955. Yet, this is what later happened to China (which was forced to switch to capatalism light), and to the Soviet Union.
You should read "Atlas Shrugged" by Ann Rand. That would answer your question.
I sometimes wonder why we don't do something similar at our land borders? We are protecting our sea borders by detecting all ships. Why not create land buoys to detect illegal crossings.
In a city like Tokyo, this would make the traffic worse. When the cars move quickly, the road will support a very high auto density. When there is a traffic jam, the road will support a much lower density.
It used to be, that some of the side roads, due to not being known, would have heavy traffic, but not enough to cause a traffic jam. These side roads were near there maximum capacity. As soon as the side road ussge increeased, the road turns into a traffic jam, and the maximum capacity of the side roads decrease
Now, the total delays for the entire system are increased. Instead of helping, these navagition systems just create traffic jams on side roads that were never had a problem before.
Give him a blaster, and Asimo could do a really good stormtrooper impression.
ASICS are the key.. At it's core, artificial intellegence is a very very complex signal processing problem. High speed signal processing algorithms are typically simulated on computers and implemented in ASICs. I would guess we could probably get 10-100x the power by implementing the AI algorithms in ASICS. The key is to get the algorithms to work first.
That's like saying Fox News doesn't have a conservative Bias or that CNN isn't liberal.
There's nothing immoral about being biased. Most newspapers slant one way or the other.
I would bet that atleast 4/5 of the employees at the Times are democrats. Most of the articles are either even handed or have a liberal slant.
Maybe the times doesn't seem very liberal to someone who is a liberal.. But to anyone else, it is easy to see.
It seems like there is only one good solution. That is for only _COMMERCIAL_ VOIP should be regulated, and even those regulations should be kept to a minimum. Free VOIP programs should have no regulations. The problem with not regulating commercial VOIP services is that most of the telephone companies are switching over to VOIP. There needs to be some sort of regulation to support a national telephone directory, emergency calling, and other services. On the other hand, free services should be able to do whatever they want. If someone wants to use an open source program to call a friend in another state, that should be their own business.
Lindbergh was NOT the first person to fly across the Atlantic. The first person to fly in an airplane was Albert Read when Lindbergh was only 17. In fact, several other pilots flew across the Atlantic before Lindbergh The Wright brothers were not the first to fly an airplane. They were just the first to be photographed while flying an airplane. Name almost any successful product that has been around for 10 years. That product was probably developed by someone else first. That's true for operating systems, search engines, cars.. Being first doesn't matter.
I finally downloaded firefox to see which I would prefer. I really like the security and some of the other features.
However, I won't be switching unless the following 2 problems are fixed.. 1- You can't read the text easily when scrolling. I'm not sure why internet explorer's scrolling works so much better. 2- The favorites panel. I really like to use the favorites panel on the left of the screen in internet explorer. The favorites panel provided by firefox is very clunky.
The problem with NASA is that it's primary mission is to increase it's funding. Why else are they talking about sending a man to Mars? It appeals to the taxpayers hearts more than a space probe ever will. These days, NASA only has a few programs that are expanding the field of science. Hubble and their space probes.
I would like to see NASA gutted except for non-manned space craft. Then I would like to see the 100 billion dollars or so left over given to private groups headed by people like Burt Rutan. Think what Burt Rutan could do with $10 billion in funding.
I think this will be a great tool for researchers at universities and companies. No longer will you have to contact the hardware group, beg and plead, and wait forever just to make a simple prototype. Instead, just send a simple an online order, and they will deliver the prototype to your doorstep.
Look at your own address zippy@cs.brandeis.edu. Isn't this a nonsense statement? "zee,eye,pee,pee,why,at,sea,s,dot,bee,our,a,en,de, e,eye,es,dot,e,de,you
I think it's impossible for voice recognition to be good enough. Let's say your name is bob Barnes. You choose an address bobb@yahoo. You say the word 'bobb'. This can either be bobby, bobbee, bobe, boby, baahbe, bahbe, etc... There isn't a one to one match between the word and the enunciation.
The problem is the form factor of the cell phone. Cell phones typically have a numeric keypad to keep the size reasonable. The easiest way to input a phone number is to enter numbers.
The biggest current trend is that everyone is switching to wireless phones. Most people don't want to carry around a phone large enough to contain a keyboard. Voice recognition works well only for words that are commonly used. For weird IP addresses, you would have to say each letter one at a time.
Imagine you meet somebody. You want to store his/her phone number your phone book on your cell phone. Which is easier? Typing 820-833-5214 or typing a 16 letter word into your 10 button keypad?
In order for Kurzweil's AI to work, it isnt enough to have 100Tflop. Kurzweil based the 100Tflop number on the idea that we would have algorithms available to replace the brain functions such as vision, pattern recognition, etc.. Just using a simple neural simulator would require many orders magnitude more of power to simulate a human brain. We are a long ways from having the needed algorithms (probably 30 years at a mininum).