I had never been to slashdot before but I had heard about it from others, so I decided to
give it a try. It was last Tuesday on my lunch break and I had finished my stewed pickles
a little quicker than usual which left me with a couple possibilities. My first option
(and it was a pretty solid option had I gone with it) was to download the asset
depreciation schedule for facility 27 and cross check it against the running tally in the
GL projection report. But, I also had another option: try out this slashdot site I had
heard about. I was feeling a little adventurous (kind of like the time I showed everyone my
impression of a cat at the office christmas party), so I made the decision to try out this
website.
I typed "slash dot" into google and just 0.43 seconds later I had a list of 63,100,000
possible hits. And this is going to be my first complaint: How is a person supposed to know
which of these 63,100,000 pages is the right one to click? I might suggest some sort of
larger font, or something so it stands out. I think an even better suggestion would be to
have each of the letters in a different color and the colors are moving from one letter to
another. This is the kind of advanced trick I don't see on the web very much these days,
but I sure would have expected a technical website to be able to pull it off, makes me
wonder how good these guys really are.
After some sleuthing I decided to click on the first one and my suspicions were correct,
this was the place "news for nerds, stuff that matters." My expectations were still pretty
high even after that rocky start, and that's when I had my second letdown. I had clicked on
the first story and was reading through the comments when it hit me: there were no avatars.
No cute kittens pawing at you as you read the comment, no barrel chested lumberjacks
strolling towards you, no mini-indi 500 cars zipping around the comment. This seemed to be
a completely dead backwater of the internet. Wasn't this a technology website? Why
weren't they employing the latest and greatest techniques in their comment forum.
At this point I was too taken aback to continue reading the site. I was stunned to think
that a website could advertise itself as about technology, while ignoring some of the most
important techniques. I decided to leave the site and quickly clicked the back button,
returning to my favorite cat forum. As for this slash dot site, my experience rated one
star and I won't return.
There have been significant advancements in ANN methods, one specific area is deep belief networks and the training algorithm (Hinton). Prior to that, it was known that multi-layer networks could out perform simpler networks for things like image recognition, but there wasn't a good way to train them.
The newer models/methods are outperforming the previous simpler models/methods.
The reason the rest of the world went to chip and pin was because their fraud rate was high (many times that of the US). The US had one of the lowest fraud rates. After many years, chip and pin brought Europe's fraud rate down to be equal to US (maybe 2013 or something) and eventually a bit lower.
The sudden interest is due to advances in learning algorithms specifically in the area of deep belief networks. Some academics (Hinton and others) came up with some methods to be able to train these types of networks, this in turn has allowed everyone to make use of these networks in areas where they perform extremely well (e.g. image recognition).
The capabilities of these new tools are definitely not hype, they are very effective, but whether you call them "AI" or not is a different discussion.
It's interesting to look at the details of the job shift due to e-commerce and then automation.
The shift to e-commerce causes a few changes related labor:
1 - Shifts the labor related to the final sale transaction from the retail store to the packing operation - this becomes somewhat of a wash for labor
2 - Removes the DC distribution labor supporting the stores - which is non-trivial but operated typically at the pallet and case level
3 - Added the piece pick labor to the DC (previously performed by the consumer in the store) which is an order of magnitude more labor intensive than previous operations for two reasons:
3.1 - Because pallet/case is at an aggregate level
3.2 - Because the DC is much larger than the store due to higher volumes and more skus causing the walk distance to increase
Without picking automation, the shift from brick and mortar to e-commerce causes a labor increase (ignoring details like idle time by store workers)
With picking automation, the shift is closer to a net wash for labor
When packing is automated (e.g. auto-baggers) then the labor drops below brick and mortar.
but machine learning/pattern matching is step one, a foundational capability that can be used as one of multiple building blocks to solve higher level problems.
What's the difference between someone who downloads a BitTorrent and someone who would never pay for the show if BitTorrent (or other options) weren't available? None.
The difference is in the state of the observer. Without the download there is unmet demand for entertainment for that person. With the download there is some portion of that demand for entertainment that is met.
It's not stealing, dumbfuck.
The dictionary definition does not always match the cultural usage of a term. When property or a service is "taken" without paying, people frequently call it stealing. The concept of physical property ownership is as arbitrary as the concept of intellectual property (meaning there is no natural law that governs property or ownership, they are made up concepts by man), so these types of arguments don't really add any value to the discussion.
Hosted solutions started back in the 60's, typically mainframe. The earliest I encountered it was mid-80's, one of our clients used a hosted financials application.
and tend to agree with your post based on my "non-expert but considerable time spent" experience.
My story:
I wanted to create intelligence via artificial life and evolution. I didn't want to create human intelligence, just tiny little creatures trying to survive type intelligence. I provided them basic sensor inputs and motor/movement raw materials but didn't program in any usage of those things, they need to figure out through evolution (how to see, move, find food, avoid getting eaten, etc.). They started with random neural nets and through generations increased in capabilities up to a plateau.
Some conclusions:
1 - There is a lot of foundational stuff I had to learn slowly and piece meal by googling etc., for example, what is a neural net doing mathematically, why/when would you use one.
2 - I thought I might figure out something interesting or clever - but the reality is that people with math backgrounds (e.g. PhD ) are the ones that are going to figure out that next clever insight. For example, someone once asked me why I was using neural nets and not support vector machines, so I read up on support vector machines and they seemed to be doing the same thing (function approximation). But I didn't have enough training to fully grasp why a support vector machine and neural are different, what are pros and cons. Reading papers online with teacher is a slow and cumbersome way to acquire knowledge in a complex area.
3 - Interesting issue not really related to the topic at hand but fun to talk about: My other conclusion was that guiding an evolving system towards intelligence is a very tricky task. My creatures hit a plateau of behavior that was at least interesting (chasing/tracking other creatures to get food, avoiding getting eaten, avoiding obstacles) , but difficult to get beyond. How would I need to change the conditions in the environment to push them beyond that into more advanced behaviors, for example hiding around a corner or hunting in packs, etc?. The initial environment needs to be favorable for guiding random brains towards some basic functionality, but then it needs to change and continue changing to keep pushing the evolution process towards more complex capabilities. That is a tricky problem, knowing which environmental conditions would reward intelligence over speed or strength or other attributes.
Well, yep, math is how we solve pretty much everything. What's more interesting than just saying it's a bunch of matrix math is understanding at a more abstract level what the math functions represent.
The reason neural nets end up being interesting is because they are essentially universal function approximators that can be adjusted/tweaked to move closer and closer to a desired function (based on input/output data) without actually knowing the function in advance.
Uhhhh, that was your response to how to sort bank records, that you "could" implement x86 on a neural network, which leads to a couple questions:
1 - If you only said "could" knowing that it's not really a good solution, why post it all?
2 - Given that you seem to agree that is not a good solution (e.g. not efficient), what is your answer to the poster that asked you about how you would sort bank records? (the implied full question of course is "how would you do it efficiently with NN compared to current computing methods?")
Probably the issue you would run into is cost of the specific machine will be greater than the cost of the general machine. For some users it would be worth it, but the market tends towards low cost not good solution, so the company selling it would probably be fighting an uphill battle unless the difference in capability was very significant.
This is what happened with mini-computers. They typically were designed for business processing workloads and had features (e.g. special disk capabilities for high IO, extra processors managing all non-cpu activity, etc.) not present in lower cost devices (e.g. PC's), but the mass market moves fast, costs lower quickly, the lower end of the existing market gets eaten up first so the niche becomes smaller, unit costs in the niche go up, etc.
You think we should add an additional layer of abstraction and computation that slows things down and eats up more energy to arrive back at the exact same spot we already were (running x86 code)?
Slashdot Mirror
I'd like to see Einstein explain bitcoin to his grandma.
I had never been to slashdot before but I had heard about it from others, so I decided to give it a try. It was last Tuesday on my lunch break and I had finished my stewed pickles a little quicker than usual which left me with a couple possibilities. My first option (and it was a pretty solid option had I gone with it) was to download the asset depreciation schedule for facility 27 and cross check it against the running tally in the GL projection report. But, I also had another option: try out this slashdot site I had heard about. I was feeling a little adventurous (kind of like the time I showed everyone my impression of a cat at the office christmas party), so I made the decision to try out this website.
I typed "slash dot" into google and just 0.43 seconds later I had a list of 63,100,000 possible hits. And this is going to be my first complaint: How is a person supposed to know which of these 63,100,000 pages is the right one to click? I might suggest some sort of larger font, or something so it stands out. I think an even better suggestion would be to have each of the letters in a different color and the colors are moving from one letter to another. This is the kind of advanced trick I don't see on the web very much these days, but I sure would have expected a technical website to be able to pull it off, makes me wonder how good these guys really are.
After some sleuthing I decided to click on the first one and my suspicions were correct, this was the place "news for nerds, stuff that matters." My expectations were still pretty high even after that rocky start, and that's when I had my second letdown. I had clicked on the first story and was reading through the comments when it hit me: there were no avatars. No cute kittens pawing at you as you read the comment, no barrel chested lumberjacks strolling towards you, no mini-indi 500 cars zipping around the comment. This seemed to be a completely dead backwater of the internet. Wasn't this a technology website? Why weren't they employing the latest and greatest techniques in their comment forum.
At this point I was too taken aback to continue reading the site. I was stunned to think that a website could advertise itself as about technology, while ignoring some of the most important techniques. I decided to leave the site and quickly clicked the back button, returning to my favorite cat forum. As for this slash dot site, my experience rated one star and I won't return.
There have been significant advancements in ANN methods, one specific area is deep belief networks and the training algorithm (Hinton). Prior to that, it was known that multi-layer networks could out perform simpler networks for things like image recognition, but there wasn't a good way to train them.
The newer models/methods are outperforming the previous simpler models/methods.
The reason the rest of the world went to chip and pin was because their fraud rate was high (many times that of the US). The US had one of the lowest fraud rates. After many years, chip and pin brought Europe's fraud rate down to be equal to US (maybe 2013 or something) and eventually a bit lower.
The sudden interest is due to advances in learning algorithms specifically in the area of deep belief networks. Some academics (Hinton and others) came up with some methods to be able to train these types of networks, this in turn has allowed everyone to make use of these networks in areas where they perform extremely well (e.g. image recognition).
The capabilities of these new tools are definitely not hype, they are very effective, but whether you call them "AI" or not is a different discussion.
Ya, he's always making crazy claims, like he's going to build a reusable rocket, and he's going to land it on a barge. What a nut!
They don't need to know that, they buy a machine that is ready to use, just plug it in and turn it on.
If only we had AI to help us understand that damn worm!
It's interesting to look at the details of the job shift due to e-commerce and then automation.
The shift to e-commerce causes a few changes related labor:
1 - Shifts the labor related to the final sale transaction from the retail store to the packing operation - this becomes somewhat of a wash for labor
2 - Removes the DC distribution labor supporting the stores - which is non-trivial but operated typically at the pallet and case level
3 - Added the piece pick labor to the DC (previously performed by the consumer in the store) which is an order of magnitude more labor intensive than previous operations for two reasons:
3.1 - Because pallet/case is at an aggregate level
3.2 - Because the DC is much larger than the store due to higher volumes and more skus causing the walk distance to increase
Without picking automation, the shift from brick and mortar to e-commerce causes a labor increase (ignoring details like idle time by store workers)
With picking automation, the shift is closer to a net wash for labor
When packing is automated (e.g. auto-baggers) then the labor drops below brick and mortar.
"...if only we could get rid of the ugly and distracting web page content"
"Coffee Lake" barely beat out "Man Bear Pig"
Red paint reflects hottest part of suns rays, increases global warming.
Top speed on straightest sections of track will be 760mph, more curvy areas will be around 300mph to reduce lateral g's.
but machine learning/pattern matching is step one, a foundational capability that can be used as one of multiple building blocks to solve higher level problems.
What's the difference between someone who downloads a BitTorrent and someone who would never pay for the show if BitTorrent (or other options) weren't available? None.
The difference is in the state of the observer. Without the download there is unmet demand for entertainment for that person. With the download there is some portion of that demand for entertainment that is met.
It's not stealing, dumbfuck.
The dictionary definition does not always match the cultural usage of a term. When property or a service is "taken" without paying, people frequently call it stealing. The concept of physical property ownership is as arbitrary as the concept of intellectual property (meaning there is no natural law that governs property or ownership, they are made up concepts by man), so these types of arguments don't really add any value to the discussion.
In other words, the title should be "$1.48 Billion has been transferred to smarter people."
If I remember right, these are the key steps:
1. Start Up
2. Cash In
3. Sell Out
4. Bro Down
Amazon and Whole Foods engaged in a normal negotiation process..
Hosted solutions started back in the 60's, typically mainframe. The earliest I encountered it was mid-80's, one of our clients used a hosted financials application.
Can we develop AI to prevent duplicate slashdot stories?
and tend to agree with your post based on my "non-expert but considerable time spent" experience.
My story:
I wanted to create intelligence via artificial life and evolution. I didn't want to create human intelligence, just tiny little creatures trying to survive type intelligence. I provided them basic sensor inputs and motor/movement raw materials but didn't program in any usage of those things, they need to figure out through evolution (how to see, move, find food, avoid getting eaten, etc.). They started with random neural nets and through generations increased in capabilities up to a plateau.
Some conclusions:
1 - There is a lot of foundational stuff I had to learn slowly and piece meal by googling etc., for example, what is a neural net doing mathematically, why/when would you use one.
2 - I thought I might figure out something interesting or clever - but the reality is that people with math backgrounds (e.g. PhD ) are the ones that are going to figure out that next clever insight. For example, someone once asked me why I was using neural nets and not support vector machines, so I read up on support vector machines and they seemed to be doing the same thing (function approximation). But I didn't have enough training to fully grasp why a support vector machine and neural are different, what are pros and cons. Reading papers online with teacher is a slow and cumbersome way to acquire knowledge in a complex area.
3 - Interesting issue not really related to the topic at hand but fun to talk about: My other conclusion was that guiding an evolving system towards intelligence is a very tricky task. My creatures hit a plateau of behavior that was at least interesting (chasing/tracking other creatures to get food, avoiding getting eaten, avoiding obstacles) , but difficult to get beyond. How would I need to change the conditions in the environment to push them beyond that into more advanced behaviors, for example hiding around a corner or hunting in packs, etc?. The initial environment needs to be favorable for guiding random brains towards some basic functionality, but then it needs to change and continue changing to keep pushing the evolution process towards more complex capabilities. That is a tricky problem, knowing which environmental conditions would reward intelligence over speed or strength or other attributes.
So multiplication of 16 bit matrices is AI?
Well, yep, math is how we solve pretty much everything. What's more interesting than just saying it's a bunch of matrix math is understanding at a more abstract level what the math functions represent.
The reason neural nets end up being interesting is because they are essentially universal function approximators that can be adjusted/tweaked to move closer and closer to a desired function (based on input/output data) without actually knowing the function in advance.
Uhhhh, that was your response to how to sort bank records, that you "could" implement x86 on a neural network, which leads to a couple questions:
1 - If you only said "could" knowing that it's not really a good solution, why post it all?
2 - Given that you seem to agree that is not a good solution (e.g. not efficient), what is your answer to the poster that asked you about how you would sort bank records? (the implied full question of course is "how would you do it efficiently with NN compared to current computing methods?")
Probably the issue you would run into is cost of the specific machine will be greater than the cost of the general machine. For some users it would be worth it, but the market tends towards low cost not good solution, so the company selling it would probably be fighting an uphill battle unless the difference in capability was very significant.
This is what happened with mini-computers. They typically were designed for business processing workloads and had features (e.g. special disk capabilities for high IO, extra processors managing all non-cpu activity, etc.) not present in lower cost devices (e.g. PC's), but the mass market moves fast, costs lower quickly, the lower end of the existing market gets eaten up first so the niche becomes smaller, unit costs in the niche go up, etc.
You think we should add an additional layer of abstraction and computation that slows things down and eats up more energy to arrive back at the exact same spot we already were (running x86 code)?
Why would you want to do that?