I think they are interesting in the same way that the first iron smelters are interesting. Sure, what came out of them was really crude - but they did boot-strap a significant advancement in technology. All subsequent ages were more advanced, because refined metal was now a possibility.
I guess you and I are going to have to agree to disagree.
I think I can see where you are coming from. Is this accurate? Your opinion is that if the goal-seeking behavior is supplied by an external mechanism, then the object is not intelligent, but rather just executing instinctual patterns.
In other words, innate behavior is not intelligent, only learned behavior is intelligent.
From my point of view, if it looks like a duck, and acts like a duck, and swims like a duck, and can learn the things a duck learns, it's probably a duck - or close enough that the differences don't matter.
I do expect that we will soon have software that will auto-optimize itself, adding new goal-seeking behavior of it's own discovery. But I admit - I don't know of any real-world examples today.
I'm going to argue that small examples of machine intelligence do exist. That is to say, there exist machines that are intelligent, however in a small space.
For example, my Subaru has a fuzzy-logic chip built into the automatic transmission. The first couple thousand miles I drove it, it was learning my driving pattern. Once it learned my driving pattern, the shifts from one gear to another were like glass.
Wall Street has A.I's that have learned to do arbitrage - see Algorithmic Trading Platforms. Read the third and fourth paragraphs to see how extensively learning A.I.'s have been deployed. This is why "day trading" is a fools market.
We have a software package at work that watches packet traffic on our network switches, and can shut down a network port if the machine attached to that port starts spewing packets at an alarming rate, or to a different set of TCP/IP ports than it "normally" does.
Each of these are examples of machine learning. Isn't the ability to learn the foundation of intelligence?
Once the foundation is conquered, the rest of it is just a matter of scale.
It's easy to discount a fuzzy-logic chip as non-intelligent, because we understand it. But I'm going to argue that just because we understand how the machine learned the patterns, doesn't mean that learning the pattern is as non-intelligent as the machines of the nineteenth century (or before). The learning is real. The goal seeking is real. The intelligence is real - it just hasn't been generalized yet.
The interesting problems: recognizing people's identity by sight/sound/patterns, comprehending language, predicting emotion - these are going to take a tremendous amount of CPU cycles. But that's just a matter of scale - not of magic. We learned it. So can the machine.
Twenty years ago, I was working on a PC that did 8 MIPS. Today, the PC I'm typing this on does 40,000 MIPS. Do I think that by 2029, the average PC will be able to learn to predict my emotion?
Don't get me wrong - if my MTA vendor builds in DKIM support, I'll use it. But I won't use it as the primary because it looks CPU heavy. A few months ago, my MTA was getting 9 connects per second per 24 hour period.
I'd be nuts to attempt do public key cryptography at that rate.
So really, I'll use RBLs and SPF and SURBLs to get rid of the obvious crap, and then validate what does get through the lightweight barriers. I don't really care what method is used on the back end - DKIM is fine by me. But DKIM won't be the primary, essentially because it would make a fine denial of service attack vector by a botnet.
In my case, I'm going to use use SPF and SURBL in pretty much the same way: as modifiers to the spam score. So I've already accepted the mail into my anti-spam system - it's just a matter of if the messages gets out of quarantine now. But you are correct - the SPF lookup is lighter weight than the SURBL, so I ought to do it first.
Yes - even though my MTA doesn't hang up on connections that fail the SPF check, at least by publishing my hard fail SPF record, I'm helping other mail systems to hang up on spam that tried to claim it was from us.
And a soon as any of my vendors support it, I'll start adding the SPF score to my ham/spam weighting.
I only get the benefit of one half of SPF - if you receive a connect from a zombie PC purporting to be one of my users, you can instantly hang up on the box, knowing it wasn't from us. But even that helped me a lot, as the number of bounce-backs I got dropped tremendously.
Heck, my SPF record is of the "hard fail" type - so if you do get a bogus connect for my domain, feel free to update the RBL of your choice.;-)
SPF isn't supposed to be the anti-dictionary-attack solution. It's a lightweight way of checking whether the incoming connection is coming from the IP address range that the message header claims to represent. That's all.
It does help me, even though my MTA doesn't support it. After I put in the "hard fail" SPF record for our domain, we got far fewer bounce-backs from other people getting bogus messages passing themselves off as us. Less bounce-backs means less work for me, explaining to my users why some mail server somewhere told them they had mis-addressed a message they never sent.
Agreed. If I'm going to throw junk away, first I'll use RBLs, then SURBLs, and then I would like to cull the next layer of spam out with SPF. I only want to try the CPU heavy stuff on the items that make it past all the lightweight barriers.
From TFA:
PayPal is deploying DKIM after already rolling out Sender Policy Framework (SPF), a complementary Microsoft-backed standard that is an extension to the Simple Mail Transfer Protocol (SMTP). SPF allows software to reject e-mail coming out of forged "from" addresses. Except that Microsoft shat on SPF because it was Not Invented Here. They tried to get the world to implement their Sender ID protocol instead.
The IETF refused to ratify SPF as an official standard because it didn't have Microsoft support.
Today, RFC 4408 is still an "experimental" protocol - due to Microsoft's hurt. Someone at Network World isn't familiar with the material they are reporting.
I think SPF addresses a real problem, and does it well; but, my MTA vendor doesn't want to spend the programmer cycles on something non-standard (they've been accused of being non-standard in the past, and don't want to risk the accusation again). I am annoyed that something so simple and easy as SPF isn't ubiquitous yet.
To extend that a little further, something that is very effective is to set up a spamtrap email address. If you have a web site, you put a hidden email address on it that is solely used as a source of spam. If people cannot read it, but web-crawlers can, then you have your bad guys.
Using that Perl script to dredge the spamtrap mailbox via IMAP makes an easy source to train against.
Something is different for you then. I receive email direct from PCs on the Comcast network all the time. Comcast does publish their dynamic IP address range to (at least one of) the Spamhaus RBL lists. So my anti-spam software throws all those Comcast emails into quarantine. If the home user configures their mail program to use the Comcast server as the SMTP relay server, then they are fine, as the Comcast mail servers are not in the RBL.
This is pretty naive. You say it would be OK, if there was oversight - but what makes you think there would be oversight?
My dad's snail mail was being read while he was active in politics. We complained to the postmaster, who did nothing, because he was of the opposite political party. (Heck, it was probably his idea). The planning meeting for California had to be moved to a different location, because the United States Post Office was violating the privacy of snail mail for political gain.
The impetus for the TFA is about a guy who told George Bush that the only way to ensure security is if people give up their privacy. Essentially, everyone is re-defined to be a terrorist.
Well, YOU may want a police state, so that nothing bad ever happens. But I know that a police state IS the bad thing that will happen.
Not knowing much at all about this whole field of software, I'm going to ask: what other software can build the model simon can use? I think that would be pretty cool, if I could use simon with something that is released under the GPL or BSD style licenses. Thanks!
You are 100% correct that the important thing is to have a team that has good cohesion. And I can see where your point about proximity providing cross-training is beneficial to everyone.
TFA comes to the conclusion that the telecommuters ought to be brought back into the office once in a while, so that the left-behinders don't feel so left-behind.
But for me, I'm certain that I could get more work done if I didn't have the distractions caused by our extrovert being visited by other extroverts. I'd be happy to throw that extra time into updating our documentation wiki, or doing a webex to explain the latest changes to our production systems.
Ultimately, I find our close-proximity to be a drain, and the collaboration benefit to be minimal. The cohesion isn't there; so I'd rather throw myself into being the super-sysadmin. I think that my extrovert co-worker probably doesn't get much from me as far as technical skill. I'm sure I make a much better (social) sounding board than tech-info database.
So dragging me back into the office, just so my extrovert co-worker has someone to regale with stories would slowly make me angry, I think.
I do sysadmin work in a space with three other guys; three of us are introverts (mostly), and one of us is an extrovert (largely). It's not easy.
Seems to me, the article talks about the effect telecommuting had on the extrovert. Well, sure. I can absolutely see where the lack of an audience is going to be a total bummer for the extrovert.
But us introverts say a prayer of thanks when the telecommute offer comes in.
The study is probably a little bit skewed, in that extroverts want to come to work, so that they do get their audience. When offered the telecommute, the extroverts probably turned it down.... Yes, they were left behind. And sure, they may be more lonely now. But given my 'druthers, I'd rather the extroverts work in Sales.
Like other people say, you need a foyer, an office for you, and a separate room for the servers and networking gear.
So the data center needs a UPS. Many servers come with two power supplies, so if you are worried about losing your one UPS, then you get two - one feeds the "A" side, the other, the "B" side power supplies. If you go high-end, you can even get remote power control for each plug. Or maybe you need a KVM over IP. Whichever, you'll be better off if you can remotely access the boxes (and power them down if need be), and not have to go inside the data center if you don't have to. Remote access from your desk also means remote access via VPN. Now there's no reason not to send you to training, right? Have laptop, will travel.
Inside the data center, make sure that ceiling mounted lighting is offset from the racks - you want the light to shine inside the cabinets from over your shoulder, not directly down on top of the cabinet, keeping the back of the machines in shadow. If you go with a 12" raised floor, you might consider running power along the floor, and data cables overhead in ladders. It sucks to be unraveling a data cable, and accidentally dislodge a power plug.
For jack fields - having them in a rack standing in the middle (not against a wall) seems to be the best way to go. The telecom guys wire up the whole rack from behind, and that feeds the building. Then the front is what you patch into the switches.
If you want to use a USB web-cam for video surveillance, plan where it will be mounted, and where the PC that runs it will be - you'd like the camera to have a straight look at the door into the data center.
And get your bosses to sign off on electronic locks to with keycards to open the doors - especially for the warehouse. Oh wait, did I forget to mention that you should plan on a warehouse?
;-)
OK really - it's just a work room, that's not the foyer, and not your office, and not the data center - but it does have shelves and a work table - and a locked door. Even if it is just where you stuff the boxes the shipper brings in late on a Friday afternoon....
As for an office for myself, all I really want is four (or more) screens and one keyboard. At the moment, I have three screens, and a KVM that switches one of the screens between four CPUs. It keeps me busy, but I dislike have to switch the KVM so often. One beefy PC with two dual port video cards, and VMWare would probably be enough. Although three dual port video cards (and six monitors) would be perfect.;-)
As an email admin, I'd like to suggest you ask your work email administrator for a change. Tell him/her this plan, and they might like it:
1) Add a nickname to your mailbox so you can receive email at both addresses.
2) Tell everyone you want about your new email address.
3) Watch for email coming in to the old address for a few months, and tell those people about your new address.
4) When you are happy that you warned everyone, give your old address to the email admin for his spamtrap account.
For the most part, I agree. However.... (There's always a "however", isn't there?)
So we were planning on a deployment of internet kiosks for job seekers, and PacBell said they could deliver the whole thing: ISDN endpoint routers, datacomm lines, and the communications server. They spec'd out about 20 Cisco 700 routers.
So I'm walking through Interop, and drop into the Cisco booth. The Cisco salesman told me to avoid the Cisco 700s, and get the Cisco 800s instead. For this small user market, Cisco acquired Combinet. The hardware was the same, but the 700s came with the Combinet OS; the 800s had IOS ported to them. Turned out PacBell was trying to help a supplier get rid of old inventory. It would have truly sucked if we had gone with the PacBell recommendation.
So yes, unless you are taking one of the training classes there, it was of marginal value. But if you had a specific plan you wanted vendors to pick apart, it was great to get the sales guys to explain how their stuff could be hooked up and give you more than what the other guy proposed.
So one of our desktop techs had just such a user, and thought that the answer was a nice, non-intimidating entry level computer book. Unfortunately, the words that left his mouth were "You need to get a book: PCs for Dummies".
Dear Mr. Haselton, have you considered (only) suing those people who are currently blacklisted by an RBL or SURBL? I would think this would be particularly effective if you used the same RBL lists that metrokc.gov uses.
"For my first witness, I'd like to call Judge Judith Eiler's IT department email administrator." "Sir (/Madam), do you protect the Judge's email account from spam? Is this spammer's IP address one you would protect the Judge from?"
Slashdot Mirror
We'll have to take up the discussion again when I can simulate that. ;-)
I think they are interesting in the same way that the first iron smelters are interesting. Sure, what came out of them was really crude - but they did boot-strap a significant advancement in technology. All subsequent ages were more advanced, because refined metal was now a possibility.
I think I can see where you are coming from. Is this accurate? Your opinion is that if the goal-seeking behavior is supplied by an external mechanism, then the object is not intelligent, but rather just executing instinctual patterns.
In other words, innate behavior is not intelligent, only learned behavior is intelligent.
From my point of view, if it looks like a duck, and acts like a duck, and swims like a duck, and can learn the things a duck learns, it's probably a duck - or close enough that the differences don't matter.
I do expect that we will soon have software that will auto-optimize itself, adding new goal-seeking behavior of it's own discovery. But I admit - I don't know of any real-world examples today.
For example, my Subaru has a fuzzy-logic chip built into the automatic transmission. The first couple thousand miles I drove it, it was learning my driving pattern. Once it learned my driving pattern, the shifts from one gear to another were like glass.
Wall Street has A.I's that have learned to do arbitrage - see Algorithmic Trading Platforms. Read the third and fourth paragraphs to see how extensively learning A.I.'s have been deployed. This is why "day trading" is a fools market.
We have a software package at work that watches packet traffic on our network switches, and can shut down a network port if the machine attached to that port starts spewing packets at an alarming rate, or to a different set of TCP/IP ports than it "normally" does.
Each of these are examples of machine learning. Isn't the ability to learn the foundation of intelligence?
Once the foundation is conquered, the rest of it is just a matter of scale.
It's easy to discount a fuzzy-logic chip as non-intelligent, because we understand it. But I'm going to argue that just because we understand how the machine learned the patterns, doesn't mean that learning the pattern is as non-intelligent as the machines of the nineteenth century (or before). The learning is real. The goal seeking is real. The intelligence is real - it just hasn't been generalized yet.
The interesting problems: recognizing people's identity by sight/sound/patterns, comprehending language, predicting emotion - these are going to take a tremendous amount of CPU cycles. But that's just a matter of scale - not of magic. We learned it. So can the machine.
Twenty years ago, I was working on a PC that did 8 MIPS. Today, the PC I'm typing this on does 40,000 MIPS. Do I think that by 2029, the average PC will be able to learn to predict my emotion?
Yes I do.
I'd be nuts to attempt do public key cryptography at that rate.
So really, I'll use RBLs and SPF and SURBLs to get rid of the obvious crap, and then validate what does get through the lightweight barriers. I don't really care what method is used on the back end - DKIM is fine by me. But DKIM won't be the primary, essentially because it would make a fine denial of service attack vector by a botnet.
And a soon as any of my vendors support it, I'll start adding the SPF score to my ham/spam weighting.
DKIM looks to be a little on the CPU heavy side, but if a message has to clear all the other hurdles first, then that shouldn't be too bad.
It was a pretty disgraceful set of moves.
I expect that from Microsoft, but then, I'm an old guy and have plenty of experience on which to draw....
Heck, my SPF record is of the "hard fail" type - so if you do get a bogus connect for my domain, feel free to update the RBL of your choice. ;-)
It does help me, even though my MTA doesn't support it. After I put in the "hard fail" SPF record for our domain, we got far fewer bounce-backs from other people getting bogus messages passing themselves off as us. Less bounce-backs means less work for me, explaining to my users why some mail server somewhere told them they had mis-addressed a message they never sent.
Agreed. If I'm going to throw junk away, first I'll use RBLs, then SURBLs, and then I would like to cull the next layer of spam out with SPF. I only want to try the CPU heavy stuff on the items that make it past all the lightweight barriers.
The IETF refused to ratify SPF as an official standard because it didn't have Microsoft support.
Today, RFC 4408 is still an "experimental" protocol - due to Microsoft's hurt. Someone at Network World isn't familiar with the material they are reporting.
I think SPF addresses a real problem, and does it well; but, my MTA vendor doesn't want to spend the programmer cycles on something non-standard (they've been accused of being non-standard in the past, and don't want to risk the accusation again). I am annoyed that something so simple and easy as SPF isn't ubiquitous yet.
Using that Perl script to dredge the spamtrap mailbox via IMAP makes an easy source to train against.
My dad's snail mail was being read while he was active in politics. We complained to the postmaster, who did nothing, because he was of the opposite political party. (Heck, it was probably his idea). The planning meeting for California had to be moved to a different location, because the United States Post Office was violating the privacy of snail mail for political gain.
The impetus for the TFA is about a guy who told George Bush that the only way to ensure security is if people give up their privacy. Essentially, everyone is re-defined to be a terrorist.
Well, YOU may want a police state, so that nothing bad ever happens. But I know that a police state IS the bad thing that will happen.
I know you may not feel that 37 minutes is worth understanding what is going on, but if you want some background information: Judge Andrew Napolitano delivers a speech against the excesses of the new security state.
TFA comes to the conclusion that the telecommuters ought to be brought back into the office once in a while, so that the left-behinders don't feel so left-behind.
But for me, I'm certain that I could get more work done if I didn't have the distractions caused by our extrovert being visited by other extroverts. I'd be happy to throw that extra time into updating our documentation wiki, or doing a webex to explain the latest changes to our production systems.
Ultimately, I find our close-proximity to be a drain, and the collaboration benefit to be minimal. The cohesion isn't there; so I'd rather throw myself into being the super-sysadmin. I think that my extrovert co-worker probably doesn't get much from me as far as technical skill. I'm sure I make a much better (social) sounding board than tech-info database.
So dragging me back into the office, just so my extrovert co-worker has someone to regale with stories would slowly make me angry, I think.
Seems to me, the article talks about the effect telecommuting had on the extrovert. Well, sure. I can absolutely see where the lack of an audience is going to be a total bummer for the extrovert.
But us introverts say a prayer of thanks when the telecommute offer comes in.
The study is probably a little bit skewed, in that extroverts want to come to work, so that they do get their audience. When offered the telecommute, the extroverts probably turned it down.... Yes, they were left behind. And sure, they may be more lonely now. But given my 'druthers, I'd rather the extroverts work in Sales.
So the data center needs a UPS. Many servers come with two power supplies, so if you are worried about losing your one UPS, then you get two - one feeds the "A" side, the other, the "B" side power supplies. If you go high-end, you can even get remote power control for each plug. Or maybe you need a KVM over IP. Whichever, you'll be better off if you can remotely access the boxes (and power them down if need be), and not have to go inside the data center if you don't have to. Remote access from your desk also means remote access via VPN. Now there's no reason not to send you to training, right? Have laptop, will travel.
Inside the data center, make sure that ceiling mounted lighting is offset from the racks - you want the light to shine inside the cabinets from over your shoulder, not directly down on top of the cabinet, keeping the back of the machines in shadow. If you go with a 12" raised floor, you might consider running power along the floor, and data cables overhead in ladders. It sucks to be unraveling a data cable, and accidentally dislodge a power plug.
For jack fields - having them in a rack standing in the middle (not against a wall) seems to be the best way to go. The telecom guys wire up the whole rack from behind, and that feeds the building. Then the front is what you patch into the switches. If you want to use a USB web-cam for video surveillance, plan where it will be mounted, and where the PC that runs it will be - you'd like the camera to have a straight look at the door into the data center.
And get your bosses to sign off on electronic locks to with keycards to open the doors - especially for the warehouse. Oh wait, did I forget to mention that you should plan on a warehouse?
OK really - it's just a work room, that's not the foyer, and not your office, and not the data center - but it does have shelves and a work table - and a locked door. Even if it is just where you stuff the boxes the shipper brings in late on a Friday afternoon....
As for an office for myself, all I really want is four (or more) screens and one keyboard. At the moment, I have three screens, and a KVM that switches one of the screens between four CPUs. It keeps me busy, but I dislike have to switch the KVM so often. One beefy PC with two dual port video cards, and VMWare would probably be enough. Although three dual port video cards (and six monitors) would be perfect. ;-)
1) Add a nickname to your mailbox so you can receive email at both addresses.
2) Tell everyone you want about your new email address.
3) Watch for email coming in to the old address for a few months, and tell those people about your new address.
4) When you are happy that you warned everyone, give your old address to the email admin for his spamtrap account.
It's a win-win for you both. ;-)
So we were planning on a deployment of internet kiosks for job seekers, and PacBell said they could deliver the whole thing: ISDN endpoint routers, datacomm lines, and the communications server. They spec'd out about 20 Cisco 700 routers.
So I'm walking through Interop, and drop into the Cisco booth. The Cisco salesman told me to avoid the Cisco 700s, and get the Cisco 800s instead. For this small user market, Cisco acquired Combinet. The hardware was the same, but the 700s came with the Combinet OS; the 800s had IOS ported to them. Turned out PacBell was trying to help a supplier get rid of old inventory. It would have truly sucked if we had gone with the PacBell recommendation.
So yes, unless you are taking one of the training classes there, it was of marginal value. But if you had a specific plan you wanted vendors to pick apart, it was great to get the sales guys to explain how their stuff could be hooked up and give you more than what the other guy proposed.
That went over well. Not.
"For my first witness, I'd like to call Judge Judith Eiler's IT department email administrator." "Sir (/Madam), do you protect the Judge's email account from spam? Is this spammer's IP address one you would protect the Judge from?"
In America, researchers apply a jolt of electricity to their wastewater bacteria. In Soviet Russia, the brew tases you! Don't Tase me, brew!