98% is pretty pathetic - 1 error in 50. Most good Bayesian filters (SpamProbe, CRM114, DSPAM) can reach at least 99.9% (1 error in 1000) with ease. Others can grow far beyond this and reach as high as 99.985%, as a recent slashdot article covered (and this one). I reset my stats a few weeks ago, and out of 1800 spams so far, 0 have made it through. The only problem with Bayesian filtering is that it's mismarketed by companies who insist they have a better solution (although it's less accurate).
And to answer your question - collaborative filtering, such as message inoculation works quite well at boosting accuracy even beyond the high levels of accuracy Bayesian filters are really capable of, whereas things like shared groups and such hurt it.
Looks like pretty clean code, a nice change from when Microsoft's source code was stolen.
Isn't this a good thing, according to open source enthusiasts? Sure we'll lose a few thousand networks during the first couple of exploits, but we'll end up with much more stable code in the end. If these guys can hack Cisco's network, they don't need IOS's source code to hack everybody else's.
Makes you wonder about whether the cost of insurance will rise as a result of this. If you get in an accident and down a street light, they'll send your insurance company a bill for a new street light. If you get in an accident and take out 5 or 6 solar-powered weather computers, your insurance company will be paying out the nose for parts.
Hmm this would converge sex and oral sex together. No wonder the project got funded. This would prevent older women from having to buy vaginal dentures.
[Kirk] Fire photon torpedoes
[Scotty] I can't sir, the bloody computer's still encrypting a message to my girlfriend - I got no power!
[Romulans] b4w h4w h4w w3 0wnz j00!
[Kirk] W3 b3 0wn3d!
Why, so SCO could lay claim to Solaris too? I can't see Sun wanting to go anywhere near Linux until issues like SCO are resolved.
Re:Well, I guess that means the stuff really works
on
Spammer Sues SpamCop
·
· Score: 1
And what's worse is he's getting off right now at the idea of being on Slashdot. Why couldn't they link to optinrealbig.com in the slashdot article? We could've taken his network out for the day.
That might not be a bad idea actually, if we all configured our spam filters to forward all spams from Richter back to his personal and company email addresses asking to be removed.
Richter claims that because the complaints fail to identify the original email sender, Optin cannot comply with the CAN-SPAM ACT, which requires the sender of an email to remove the address of any person who does not desire to receive any further email.
In Soviet Russia, the spammer complains about forged headers.
Why don't you download the first couple parts of the course; i'm sure you'll find their comments in there somewhere. Do they also teach you how to flame before you think at MIT or did you pick that up on your own?
Funny that they should mention that, being that Abelson & Sussman from MIT's own CS courses (available via video lecture make the comments that programmers who have to go through a planning phase before they program aren't real programmers. It looks like MIT made most of the programmers who have no respect for the software development cycle.
Slashdot Mirror
98% is pretty pathetic - 1 error in 50. Most good Bayesian filters (SpamProbe, CRM114, DSPAM) can reach at least 99.9% (1 error in 1000) with ease. Others can grow far beyond this and reach as high as 99.985%, as a recent slashdot article covered (and this one). I reset my stats a few weeks ago, and out of 1800 spams so far, 0 have made it through. The only problem with Bayesian filtering is that it's mismarketed by companies who insist they have a better solution (although it's less accurate).
And to answer your question - collaborative filtering, such as message inoculation works quite well at boosting accuracy even beyond the high levels of accuracy Bayesian filters are really capable of, whereas things like shared groups and such hurt it.
Looks like pretty clean code, a nice change from when Microsoft's source code was stolen.
Isn't this a good thing, according to open source enthusiasts? Sure we'll lose a few thousand networks during the first couple of exploits, but we'll end up with much more stable code in the end. If these guys can hack Cisco's network, they don't need IOS's source code to hack everybody else's.
Makes you wonder about whether the cost of insurance will rise as a result of this. If you get in an accident and down a street light, they'll send your insurance company a bill for a new street light. If you get in an accident and take out 5 or 6 solar-powered weather computers, your insurance company will be paying out the nose for parts.
Vendor Trinkets.
Hmm this would converge sex and oral sex together. No wonder the project got funded. This would prevent older women from having to buy vaginal dentures.
This would do wonders for the pr0n industry... It's a bird, it's a plane...it's .... penis and ass man!
Anywhere there's a Denny's, Cracker Barrel, or a Howard Johnson's
I hope their web server's not using it. Nothing like a good slashdotting over a slow photonically encrypted connection.
Didn't they try and use this same technology to deliver oil over the Internet a few years ago? I don't recall that working very well.
[Kirk] Fire photon torpedoes
[Scotty] I can't sir, the bloody computer's still encrypting a message to my girlfriend - I got no power!
[Romulans] b4w h4w h4w w3 0wnz j00!
[Kirk] W3 b3 0wn3d!
The heavy metal pc takes compting to maybe not a new but different dimension.
Does it support spell check?
Isn't anybody worried they're gonna break the @!#$!@ moon and kill us all with massive tidal waves? Or am I just being paranoid?
Be the first one to use the robotic arm to throw your partner and hit the nebula and win immunity.
Why do I do this? Because the money's good, the scenery changes, and they let me use explosives, okay?
We're stayin we're goin' make up your mind... I vote we all stay and die.
This will be the next revolutionary product from Billy Mays.
Why, so SCO could lay claim to Solaris too? I can't see Sun wanting to go anywhere near Linux until issues like SCO are resolved.
And what's worse is he's getting off right now at the idea of being on Slashdot. Why couldn't they link to optinrealbig.com in the slashdot article? We could've taken his network out for the day.
That might not be a bad idea actually, if we all configured our spam filters to forward all spams from Richter back to his personal and company email addresses asking to be removed.
Richter is claiming to be "opt-in" only. Apparently the whole world has opted-in and Spamcop is sorely mistaken.
Richter claims that because the complaints fail to identify the original email sender, Optin cannot comply with the CAN-SPAM ACT, which requires the sender of an email to remove the address of any person who does not desire to receive any further email.
In Soviet Russia, the spammer complains about forged headers.
My l33t b00k has pictures. How can you read that book without any pictures?!?
Why don't you download the first couple parts of the course; i'm sure you'll find their comments in there somewhere. Do they also teach you how to flame before you think at MIT or did you pick that up on your own?
Pay no attention to the laptop behind the curtain.
Funny that they should mention that, being that Abelson & Sussman from MIT's own CS courses (available via video lecture make the comments that programmers who have to go through a planning phase before they program aren't real programmers. It looks like MIT made most of the programmers who have no respect for the software development cycle.