Just had a telemarketing call. They had the right number, but the wrong name and address. Considering we've been here for a couple of years, they must be working from really old call lists
( Read more... )
And the best bit would be that the filterboxes would all be tied into a free-call number hooked to a central filter system, so all filterboxes would act pretty much as one giant filter for the country, updating and synchronising their filter lists as new telemarketing numbers got through and people pressed the big red "Filter!" button on their phones. More than a certain number of tagged calls from the same number would spill it over into the public filter list until no more tagged calls from that number had been registered for a certain amount of time.
I also forsee a market for a US cellphone version of the filter, as apparently over there they pay to receive calls, not make them. Thus telemarketing costs the call receiver both time and money.
Not to mention that with the Eliza idea, each filterchip would pull down a slight variation of the human-spoofing software and test it against telemarketing calls, then report the settings and success (amount of time wasted) back to the central system, which would then use a genetic algorithm to search for the most effective time-wasting set of parameters and upload those to each filterbox, honing its timewasting abilities with each telemarketing call.
As a particularly delicious side-effect, there would be ways to spoof the system and partially whitelist telemarketing numbers. However, the only way to do this would be to buy a lot of the filterboxes (each one with a unique ID) and set up a rather complex and expensive PABX-alike or phone-spoofing system.
However, the more unique IDs (customers) flag a number as a telemarketing number, the more filterboxes will have to be bought and set up on spoofing systems to counterbalance that - and whitelisting would require resources proportional to the square of the amount of blacklisting. This means that minimally-telemarketing numbers could be whitelisted by buying a couple of filterboxes and support systems, medium-telemarketing numbers would require a boatload of filterboxes to counter, and numbers used for massive amounts of telemarketing wouldn't be able to whitelist themselves by any process other than not using that number at all for six to 24 months.
This means that anyone wanting to whitelist their numbers will have to try buying lots of filterboxes. And while buying one might be cheap, buying dozens or hundreds certainly won't be. Another way to funnel money from telemarketers to the filterbox company.
And the best bit would be that the filterboxes would all be tied into a free-call number hooked to a central filter system, so all filterboxes would act pretty much as one giant filter for the country, updating and synchronising their filter lists as new telemarketing numbers got through and people pressed the big red "Filter!" button on their phones. More than a certain number of tagged calls from the same number would spill it over into the public filter list until no more tagged calls from that number had been registered for a certain amount of time.
I also forsee a market for a US cellphone version of the filter, as apparently over there they pay to receive calls, not make them. Thus telemarketing costs the call receiver both time and money.
Not to mention that with the Eliza idea, each filterchip would pull down a slight variation of the human-spoofing software and test it against telemarketing calls, then report the settings and success (amount of time wasted) back to the central system, which would then use a genetic algorithm to search for the most effective time-wasting set of parameters and upload those to each filterbox, honing its timewasting abilities with each telemarketing call.
As a particularly delicious side-effect, there would be ways to spoof the system and partially whitelist telemarketing numbers. However, the only way to do this would be to buy a lot of the filterboxes (each one with a unique ID) and set up a rather complex and expensive PABX-alike or phone-spoofing system.
However, the more unique IDs (customers) flag a number as a telemarketing number, the more filterboxes will have to be bought and set up on spoofing systems to counterbalance that - and whitelisting would require resources proportional to the square of the amount of blacklisting. This means that minimally-telemarketing numbers could be whitelisted by buying a couple of filterboxes and support systems, medium-telemarketing numbers would require a boatload of filterboxes to counter, and numbers used for massive amounts of telemarketing wouldn't be able to whitelist themselves by any process other than not using that number at all for six to 24 months.
This means that anyone wanting to whitelist their numbers will have to try buying lots of filterboxes. And while buying one might be cheap, buying dozens or hundreds certainly won't be. Another way to funnel money from telemarketers to the filterbox company.
Not that I've thought about it at all. Really.
Reply
Leave a comment