Gates predicts death of spam in 2 years

Pioneer Press ^ | January 24, 2004 | KEN MORITSUGU

[sarcasm]

DAVOS, Switzerland — Microsoft Chairman Bill Gates on Friday predicted the death of spam in two years and said his company is working on a new software that would make it impractical and uneconomical for marketers to send the mass e-mails that annoy so many e-mail users.

Gates made the bold prediction at the World Economic Forum, an annual gathering of business leaders in the Swiss skiing village of Davos.

Current anti-spam software uses filtering to remove e-mails with certain words that frequently show up in spam.

Some marketers have responded by leaving such words out of the e-mail subject line and providing only an Internet link to the marketing message.

The new software would require that any incoming e-mail from an unfamiliar address — one not in a user's address book — prove that it isn't spam, Gates said.

One possible approach would send a puzzle back to the sender. The puzzle would be designed so that only a human could solve it. The e-mail would be accepted only if the puzzle were solved.

[Focault's Pendulum]

One possible approach would send a puzzle back to the sender. The puzzle would be designed so that only a human could solve it.

Might slow things down a bit.

[goldstategop]

I've news for Bill Gates. Phil Goldman's http://www.mailblocks.com already utlilizes active "human response" technology. And its already cut back on the spam that flowed into my MSN mailbox - an account by the way, that's I've never used!

[Drango]

Two years? I'll take this bet.

[AzJohn]

And its already cut back on the spam that flowed into my MSN mailbox - an account by the way, that's I've never used!

Even without using any special software, the amount of spam I've gotten on MSN/Hotmail accounts has gone way down in the last couple months. Less spam overall and the stuff that does get through is more likely to go into the junk mail folder rather than my inbox. They've changed their setup somehow to cut back on this stuff.

[George Smiley]

The only way that Microsoft could take credit for this would be if it occurred as an "unintended consequence" of a new Microsoft operating system:

Everybody's PC (not to mention mailservers) would crash so frequently that the vast majority of spam [and legitimate e-mail as well] would never reach its destination.

There are clearly some folks writing code in Redmond who are incapable of changing their own underwear.

[Drango]

LOL

None of this matters. Congress has passed a law regarding SPAM which will take care of it.

[Larry Lucido]

The puzzle would be designed so that only a human could solve it.

"Ha, I can solve any puzzle any puny human could solve!"

[secret garden]

Yes but how much is it? It reminds me of the irritation factor of paying extra every month to keep my phone number unlisted.

[SamAdams76]

Yeah, but I'm waiting for the Microsoft package to come out. I try to keep my computers with 100% Microsoft software so that my computers run pretty much flawlessly. They are the only software company I fully trust.

[kingu]

It'd be better if they actually designed their mail servers to recognize that three emails that are the same likely indicates that it is spam and held it for human (or electronic) review.

I'll top Bill Gates' prediction.. Three days after his software is released, it will come out that spammers are buying licenses from Microsoft to send spam mail and Microsoft will shrug and go, "They're legitimate bulk mailers who applied for and were granted licenses to fill up people's mailboxes." Fees will of course just go to Microsoft and licenses will range around 25,000 USD each.

[Gorzaloon]

I've news for Bill Gates. Phil Goldman's http://www.mailblocks.com already utlilizes active "human response" technology. And its already cut back on the spam that flowed into my MSN mailbox - an account by the way, that's I've never used!

Challenge-response whitelisting schemes have been around for a couple of years. They need work, because of subscribed mailing lists where no "human" is doing the sending. But it's a step in the right direction.

Not as good a step as impaling spammers by the roadside.

My domains allow me to create a large number of addreses, all of which I then forward to my main account. For example, "EBAY@domain.com", etc. In this way I can at least find out who is harvesting or selling addresses, or otherwise releasing them by default, and then blackholing everything from them till the end of time and giving them no further business. That address is then killed. I went from hundreds of spams a week to one or two.

But my contention is, "Why should I even have to do this at all?"

[Cvengr]

The puzzle would be designed so that only a human could solve it. The e-mail would be accepted only if the puzzle were solved.

Sortof the same strategy as translating Arabic chat room messages from suspected Al Qaida members. Gee, I wonder if the same programming time spent might have dual use.

[jimtorr]

This is the same man who once said that nobody would ever need more than 64k RAM in their desktop PC.

[Fester Chugabrew]

Anyone who sends an email under false identity ought to be subject to a stiff penalty under law. As with phone calls and mail, those who use these as a means to harass the public should be spanked with fines (and jail time if need be).

[ladyjane]

Was it Gates who said that? I thought it was Ken Olson of DEC who said something similar to that. John Watson of IBM said there was a world market for only 5 computers. Of course Lewyt of vacuum cleaner fame predicted that by 1960 everybody would have nuclear powered vacuum cleaners.

[ml/nj]

I still say: Make any debt arising out of acceptance of a spam solicitation uncollectable in the United States, and commercial spam will dry up in a couple of months.

No one listens to me though.

ML/NJ

[11th Earl of Mar]

Like I believe MicroSoft wants to stop spam!

That's like saying that the phone company wants to stop telephone solicitations by telemarketers.

[Tijeras_Slim]

Nonsense. I've got a can of this stuff in the pantry that's atleast 5 years old. It's probably still good.

[Rifleman]

That would help. Fairly simple Baysian (just means that the prog does a word count and when you tell it what is spam and what is not, assigns a score) filters eliminate .999 of spam and have false positives of < .005 with very little training. They also get better with time, and match _your_ email usage patterns.

This challenge and response stuff has potential to be a PITA. Once it is trained, which in a properly designed email client consists of clicking a checkbox on emails that you recognize as spam, the baysian filters require no user intervention. The email client should have these built in and on by default.

Gates will use this to create more lockin on his victim<<<<<<customers. The implimentation will probably require a microsoft email client to do the reply.