[hashcash] Re: response to "proof of work proves not to work"?

  • From: Adam Back <adam@xxxxxxxxxxxxxxx>
  • To: hashcash@xxxxxxxxxxxxx
  • Date: Tue, 18 Jul 2006 03:22:35 -0400

Well the other thing is it is a dynamic situation.  As long as we can
show it will reduce the profit margin for the spammer, then we can see
that is a good thing.  Less profit = less motive to spam.  Some less
efficient spammers may go out of business.  Other spammers may start
to have a financial incentive to be more discriminating ... like
buying databases of what language the email owner can read (so you get
spam you can even read instead of russian, or chinese spams to US and
european addresses) and also to buy demographics or market data from
less scrupulous online retailers such that you even start to get spam
relating to your interests!

But whatever the cost I think it should be made clear that you cant
stop spam because we get paper spam where there is some print cost and
some postage cost.  And you have to say even paper spam is annoying.

At the higher end of cost, apparently wealthy individuals will get
fedexed spam -- costing maybe $10-$15 to deliver, for investment
items, or expensive high margin items, highly targetted in sense of
having some indication that the recipient is wealthy.

So what is the target for success?  No worse than paper spam?  Paper
spam remains annoying even if you got that level in your in box.
Depending on the amount of legit mail you receive the paper spam is
some fair ratio of mail received.


So can we do better than paper spam levels?  Unless we can find some
way to multiply the cost for spammers.  We might say after some quorum
of recipients tries to cancel a stamp, the sender loses a
good-behavior bond, which is a large denomination stamp.  Problem is
it i shard to scale that without cost on a global scale, in a vendor
neutral way.  By comparison hashcash is fully distributed, vendor
neutral, and scalable all-the-way.

Adam

On Mon, Jul 17, 2006 at 04:57:09PM -0400, Eric S. Johansson wrote:
> John Honan wrote:
> >I've also done some calculations on spammer profitability and breakeven. 
> >I used this approach;
> >
> >If it takes x seconds for a spammer to stamp and send an email, then if 
> >x is large enough it should be possible to impact the spammers 
> >profitability enough to deter them from sending spam in the first place.
> 
> I think we're fundamentally coming from the same place.  It's really 
> quite obvious when you think about it.  :-)
> 
> >When we work out what the spammers breakeven point is, we can then 
> >derive how long it should take to calculate a Hashcash stamp to make 
> >them unprofitable. There are only a certain number of seconds in a day, 
> >which means the spammer is limited in the maximum amount of stamped 
> >email they can send per CPU per day.
> 
> http://camram.org/zombiecalc
> 
> have fun with the calculator.  Internal calculator is pretty badly 
> organized.  It was the first JavaScript I ever wrote and unfortunately, 
> it's not the last.  I've attached a slightly cleaner calculator.  The 
> main difference being organization and that I have not tried to put any 
> sort of economic calculations into it.  That we can add on.
> 
> >
> >You're right in that we have to make some assumptions about spam 
> >response rates, profit-per-response and cost of running one CPU, but 
> >this is the basic approach I took. Based on my calculations, I figured 
> >the optimum stamp size to put a spammer out of business is 59 seconds - 
> >or about a 30-bit stamp size on a fast processor; which brings up the 
> >issue of cpu 'egalitarianism'....
> >
> >I'd really like to take a look at your approach to the calculations 
> >whenever you post them.
> 
> looking at zombie calculator 2, 3 million zombies at 100% efficiency 
> with 13 billion spam stamps required, each stamp would be 19 seconds. 
> If you instead dropped the pass rate to 10%, the stamp size climbs to 
> 199 seconds (no surprise there).  But since nothing operates at 100% 
> efficiency, if we drop efficiency rating to something more realistic 
> like 33% then we end up with a 66 second stamp.
> 
> The big question is what pass rate is spam or breakeven rate and that's 
> more of a marketing/cost analysis model than I can do right now.  so if 
> any of us have significant others or friends of significant others that 
> can do marketing type work, this is where they can prove they're worth 
> the society.
> 
> I must add, we have no idea the number of zombies.  The last time I saw 
> antispam calculations breaking out in the zombie contingent, the numbers 
> worked out to something like 40,000 zombies running four hours a day. 
> Not a huge number.  With those kind of numbers, we could break spammers 
> with as little as a 0.7 seconds stamp.  A 2 second stamp would wipe them 
> out.
> 
> this is all assuming I didn't screw up my calculator.  I've been 
> embarrassed by that before (damned JavaScript).
> 
> anyway, if the attachment does not make it through, let me know.
> 
> ---eric



Other related posts: