That sounds like you're talking about subscribers with no data plan, the most expensive overage fee for data plans is $59.96 per GB (not GiB as you mistakenly gave the price for).
However for most of the dataplans it is $10 per GB as per
http://www.att.com/shop/wireless/plans/dataplans.html
Elsewhere on their site (burried in http://www.att.com/shop/wireless/data-plans.html It looks like they may be going up to $15 per GB. All of these prices, even the highest of $59.96 per gb are far lower than your listed $256 per GB(sic)

AT&T revised the 'unlimited' contract terms more than two years ago and you agreed to them when you signed a new contract

You've described AT&T's only data plans new subscribers can get now.
The capped plans are the old grandfathered 'unlimited' plans that no one can subscribe to now. All new subscribers subscribe to say 5 gigs per month and after 5 gigs they charge like $10 per gig

You should probably let the US Treasury know that as they disagree with you.
http://www.treasury.gov/resource-center/faqs/Currency/Pages/legal-tender.aspx

I dont believe that driver is from FTDI. A) Russ Dill is a known kernel developer. B) from what I udnerstand that patch would actually brick legit FTDI FT232RL devices as well. He submitted the patch as a joke but it looks like it just overwrites the checksum located at an even address rather than using a collission by modifying only odd addresses. That even address write would cause a write on the real FTDI chips as well.

Actually in the uk gun crime has doubled as well since the ban.

The far more interesting thing is violent crime has gone up for those areas that enacted gun bans while for the rest of the world violent crime has actually gone way down.

In his defense going out and shooting at little bits of paper is likely a more productive use of his time, and an excellent alternative to sitting on the computer when your internet is too slow.

US deer are grazers so their heads tend to be low to the ground despite standing much taller. We dont have carnivorous deer that go for our hearts like you seem to have in Germany.

Aha, so you missed the original quote, i'll try bolding the relevant parts this time.

I was specifically replying to that part, as TLS and TTLS both have the same degree of mitm vulnerability with properly configured clients.
If the server cert fails in TLS or TTLS then MITM is a possibility, you dont need the username/password or client cert to mitm a TLS connection, just the server cert.

Actually for that matter wouldnt a compromised server certificate leave you vulnerable to a proxy attack anyway where you would use the compromised server cert to pretend to be the access point communicating with the proper radius server thus giving MITM on TLS or TTLS the same? You might not get the actual client cert on TLS but you would have their traffic all the same.

If server cert validation has failed chances are your CA was compromised, in which case the attacker could just generate client certs at will anyway....

TTLS properly configured is no easier to MITM than properly configured TLS, you should be using server cert validation with either.

I dont think a cert server works the way you think it does.

I mean technically it has costs... but theres not a lot of reason you can't use a $300 convertible tablet pc handle your ca cert virtually indefinitely, it doesn't have to be turned on after you finish signing certs until its time to sign another batch...