Bitcoin network is vulnerable like never before

[lucif]

Based on A successful DOUBLE SPEND US$10000 against OKPAY this morning

Two factors.

1. Transition to ASICs
Network hashrate is low now, and interested parties can get >50% of hashrate power due to new asic technology.
With that they can create financially effective 51% double spend attack:
- Create a bitcoin transaction $tx1 to own address $a2 from own address $a1
- Wait 1-3 blocks
- Turn on hash power, start mining from 1-3 earlier blocks telling your miner to ignore your transaction $tx1. Enough hash power will create chain fork where $tx1 is not confirmed.
- Create a double-spended bitcoin transaction $tx2 from own address $a1 to some financial service/marketplace address $a3
- As long as your hashpower is up and it ignores $tx1, service/marketplace thinks that $tx2 is valid and confirmed.
- Make clearing to fiat that $tx2
- Configure miner to start mining back from parallel chain fork where $tx1 is confirmed
- Service/marketplace looses fiat.

2. maxblocksize
Obviously this shit should be changed soon or later. With following hard fork and another "A successful DOUBLE SPEND" topics.

[Bit_Happy]

F.U.D? 
It was not really successful, according to that thread.

[Maged]

F.U.D? 
It was not really successful, according to that thread.

It was. Part of the problem is that we always assumed that the solution to hard forks would be for people to upgrade their client. It never occurred to anyone to add protection against newer clients being incompatible with the older clients. If Bitcoin-Qt had (hard) fork detection on the side that accepted both chains, this could have triggered safe mode for everyone, preventing double-spent coins from being accepted in the first place.

[benjamindees]

Yes, this is true.  I've been thinking for a while that there should be some concessions made to retain as many GPU/FPGA miners as possible until ASICs are more widely distributed.  Perhaps leaving 'maxblocksize' as is for now would accomplish this?