In an interview at The Internet Engineering Task Force 93, former CIA employee and NSA contractor Edward Snowden spoke about privacy and security. Interestingly, Snowden focused his talk on Bitcoin and its flaws and weaknesses.

Snowden began by briefly introducing the 51% attack concept, and stated that the Bitcoin network could be vulnerable to attacks and manipulation if “a pool has over 50% of the computing power to demonstrate adequate proof of work to add entries to the blockchain ledger.”

Snowden commented,

“Obviously, Bitcoin by itself is flawed. The protocol has a lot of weaknesses and transaction sides and a lot of weaknesses that structurally make it vulnerable to people who are trying to own 50 percent of the network and so on and so forth.”

He continued and said, “Focusing too much on bitcoin, I think is a mistake. The real solution is again, how do we get to a point where you don’t have to have a direct link between your identity all of the time? You have personas. You have tokens that authenticate each person and when you want to be able to interact with people as your persona in your true name, you can do so.”

Edward Snowden

Implementation of Proof of Work

Despite his concern for the digital currency, Snowden believes the concept of tokenization and proof of work could be implemented to create “some very interesting things,” which “people can basically pay for access other than direct transfers of currency that originated with an association to their true name.”

Snowden suggested a similar concept of Tor, where bitcoin transactions are “tumbled” and changed once it is sent, like DASH’s Darksend. Darksend mixes a pool of transactions and its addresses randomly, making the transactions completely anonymous and untraceable.

The anonymity of DASH allows users to have multiple “persona” – “a common persona, an anonymous persona, a shared persona” to send transactions that cannot be tracked by the NSA or any government agencies.

Snowden explained, “A lot of these are difficult problems particularly when we talk about the metadata context, the signalling context. And there are actually some really bad proposals, I think, and this is in no offense to anybody who works on these particular problem spaces, but again, it gets back to the middlebox space.”