Blockchain: Blueprint for a New Economy

transactions will be executed, such that

malicious players would not be able to

get their transactions executed or

recognized on the network. Any

important transaction regarding resource

access and use might require assent by

consensus models. Thus, the way that

friendly AI could be enforced is that

even bad agents want to participate in

the system to access resources and to do

so, they need to look like good agents.

Bad agents have to resemble good agents

enough in reputation and behavior that

they become indistinguishable from good

agents because both behave well. A

related example is that of sociopaths in

real-life society who exist but are often

transparent because they are forced into

good player behavior through the

structure and incentives of society. Of

course, there are many possible

objections to the idea that the blockchain

structure could enforce friendly AI: bad

agents might build their own smart

networks for resource access, they might

behave duplicitously while earning trust,

and so on. This does not change the key

point of seeing blockchain technology as

a system of checks and balances for

incentivizing and producing certain

kinds of behavior while attempting to

limit others. The idea is to create

Occam’s razor systems that are so useful

in delivering benefits that it pays to play

well, where the easiest best solution is

to participate. Good player incentives

are baked into the system.

Some of the key network operations that

any digital intelligence might want to

execute are secure access, authentication

and validation, and economic exchange.

Effectively, any network transaction that

any intelligent agent cares about to

conduct her goals will require some

form of access or authentication that is

consensus-signed, which cannot be

obtained unless the agent has a good—

which is to say benevolent—reputational

standing on the network. This is how

friendly AI might be effectuated in a

blockchain consensus-based model.

Smart Contract Advocates on

Behalf of Digital

Intelligence

Not only could blockchain technology

and consensus models be used

potentially to obtain friendly AI

behavior, the functionality might also be

employed the other way around. For

example, if you are an AI or a digitally

uploaded human mindfile, smart

contracts could possibly serve as your

advocate in the future to confirm details

about your existence and runtime

environment. Another long-standing

problem in AI has been that if you are a

digital intelligence, how can you confirm

your reality environment—that you still

exist, that you are sufficiently backed up,

that you are really running, and under

what conditions? For example, you want

to be sure that your data center has not

shoved you onto an old DOS-based

computer, or deleted you, or gone out of

business. Smart contracts on the

blockchain are exactly the kind of

universal third-party advocate in future

timeframes that could be used to verify

and exercise control over the physical

parameters of reality, of your existence

as a digital intelligence. How it could

work is that you would enact smart

contracts on the blockchain to

periodically confirm your runtime

parameters and decentralized back-up

copies. Smart contracts allow you to set

up “future advocacy,” a new kind of

service that could have many relevant

uses, even in the current practical sense

of enforcing elder rights.

Speculatively, in the farther future, in

advanced societies of billions of digital

intelligences living and thriving in smart

network systems, there would need to be

sophisticated oracles, information

arbiters accessed by blockchain smart

contracts or some other mechanism. The

business model could be “oracles as a

service, a platform, or even as a public

good.” The Wikipedia of the future could

be a blockchain-based oracle service to

look up the current standard for digital

mindfile processing, storage, and

security, given that these standards

would likely be advancing over time.

“You are running on the current standard,

Windows 36,” your smart contract

advocate might inform you. These kinds

of mechanisms—dynamic oracle

services accessible by smart contracts

on universal public blockchains—could

help to create a system of checks and

balances within which digital

intelligences or other nonembodied

entities could feel comfortable not only

in their survival, but also in their future

growth.

Blockchain Consensus

Increases the Information

Resolution of the Universe

In closing, there is ample opportunity to

explore more expansively the idea of the

blockchain as an information technology,

including what consensus models as a

core feature might mean and enable. A

key question is what is consensus-

derived information; that is, what are its

properties and benefits vis-à-vis other

kinds of information? Is consensus-

derived information a different kind or

form of information? One way of

conceiving of reality and the universe is

as information flows. Blockchain

technology helps call out that there are at

least three different levels of

information. Level one is dumb,

unenhanced, unmodulated data. Level

two could be posed as socially