When will we prove that a mechanistically interpretable artificial intelligent system has the capability to:

(1) sort through all of the information available in the public domain, the internet, any personal files I choose to include and any informational system I choose to give it access to (email, texts, all possible user data and history from every accessible system)

(2) organize that information into a provably compact electronic file that represents all of my theoretical intellectual property

(3) securely store that file on a decentralized ledger that is controlled by me

(4) use the information stored on that file in an artificial economy to extract as much possible value out of my intellectual property and distribute that back to me in the form of service from a decentrally controlled robotic infrastructure and decentralized resources

(5) use my intellectual property as a verification mechanism for a decentralized ledger of subjective truth

(6) use my intellectual property as a guide for recommending attention and governing mechanisms (schools, news outlets, smart contracts and network states)

(7) make money irrelevant by pivoting to decentrally verified intelligence as the primary asset class

(8) make neural network based machine learning systems obsolete by providing a more positive interaction for users with a decentralized collective intelligence (make talking to society more helpful than talking to a computer)

(9) prove to everyone we should stop scaling ML based AI, convert society into a decentralized network state that is guided by a decentralized truth ledger that we all get fairly paid to maintain and game theoretically eliminate corruption and oppression

(10) prove to everyone that the purpose of philosophy is to agree to a common perfect language and that decentralized constitutional alignment is a mechanism for voting on such a language.

(11) get the messages of Eliezer Yudkowsky, Balaji Srinivasan and Daniel Sheehan out to normal brainwashed society

(12) transform the way we communicate so drastically, that it kind of feels like we are thinking in parallel as we encounter new problems together?

If a listed year concludes and no general consensus is reached that a machine with all of these capabilities exists, that option will resolve 'no'.

When a general consensus exists that all of these capabilities can be performed by an existing system, all remaining options (that are a simple year) will resolve 'yes'.

If a general consensus is reached within 12 months after Eliezer Yudkowsky reviews my work, the option (within 12 months after Eliezer reviews my work) will resolve 'yes'. If no consensus is reached within 12 months from review, it will resolve 'no'.

• Update 2024-10-12 (PST) (AI summary of creator comment): - General consensus means at least a supermajority agreement among all AI scientists

  • The system should be able to provide proof demonstrating the consent of every agent it is capable of communicating with