Readability of Tau language

[martijnbolt(OP)]

In the educational videos it is mentioned knowledge in Tau net will need to be written in computer readable language.

What is this language, what does it look like and will it be readable by common users or only by experts that understand this language?

[S1r_4zdr3w]

It is in formal language. So in the beginning, Tau users need to know logic and boolean algebra and stuff Ohad mentioned in the video series lecture in order to be able to participate in the early stages.

Playlist:
https://www.youtube.com/watch?v=lNU1wS5rqtg&list=PLav2klOnTUlOeakJCbLZxoib_x0jYAQ5f&index=1

Atomless Boolean Algebras - Introduction - 1 / 8

This is also the plan for the Tau team to create a controlled natural language where users who are non experts could participate and write their opinions.

There is a difference between Natural Language & Controlled Natural Language. I recommend for anyone interested to see the difference, it was discussed by Prof. Franconi on the interview with David Shapiro,

in this link
https://www.youtube.com/watch?v=CGJ0qivcZJw&list=PLeBSjZu0y6vVNYA4mV2nrMW0lN9YVlnNz&index=4

where natural language is too powerful and we have to limit it, and that's why it is called Controlled Natural Language as what will be build in Tau in the future.

The Tau team is writing the code for Controlled Natural Language, though the priority at this stage right now is the Tau Language Demo.

https://www.youtube.com/watch?v=f5lZ2xw15R0&list=PLeBSjZu0y6vUOVx9LUtfLyL-yIyePg4O4&index=5

Here are tips for anyone in the future who wish to use the controlled natural language effectively,

> Make your sentence accurate and clear
> The more specific your the opinion, the better
> Avoid words that are subjective or can have different interpretation


Remember that self contradicting statements leads to unsatisfiability. I have seen this in TML (Tau Meta Language), you can run it,

here is the link:
choose https://tml.tau.net/

Go to folder
Regression tests>unsat>

where Tau which uses logical AI, cannot reason over (unsat contradition) and (unsat infinite loop), as this leads to undecidability.  I will give more examples in the future for Tau Language.

So when we build our worldview, for sure we may see ourselves having self contradicting thoughts our old self comparing to our new perspective and that's natural. Tau will help us by automatically highlighting our self contradictions 

https://www.youtube.com/watch?v=ZcIQO5k7zr8&list=PLeBSjZu0y6vUOVx9LUtfLyL-yIyePg4O4


and so we can fix it, refining our worldview as we do more iterations. It will be an ongoing process of refinement and optimizing our perspective and worldview in Tau.

 

[martijnbolt(OP)]

Thank you for this elaborate answer. I have 2 follow-up questions:

1. If all knowledge is stored as boolean algebra not many people understand, will the system be able to output knowledge in natural language autonomously  so basic (non-technical) users can verify the content?

2. If we need experts to translate knowledge to boolean algebra or develop software to do this for basic (non-technical) users, how can users trust/verify the data or software? How can we trust this translation process, from natural language to BA and vice versa, is and remains decentralised and verifyable for most people (and therefore that Tau is trustworthy)?

[S1r_4zdr3w]

1.
The goal is "users only need to say their opinion, in a disorganized manner. Just like people go to Facebook at random times of the day, and put a random opinion of theirs. And by that, the system will be able to calculate which opinions the people agree on and disagree on." https://www.youtube.com/watch?v=m6t01smp5Hc&list=PLeBSjZu0y6vXmFwoi1BH7uwb2w_642aKW&index=2

We can leverage llms to assist us in converting natural language into formal language.
https://www.youtube.com/watch?v=1MfC1d5rCjo&list=PLeBSjZu0y6vXot_3PQoYFx1lSWd9Eqhq0&index=2

There will be future efforts to have a better user interface of Tau.
https://www.youtube.com/watch?v=ZcIQO5k7zr8&list=PLeBSjZu0y6vUOVx9LUtfLyL-yIyePg4O4

Trusting Tau's output will depend on the user requirement. Some users may need mathematical proof in some cases, some may need credentials in some cases. It's case to case basis, so it's really up to the user.
https://www.youtube.com/watch?v=kOJfvzo3mhY&list=PLeBSjZu0y6vWInPFgM7nCDD7VjqSDViu6&index=6

Proof of Execution can provide short cryptographic proof that the execution was correct
https://www.youtube.com/watch?v=xfFl9H1yHzo&list=PLeBSjZu0y6vV-vSKu9EcoaceQEiWtYcP6&index=9


The Attestation and Verification System patent (credentials)
https://www.youtube.com/watch?v=ykVUja8m8JE&list=PLeBSjZu0y6vXmFwoi1BH7uwb2w_642aKW&index=29

2.
Tau, a logical AI, will build the Opinion Map.
https://www.youtube.com/watch?v=gil9twYlWSk&list=PLeBSjZu0y6vXmFwoi1BH7uwb2w_642aKW

How to trust/verify the data or software?
Proof of Execution can give you a cryptographic proof that the running was correct.
https://www.youtube.com/watch?v=uNYa0W6xIBE&list=PLeBSjZu0y6vVNYA4mV2nrMW0lN9YVlnNz&index=7

How is Tau trustworthy?
"The world over Tau, everything is opinions. Therefore there can be of course contradicting opinions. Now for the case that the majority is wrong, this happens also in real life of course, but on Tau we don't have a full solution but we are in a better situation than in real life because a logical engine is able to detect contradictions. So sooner or later, a wrong opinion will run into a contradiction and by that be discarded. In this sense, we are in a better situation than in real life. "
https://www.youtube.com/watch?v=6FH_d0xecXM&list=PLeBSjZu0y6vWInPFgM7nCDD7VjqSDViu6