Tau-Chain and Agoras Official Thread: Generalized P2P Network

[kjn311]

Just tell me you are building Skynet: https://en.wikipedia.org/wiki/Skynet_(Terminator)

Skynet is a fictional neural net-based conscious group mind and artificial general intelligence (see also superintelligence) system

tau is not neural net. we'll reach skynet once we implement a neural net over tau (easy) and train it (hard).
the direct training of tau is not by machine learning methods but more like "rule based AI" methods

Last question and I'm sure many here would like to hear the answer to...I believe I heard this in a podcast...

I can use Bitcoin over Tau?

[1714197892]

1714197892

[1714197892]

1714197892

[1714197892]

1714197892

[1714197892]

1714197892

[ohad]

I can use Bitcoin over Tau?

of course. but not over the first alpha, which will support theories in more general fashion, but not support "processes" yet. the next alphas will deal with such transition, generalizing knowledge to programs. we're able to do so thanks to releasing ourselves from the curry-howard approach.

[dmitryshech]

Tau needs a use case and just market that. Ethereum use case was smart contracts. To me, Tau "main" use case is Artificial Intelligence. Market AI when talking about Tau. Maybe even change the name to something cool and robotic like "Skynet".  

indeed an important motivation of tau, is to bring the power of AI to everyone altogether (and by everyone altogether), rather to be held by centralized entities

Decentralized AI it's the whole different story! Skynet has some bad connotations though
In my opinion, right branding and marketing it's very important, maybe even the most important part. No doubt, tau needs the right story that can spread the idea, story that solves needs or wants of people. Most of the crypto projects today are not even close to the ideas of tau but with the right marketing some of them succeed in wrapping it in a very desirable package, the result of it is attention and investments.
In case of tau that could be 100% authentic story because the value that tau can give to the world is truly game changing.

tau is not neural net. we'll reach skynet once we implement a neural net over tau (easy) and train it (hard).
the direct training of tau is not by machine learning methods but more like "rule based AI" methods

we will have an intelligent social network for not only serious stuff, but also for fun.

Is this to quotes can be combined? Decentralized AI that we going to train all together in a kind of game?

[ohad]

tau is not neural net. we'll reach skynet once we implement a neural net over tau (easy) and train it (hard).
the direct training of tau is not by machine learning methods but more like "rule based AI" methods

we will have an intelligent social network for not only serious stuff, but also for fun.

Is this to quotes can be combined? Decentralized AI that we going to train all together in a kind of game?

you're getting there.. not fully though, here

every team will "train" (collaboratively and comuter-aided-ly build a theory) their own project.
in parallel to that, every person/profile will have their own theory, perspective, built from their own sources, and by agreeing/disagreeing to others, etc.
add to that the special root team, which defines the system itself.
then we can use the machine to query theories, or prove assertions, or discover new knowledge, or check consistency, or calculate the consensus, or synthesize a model/program that admits a theory/spec...

can ask the following question:
how can one make sense of thousands of people giving advice on how to build a specific software?
tau is all about that.

[Cibervision]

TAUCHAIN -  the power of AI to everyone altogether (and by everyone altogether), rather to be held by centralized entities
Sounds really great, this sentence is Amazing Ohad. Explain tauchain shortly and powerfully

10 points for you

[dmitryshech]

tau is not neural net. we'll reach skynet once we implement a neural net over tau (easy) and train it (hard).
the direct training of tau is not by machine learning methods but more like "rule based AI" methods

we will have an intelligent social network for not only serious stuff, but also for fun.

Is this to quotes can be combined? Decentralized AI that we going to train all together in a kind of game?

you're getting there.. not fully though, here

Finally I am getting somewhere, and I am not the laziest one... , some could give up

every team will "train" (collaboratively and comuter-aided-ly build a theory) their own project.
in parallel to that, every person/profile will have their own theory, perspective, built from their own sources, and by agreeing/disagreeing to others, etc.
add to that the special root team, which defines the system itself.
then we can use the machine to query theories, or prove assertions, or discover new knowledge, or check consistency, or calculate the consensus, or synthesize a model/program that admits a theory/spec...

can ask the following question:
how can one make sense of thousands of people giving advice on how to build a specific software?
tau is all about that.

Hmm... Can't even imagine where it can get us. It's like with the time bringing all the world's knowledge in one place and having power and abilities to properly sort it, operate it and constantly grow and improve it.  Is it something that already tested? Can you estimate when the first code will be ready?

The idea is amazing, and clearer now. thank you Ohad for your efforts!!!

[ohad]

Is it something that already tested?

there are some similar concepts
https://en.wikipedia.org/wiki/Semantic_social_network
https://en.wikipedia.org/wiki/Social_Semantic_Web
(they're more like the first alpha, nothing like the next alphas)

but i didn't find anywhere (not even in game theory yet) something about collaborative theory formation (in contrast to numerical rather logical data. and in contrast to players each with their own preferences and score. here we're about groups of people each having more or less shared goal, and a logic-consensus driven process. even nomic is not about how to build a theory, but about how to build the game itself). seems like we're establishing a new field here.

Can you estimate when the first code will be ready?

very hard to estimate. at least for me. i must admit that part of the delay is due to new ideas that don't stop coming. a recent one was to convert your theory to a wiki, so you could browse the encyclopedia of your own worldview. or let others browse it. (or compare it with another one's, for the sake of partnership, or dating, or common interest.. but that's an old idea, the new idea is the wiki format). and obv most of the time i think of the math aspects of this all which are nontrivial. especially how to write a scalable solver

[kjn311]

How do you secure the network without a native coin? The thing about Ethereum and Bitcoin is it cost large sums of money to attack.

[ohad]

How do you secure the network without a native coin? The thing about Ethereum and Bitcoin is it cost large sums of money to attack.

forming the blocks will begin centralized so no problem for the beginning.
moreover, many of the network's features, will not require blockchain at all (anything that doesn't care which rule came first).
in the process of forming the decentralized version using the centralized one, we'll have to consider mining reward. the amount of miners we'll need will be more or less proportional to the cap we intend to support. i mentioned few times that i'm for "miners tax", but that's my personal opinion on an issue to be agreed collaboratively.

[klosure]

Really interesting interview of Arthur Breitman, the lead developer of Tezos.
Tezos is one big leap forward in the direction of Tau. It has a lot of common points with Tau. For instance, the protocol is self amending following a very similar process to Tau that is also inspired by the nomic game. Nomic is actually being discussed in the interview. Another important similarity is the use of a custom functional language practical enough to develop with and that comes with built-in formal verification capabilities, the goal being to ship code with a mathematical proof of its properties. Another very similar aspect and that is beautifully well explained in the interview is that blockchain networks are essentially all following the same protocol at a very high level, when you abstract specific details like the exact nature of the state that is being validated, and the exact way that the canonical version is being discriminated, blockchains are essentially all doing the same thing, so Tezos is coming with built-ins that can handle all the low level stuff like i/os as well as generic blockchain mechanics, and all the specific blockchain rules are left for the users to decide in the genesis block (they call it "seed protocol") and subsequent protocol amendments. This allows Tezos to implement Bitcoin, Ethereum and any other blockchain. Even the idea of writing program specifications as a set of constraints and letting the compiler come up with an implementation that is guaranteed to meet the constraints is addressed. A last thing: Breitman alludes at some point to the fact Ethereum is doing a mistake by attempting to be a "universal computer" that can compute everything, and explains that a blockchain network doesn't need full spectrum computation and that a subset of what Ethereum can do is sufficient. Although he doesn't point specifically to concepts such as turing completeness and decidability, and doesn't state where Tezos stands on that aspect, it sure seems like he is referring to that but keeping things simple to avoid perplexing the non-technical audience. Is Tezos decidable too?

Now, that makes a hell lot of common points with Tau, doesn't it? And they are planning to release this year!
Ohad, are you following Tezos closely enough? What's your take on their technology?

[klosure]

Actually they already have a functional system and released their code of Github. Arthur Breitman  started a thread here earlier this month and it seems that they are looking at doing a crowdsale in Q1 or Q2, so pretty soon. Let's ask him directly if Tezos is decidable

[ohad]

it is indeed a decent project, i saw him speaking several months ago and was impressed. it might be the only thing out there that is close to tau. that said, there's still a big gap. i'd ask two things before decidability: expressibility, and which parts of the system are amendable (is it the whole code no matter what?). and that, taking into account the old tau. the new tau refines the process much more than nomic.

[ohad]

two more points:
1. need to understand what they mean by "ledger" and in what it's different than "generic knowledge". EDIT: indeed they seem to focus on contracts
2. ocaml impl?! ... the easy but eventually-useless way?

[ohad]

.. and after listening to the interview:
he says that people will prepare their proposals and then by some vote (60%) they'll decide which proposal to accept.
my bigger question is, and is a question "against" the old tau as well, and "against" nomic:
how do you expect people to form well-formed proposals, that actually express what they meant and so on?
do you expect this to be an individual process for experts?
on the new tau we consider this a collaborative process for non-experts, the process of forming a proposal, aside the process of accepting it.

a smaller point would be regarding votes. once you take a close look, you don't need them anymore

[ohad]

two relevant links:
https://bitcointalk.org/index.php?topic=950309.msg14904326#msg14904326
https://steemit.com/crypto-news/@dana-edwards/tezos-the-generic-self-amending-crypto-ledger (see also comments)

[klosure]

It might be the only thing out there that is close to tau.

Well, now there is also Autonomic

that said, there's still a big gap.

I agree. Although there is quite a bit of overlap on the fundamentals, Tezos seems to be specifically aiming at protocol governance, whereas Tau is both much more general in scope and closer to the metal due to the versatility of its RDF syntax which gives it the ability to blend naturally with semantic structure of which it is only a flavor. I think we can probably learn a lot from the launch of Tezos, their approach, their mistakes, and how things pan out with the nomic game.

i'd ask two things before decidability: expressibility,

I asked that already

and which parts of the system are amendable (is it the whole code no matter what?).

In the EpiCenter interview, he talks about layers, of which only the topmost layer that contains functional rules about the way the blockchain is working and the business logic would be self-amending, which includes consensus rules and voting. Based on that, I think all of the lower level stuff like network communication, overlay network management, DHT etc are hard coded and evolve following a regular software life cycle. This is actually made more explicit later in the interview where Arthur Breitman explains that only changes that involve governance issues are really controversial and subject to on-chain consensus, whereas more technical issues such as bug fixes are following the normal software life cycle and adopted following the normal soft/hard fork rules as people update their clients.

1. need to understand what they mean by "ledger" and in what it's different than "generic knowledge"

probably ledger = blockchain, generic knowledge = whatever state the blockchain is used to maintain?
Can you provide the context?

2. ocaml impl?! ... the easy but eventually-useless way?

He mentioned in the interview that he was inspired by Coq for the formal verification. Coq is written in OCaml so going with OCaml would probably have saved quite a bit of time. I can't find back the detailed team page of Tezos, but if I remember well there were some people from Inria, so that would also explain why they went for OCaml.

Just noticed that Andrew Miller is an advisor of the project. Wasn't he also following Tau? And there is Zooko too! What a small world. I wonder if HMC knows the tezos guys.

[klosure]

how do you expect people to form well-formed proposals, that actually express what they meant and so on?
do you expect this to be an individual process for experts?
on the new tau we consider this a collaborative process for non-experts, the process of forming a proposal, aside the process of accepting it.

I think it's wishful thinking to imagine that non-experts will be able to amend the protocol or understand deeply what they are voting for. Tezos's solution apparently is to allow people who don't understand what they are asked to vote for to delegate their votes to someone they believe knows better and shares their perspective. Not a bad idea. At best, you could have experts rephrasing the decision in simpler terms in plain english, but then you would still need to trust that they are not misrepresenting the problem or showing a bias, and you would still need people to really try to understand what they say at a logic level which isn't a given, and then there is the laziness. As pointed in the interview, the DAO was a good example of how in practice, most people won't really bother voting.

a smaller point would be regarding votes. once you take a close look, you don't need them anymore

Would that be a solipsistic look or a totalitarian look or an omniscient look ?

[ohad]

1. need to understand what they mean by "ledger" and in what it's different than "generic knowledge"

probably ledger = blockchain, generic knowledge = whatever state the blockchain is used to maintain?
Can you provide the context?

i saw it on their website. by generic knowledge i basically refer to the expressiveness of their language.
they do require you to write code. code can be seen as a proved statement, which is fine, but is up to the programmer to prove in many cases, however it requires more than that. this has structural implications on how you can say what you have to say. it's a point i'm frequently strugging to explain, but let's try the following. when we write a contract, or a book, we mean that all sentences written there are true. we implicitly mean that the reader should take the *conjunction* of all sentences/clauses/paragraphs. i refer to this as the "conjunctive nature" of how we express ourselves. proofs and programs, on the other hand, are of compositional nature. think i'd like to express an in idea for you in writing, so i write a paragraph with many holes such that you cannot understand a thing out of it, and then i give you words and half-sentences to substitute on those holes, many of such sets of substitutions, and you'll have to read that first paragraph again and again but with different substitutions in the hole places. you'd hate that. but machines love that. i believe that we should stick to our conjunctive nature, and let machines translate it by themselves into a compositional nature.
what this practically means is,
specification is also of conjunctive nature. typically, you want a program that do/satisfy "this AND that AND that". the order doesn't matter. unlike the curry-howard approach in old tau (similar also to prolog), on the new tau executing the program does not mean "executing the spec/logic", namely the operations taken have nothing to do with the reasoning process, and we don't invoke curry-howard correspondence at all. the logic is about the program, is not the program itself. and has a conjunctive nature, especially when it comes to large number of participants, each stating their opinion.
there's much more to say, and i'll say slowly.

2. ocaml impl?! ... the easy but eventually-useless way?

He mentioned in the interview that he was inspired by Coq for the formal verification. Coq is written in OCaml so going with OCaml would probably have saved quite a bit of time. I can't find back the detailed team page of Tezos, but if I remember well there were some people from Inria, so that would also explain why they went for OCaml.

Just noticed that Andrew Miller is an advisor of the project. Wasn't he also following Tau? And there is Zooko too! What a small world. I wonder if HMC knows the tezos guys.

indeed we don't follow the coq way in many senses. not to say anything bad about coq, the contrary. it's just for slightly different purposes.
yes Miller had a lot to do with "proof of execution", economically hashing the execution tree for selected datatypes. Zooko is a very knowledgable guy as well.

how do you expect people to form well-formed proposals, that actually express what they meant and so on?
do you expect this to be an individual process for experts?
on the new tau we consider this a collaborative process for non-experts, the process of forming a proposal, aside the process of accepting it.

I think it's wishful thinking to imagine that non-experts will be able to amend the protocol or understand deeply what they are voting for. Tezos's solution apparently is to allow people who don't understand what they are asked to vote for to delegate their votes to someone they believe knows better and shares their perspective. Not a bad idea. At best, you could have experts rephrasing the decision in simpler terms in plain english, but then you would still need to trust that they are not misrepresenting the problem or showing a bias, and you would still need people to really try to understand what they say at a logic level which isn't a given, and then there is the laziness. As pointed in the interview, the DAO was a good example of how in practice, most people won't really bother voting.

think of the "facebook with controlled english" metaphore.

a smaller point would be regarding votes. once you take a close look, you don't need them anymore

Would that be a solipsistic look or a totalitarian look or an omniscient look ?

by no means the system will tell people what to think. but votes isn't necessarily the best way. your opinion can be extracted from all your other opinions that you stated, or that others stated and you agreed/disagreed with. can you imagine people voting for proposals all day, rather state their worldview "once" and forever? (they may change their mind too). that's in very very short

[ohad]

voting is such a black-or-white thing. if you propose something, am i allowed to say only "yes" or "no"? how about a system supporting a more constructive yet formal discussion? and what if your proposal is incomplete, yet not worthless, and you want to propose others to offer you a completion?
and what if your proposal is contradictory, yet not worthless? there might be, for example, several ways to resolve the contradiction, and you'd like your friends (or team members) to suggest you which to choose.

[ohad]

in real life, forming a theory together, is a live social process, that inevitably, contains many non-theories, while everything else would be non-realistic for humans. that's one of the biggest recent lessons