DACs

[coinrevo]

there is some buzz about the concept of a distributed autonomous corporation. I wanted to find out who invented the concept, but as these so called whitepapers floating around don't use scientific standards for citation I haven't been able to identify where the term comes from. google scholar doesn't return any reference on a keyword search.

Szabo wrote this paper called "Multinational Small Business " in 1993, which comes very close to the concept: http://szabo.best.vwh.net/multi_small.html . So the best of my knowledge this is the source of the concept, which might have roots in earlier works of art. The idea is bit different, and more well thought of, in that in refers specifically to the multi-jurisdiction problems, and acknowledges that cooperations are legal constructions, not just bits on the internet.

The rise of virtual nations. Multinational small businesses might speak entirely English, Japanese, Mandarin, etc. Their employees might live primarily within a single cultural milieu, dispersed thru a large number of small ethnic communities around the world, keeping close culture-specific, multimedia communications links between the communities.
[...]
Businesses will learn to share the information needed to attract investment and sales, only to those investors and customers, without jeopardizing their legal status in any major market in the maze of obscure jurisdictions they operate in. The companies that first bring these capabilities to international small business at affordable prices stand to reap large fortunes. The new paradigm of smart contracts may provide the cornerstone for building these tools.

[Mike Hearn]

"DAC" is Vitalik Buterin's rebranding of what I started calling agents:

https://en.bitcoin.it/wiki/Agents

That page contains a link to the first discussion I'm aware of around the concept, in which Gregory Maxwell described a system he called StorJ.

[coinrevo]

thanks for the info. interesting, I didn't know gmaxwell's storJ proposal. The thread is here: https://bitcointalk.org/index.php?topic=53855.0

In some ways these concepts could have only evolved after BTC.   I assume bitshares were the first attempt at constructing these, although storJ refers to something quite different.

I think a lot will be gained by understanding that cooperations are legal constructs. The article above explicitly refers to the problem of multi-jurisdictions. agency granted by law (sometimes even called artificial people), implies rights and responsibilities. Is AAPL a US company? it operates globally so it is a multi-national. but there is no entity granting rights to multi-national cooperations, besides nation to nation agreements.

[coinrevo]

Here is a recent article, quoting Mike. This uptake in the concept can be attributed to the last couple of weeks.

http://www.economist.com/blogs/babbage/2014/01/computer-corporations

Others disagree. "Lots of people will throw the term [DACs] around without really understanding it," says Mike Hearn, a Google engineer and Bitcoin developer. He prefers the term "autonomous agent" as a more useful metaphor. For such agents to exist, he says, "you need trusted computing to work well and it never has. So it'd require new hardware to be deployed."

In Mr Hearn's view, no true DAC currently exists. He speculates that the first will be some form of decentralised online storage, a kind of "distributed DropBox".

Interestingly the articles start with this:

MITT ROMNEY, the defeated American presidential candidate, once declared while campaigning that “corporations are people"

[bitfreak!]

The first time I read about the concept of DACs and the first time I heard the term DAC used was a few months ago when Invictus Innovations started developing their BitShares project.

Their website seems to have some good info on DACs:
http://invictus-innovations.com/i-dac/

And a longer article by the president of Invictus Innovations:
http://letstalkbitcoin.com/bitcoin-and-the-three-laws-of-robotics/

[Carlton Banks]

"DAC" is Vitalik Buterin's rebranding of what I started calling agents:

https://en.bitcoin.it/wiki/Agents

Neither are good descriptions of what's actually happening. It's an autonomous service, and some of those services can interact with each other. Not very sexy sounding, but at least it's not euphemistic or overblown for the sake of it. And the legislative angle is misleading. These services will re-write the law, on the basis that the resources of those enforcing the law aren't effective enough to prevent the service operating. Pirate Bay and bitcoin have literally made that happen in The Netherlands and New York state just this week.

[Mike Hearn]

Well, what I wrote is a little bit misleading for another reason - I found Vitalik's articles a little hard to follow but it sounds like he was imagining something a bit different from StorJ and julz/gmaxwell style agents. A DAC is described as a single entity spread over thousands of machines. An agent is a single instance of a program on (most likely) a single computer, which may be a node in a p2p network, and which competes in the open market against other agents and humans.

Regardless, it's cool that the Economist is picking up on such strange and fantastical ideas.

I updated the Agents page on the wiki to credit julz (re-reading the original thread, I see that he laid out the same ideas as gmaxwell) and to link to Vitalik's articles also.

[vokain]

I wanted to find out who invented the concept, but as these so called whitepapers floating around don't use scientific standards for citation I haven't been able to identify where the term comes from. google scholar doesn't return any reference on a keyword search.

https://github.com/DavidJohnstonCEO/DecentralizedApplications

There's also this, I don't know if you've read it but I think it does fall into your above category, though still informative enough to be worth reading

[CIYAM]

At this early stage I think that DAC is mostly just being used as a "buzz word" to attract investors rather than anything really groundbreaking.

For such an entity to work it would need to be 100% controlled by "math" rather than by individual whim and I don't think we've yet seen anything that is doing this (although maybe such a thing will be appearing soon).

[Carlton Banks]

Well, what I wrote is a little bit misleading for another reason - I found Vitalik's articles a little hard to follow but it sounds like he was imagining something a bit different from StorJ and julz/gmaxwell style agents. A DAC is described as a single entity spread over thousands of machines. An agent is a single instance of a program on (most likely) a single computer, which may be a node in a p2p network, and which competes in the open market against other agents and humans.

Regardless, it's cool that the Economist is picking up on such strange and fantastical ideas.

I updated the Agents page on the wiki to credit julz (re-reading the original thread, I see that he laid out the same ideas as gmaxwell) and to link to Vitalik's articles also.

Sounds good. You could argue that the DAC concept could consist of multiple agents working on differentiated tasks as assigned by a presiding agent directing the strategy, also autonomously. But it's a totally abstract idea until someone develops a working example that fulfils the concept. Still, the key takeaway is that cryptocurrency permits it to happen, the banking system isn't structured to allow that sort of thing (with the near exception of the algorithmic trading in financial markets, but they're constantly overseen with parameter changes, design changes or scrapped outright).

The Economist's angle on cryptocurrency as a whole has a curiously targeted approach. They're covering developments in the technology space pretty astutely, but anything that spills over into legal contracts, or economics itself, is peculiarly absent (or only addressed in a deflected or indirect fashion). I suspect they'll end up doing some good coverage in the end though, it's an area that's ripe for one those pull-out reports they often have.

[Mike Hearn]

The way it seems to work is that a network of freelance journalists pitch them stories and they decide what to buy. So they get a lot of techies coming to them with stories around cool stuff that made them excited, but the legal/financial side is less covered in this way.

[Sukrim]

Well, a simple pyramid scheme might already work with some simple contract (e.g. pay 1% of all holdings to everybody who contributed to the stash once a week until you run out of money).

Slightly more complex stuff (an example could be to operate an autonomous blog, where you can get paid as author depending on how popular your contribution is and how much ad money/other income it generates) is probably also possible - but once you get into even more complex scenarios defining the algorithms etc. is probably much more work than to just trusting someone to do it in meatspace.

[CIYAM]

...but once you get into even more complex scenarios defining the algorithms etc. is probably much more work than to just trusting someone to do it in meatspace.

Exactly why I think a lot of this DAC talk is about using a new "buzz word" more than using a new useful piece of technology.

[Timo Y]

...but once you get into even more complex scenarios defining the algorithms etc. is probably much more work than to just trusting someone to do it in meatspace.

Exactly why I think a lot of this DAC talk is about using a new "buzz word" more than using a new useful piece of technology.

You seem to misunderstand the purpose of DACs. Their purpose is not do take away work from humans, but to take away power from humans.  That is what differentiates them from robots/AI. To achieve this, they don't even need to be exceedingly complicated (not to be confused with complex).

Many organizations in the world are lead by "Monarch CEOs". These are essentially semi-skilled workers whose role is not so much making strategic decisions as acting as an authority figure who encourages unity once a decision has been made.  The problem with putting a human in this role is that power corrupts. A CEO will always be temped to use his power for their personal best interest rather than the interest of the organization. And even an honest CEO is prone to making irrational decisions because he is a slave to emotions like jealousy and pride.

The big innovation here is that the CEO himself can be replaced by an algorithm that is guaranteed to act objectively and in the best interest of whatever "charter" has been hard coded into the DAC.

I don't think something like fully decentralized power has ever existed. Democratic government and certain religions have attempted this, but ultimately they must always rely on some human as a power custodian. If DACs can be pulled off, they will be one of the biggest innovations in social/political organization in hundreds of years.  Buzzword or not.

PS. Daniel Suarez' novel Daemon from 2006 was the first place where I saw this idea explored in detail. Highly recommended!

[CIYAM]

The big innovation here is that the CEO himself can be replaced by an algorithm that is guaranteed to act objectively and in the best interest of whatever "charter" has been hard coded into the DAC.

It is without doubt an interesting concept but...

PS. Daniel Suarez' novel Daemon from 2006 was the first place where I saw this idea explored in detail. Highly recommended!

the question is when it will move from the world of fiction to the world of reality (am hoping sooner rather than later myself but still think it will take quite a while and if it needs to be an AI then that "quite a while" could easily be another 50 years away).

[Mike Hearn]

Actually I believe an earlier fictional exploration of the idea of automating away management is Manna, by Marshall Brain:

http://marshallbrain.com/manna1.htm
http://en.wikipedia.org/wiki/Manna_(novel)

It dates from 2003.

[CIYAM]

Hmm... well wasn't HAL 9000's role to basically manage the spaceship (and that was written in the 1960's)?

One interesting quote from the Wikipedia entry on HAL 9000 is "For example, AI pioneer Herbert A. Simon at Carnegie Mellon University, had predicted in 1965 that "machines will be capable, within twenty years, of doing any work a man can do",[19] the overarching premise being that the issue was one of computational speed (which was predicted to increase) rather than principle."

Well - we still don't have a HAL 9000 even 50 years since it was thought of (and I don't see one about to appear in the next 10 years either) so this is why I find some of this DAC stuff a little bit like the whole AI stuff of the past.

[CIYAM]

Actually I now know of at least one New York lawyer currently working on creating a "true DAC" so unlike the AI stuff I was talking about this actually seems to not only be potentially possible but also may in fact appear as early as late this year or early next year.

Whether the US (or any other) legal system will *accept* such a DAC is something that will probably take a fair bit longer.

[CIYAM]

I don't consider *every cryptocoin to be a DAC* so the platforms that would be used for implementing this are only what are referred to as "crypto 2.0" ones (such as Ethereum, Mastercoin and Nxt).

Calling any crypto currency a DAC IMO opinion would make the term DAC *pointless*.

A "corporation" needs to be able to "do the normal sorts of things that a corporation needs to do" - this requires "votes" and typically also "share classes" and even things like "boards".

[CIYAM]

You should use DAE (distributed automated entities) as the word "corporation" does *not apply* to the military or other things you are listing.

The word "corporation" is *well understood* and shouldn't be "misused" to "mean anything you want".

So to be clear - you are talking about DAEs but this topic is about DACs.