I have been thinking about this exact thing for a while now.
With Decentralized Autonomous Organizations, the resource utilization of a business is simply an algorithm, which can use any number of APIs to return profit and grow the entity.
All of these APIs we are creating for various functions may soon replace us. Many APIs already do directly dispatch humans to do certain tasks. The scary part is that when someone is providing a service to you, they may not be aware of whether their superior is a human or an algorithm.
A minimum viable autonomous corporation could be a lemonade stand that is managed entirely by an algorithm; with virtual currency it can hire a person to set up a lemonade stand at a certain datetime, and take a cut of earnings. If sales are slow, it could hire someone to design a poster, and another to post them in the area; or even an analyst to decide what the best course of action is. Over time, with many locations and employees, it would be able to learn the best locations and conditions for optimal profit. Perhaps also deciding which employees to continue working with and which ones to fire.
These "management algorithms" would likely need to be developed by someone. But what if the algorithm hired the person to write the algorithms, updates, and performance tweaks?
There could be a bootstrap algorithm that is simply an idea (these could be automatically generated based on search trends, market fluctuations, etc). These seeds would be open to the public for investment with virtual currency (bitcoin could be used to allocate the cap table as well), the program would be able to hire talent to build out the idea, and the investors would receive dividends.
I see this evolving as a kickstarter for viral autonomous businesses. And I don't see why it couldn't be viable with today's technologies and APIs.
The 'Daemon' dream falls down in the fact of fraud and/or entropy. Ask PayPal about their 100% automated fraud detection algorithm (did not work) or Google about driving on umapped roads or through snow.
Will humans on the "receiving" side of an API remain unaugmented, i.e. operating at human speeds with a slow touchscreen interface? Or will they develop their own automation, so that millisecond-speed algorithms broker between the "manager" and "worker"?
How much ad and click revenue today is actually human? Bookstores did not sit idly by when Amazon decided to take on the entire industry, how is Amazon's publishing imprint doing? Humans do eventually wake up and compete with algorithms.
What a great book. I love it because even though everything in it is currently future-tech, almost everything seems almost within reach. It all seems so plausible.
"My opinions have shifted in the past 10-15 years and I now think that permitting autonomous corporations to exist -- or even continuing the doctrine of corporate personhood -- are a terrible idea for humanity (because they're effectively AIs that compete with us, and not in a good way)."
Rats! I was going to make the Accelerando connection too! While we are on the subject of being outcompeted, Saturn's Children is also a pretty good read.
"A society that runs on robot slaves who are, nevertheless, intelligent by virtue of having a human neural connectome for a brain, is a slave society ... The human society underpinning "Saturn's Children" got into bad trouble, relying on robot slaves as labour and disappearing up its own arse in the pursuit of virtual luxury."
Human-level performance in this domain will probably take more than 15 years to be created, but I'm pretty confident it will be done. We did solve chess, after all.
Nitpick: We solved nim. We created an algorithm that, when applied, guarantees the first player a win if some conditions are met, and the second player a win otherwise.
There is no equivalent for chess. Instead, we created programs that can apply more general game-theoretic algorithms with ad hoc modifications for chess (for example, minimax with alpha-beta pruning and lookup tables to recognize book openings, closings, and common scenarios) which can, at current CPU speeds and RAM capacities, reliably beat the top-ranked human players.
It is mathematically possible to beat the best computer chess player; it might not, however, be humanly possible. OTOH, it is mathematically impossible to beat a simple nim program which has been programmed with the perfect play algorithm, assuming the game is such that the program will win if it applies that algorithm.
My point is, if something is solved, an adversary throwing more computing power at the process is meaningless. It would merely hasten the inevitable. However, for a problem like chess, an adversary suddenly getting better computers could be a game changer.
Super excited for the potential of DAOs to reshape the world.
Something I've been thinking about with any blockchain technology is that it will likely end up sitting on top of the current Internet platforms, but still behind the scenes and hidden from both enterprise and consumers.
If so, will new blockchain-based startups disrupt industries and create their own competing consumer services, or will the incentive to dramatically cut costs (labour mostly) push existing industries to transition their operations to run on a blockchain?
eg the average user should never come across the word "bitcoin", they just continue to transact and pay with "dollars" as they already do. But behind the scenes, the entire financial system operates on a blockchain.
eg2 existing companies looking to cut costs could automate much of their staff by transitioning their core database and systems to run on a blockchain. However all customer-facing UI and interactions would operate as normal. A company like Uber could run entirely on the blockchain as a DAO, and have a small team responsible for maintaining the consumer-facing apps
see peertracks.com for an example of what you are describing. its launching march 2015. Users will buy songs in usd and the usd will be turned into noteUSD on the blockchain. All song sales and streams etc will also be tracked on the blockchain. The frontend that interacts with consumers will be run traditionally through a private company etc.
> All of these APIs we are creating for various functions may soon replace us.
Good. Then we can get that promise of more leisure time that's been hanging around since the 50s.
Imagine if everyone had more time to learn another language, play guitar and ponder the meaning of life. APIs and robots can grow my food and wash my clothes for me, I have better things to be doing.
The promise have been around since the 1840's or before. The origin of socialism was the idea that the rise of industrialisation would increase efficiency to the point that scarcity would eventually become a thing of the past - at least for everything needed to meet normal human needs. People like Saint Simon, Fourier, Owens, Marx, Proudhon and many others started with those assumptions.
Large parts of Marx entire theory of how socialist revolutions would come about is based around the combination of this idea and the idea that the capitalists won't willingly grant it.
>scarcity would eventually become a thing of the past - at least for everything needed to meet normal human needs.
We're really there aren't we. Almost everyone in the industrialised parts of the world have access to food, shelter and other basic human needs. Damn, in Western Europe even the romanian beggars have iPhones.
> APIs and robots can grow my food and wash my clothes for me, I have better things to be doing.
No, they're do that for their owners, the capital holders; you'll be left out in the cold to starve. This is a labor based world, no labor, no share of the pie.
Contract work on modular functions, not a monolithic code base. Spawn a child, apply contractor's code, watch what happens. It would be a mixture of Genetic programming (fitness tests) and intelligent design :)
Up to a point; often the module boundaries are wrong (I've worked in a company that failed mostly for this reason - team A worked hard on module A, team B worked hard on module B, but neither was able to challenge the insanity of calls that had to ping-pong back and forth between A and B several times over to get a result). And you can't touch the genetic programming structure itself; the program can optimize itself, but it can't optimize its optimizer, and ultimately it may be outcompeted by programs with better optimizers.
In traditional capitalism the buck stops with the board, and the shareholders will fire them if they're losing too much money (something that's become more important in recent years, even as more and more day-to-day stock trading is algorithmic). Who takes that role in the case of an autonomous algorithm - would the algorithm itself be a director?
if the NSA could hire people to do it's algorithmic work and yet not have them know what they are "truly" working on, i m sure you can do it for a business.
When I think of a minimum viable autonomous corporation I think of a bot that crawls twitter for trending slogans; puts the trending slogans on shirts; sell the shirts on Amazon; place manufacturer orders for shirts only after they sell on Amazon. It can branch out in to other merchandise from there.
With Decentralized Autonomous Organizations, the resource utilization of a business is simply an algorithm, which can use any number of APIs to return profit and grow the entity.
All of these APIs we are creating for various functions may soon replace us. Many APIs already do directly dispatch humans to do certain tasks. The scary part is that when someone is providing a service to you, they may not be aware of whether their superior is a human or an algorithm.
A minimum viable autonomous corporation could be a lemonade stand that is managed entirely by an algorithm; with virtual currency it can hire a person to set up a lemonade stand at a certain datetime, and take a cut of earnings. If sales are slow, it could hire someone to design a poster, and another to post them in the area; or even an analyst to decide what the best course of action is. Over time, with many locations and employees, it would be able to learn the best locations and conditions for optimal profit. Perhaps also deciding which employees to continue working with and which ones to fire.
These "management algorithms" would likely need to be developed by someone. But what if the algorithm hired the person to write the algorithms, updates, and performance tweaks?
There could be a bootstrap algorithm that is simply an idea (these could be automatically generated based on search trends, market fluctuations, etc). These seeds would be open to the public for investment with virtual currency (bitcoin could be used to allocate the cap table as well), the program would be able to hire talent to build out the idea, and the investors would receive dividends.
I see this evolving as a kickstarter for viral autonomous businesses. And I don't see why it couldn't be viable with today's technologies and APIs.