The Singularity can be the ultimate means to power. Who gets there first gets the world, maybe even the Universe, increasing their lead before the others exponentially. This attracts different entities into a “winner takes all” race – the Singularity race.
The civilization progress, if viewed as having levels for the analysis purpose, builds each of its levels upon the previous one. Thus, it related to a von Neumann machine: both are von Neumann-type processes. All such processes can be, and usually are constrained from being exponential for too long by some resource – either of its exhaustion, or of reaching the limit of its supply per an unit of time.
The progress is not an exception to this rule. In a world where competition is strong and the progress level is a key asset, the constraints that will be met first are the speed and the capacity of the engines that move the progress – the human brains. All other currently known constraints are much more distant, and will probably be easy to circumvent with the means of the progress level at the moment they are reached.
The brainpower constraints can be pushed somewhat further by better organization of the scientific process, social and economical adjustments etc. However, this degree appears insufficient to achieve the Singularity within reasonable time and investments – that needs a sufficiently developed artificial intelligence (AI). Therefore, the key to winning the Singularity race is the earliest development of AI that is able to increase its own abilities faster than human brains can do. So, the Singularity race is effectively between the entities who have the best base for developing AI, and invest the most in it.
The contenders
Currently (end-2015) the most serious contenders are some big American IT companies and some intelligence agencies. There are also several types of minor contenders.
The companies
These currently include Google, Apple, Microsoft, Yahoo, Facebook and IBM. Other company-type players might participate too silently to be noticed, or might join the race at a later stage.
Opened the Singularity race, and currently appear to lead it. They have substantial finances, powerful IT infrastructure, a lot of talent, the best Singularity visionaries (both among the personnel and the owners) and a bind to the early adopters. Develop a lot of different AI-containing products, which together cover a wide range of AI aspects. Have amassed the probably biggest cache of information, including quality knowledge and human personal data, in the IT industry – and this is a key super-AI asset. If marketing mistakes like Google Glass are avoided, they have good chances to win the race.
What might hinder Google is neglect for human values. The quiet abandoning of the principle “Don’t be evil” hasn’t gone unnoticed by the supporters, and they are who gives Google the best talents. (And an important market leverage, especially on emerging markets – and the future income of Google will depend on these.) Compared to the power of a big company, the difference caused by the grassroots support usually appears negligible. However, this evaluation is often misleading, and in a close race even a small difference might prove crucial.
Apple
Despite the late start, they appear to catch up quickly. Have far more financial resources than any other big IT company, maybe even than all of them together. Have also a huge cache of human personal data and well-chosen quality knowledge. Given that the strategy of simplifying the users interaction with gadgets naturally puts them on the AI path, Apple might become the leader within a decade. Also, they lately show taste for acquiring companies developing key AI technologies, including outside of its current interest areas, which shows increasing Singularity awareness and tendency to develop in its direction.
On the detrimental side, Apple appears to mostly lack Singularity visionaries among both personnel and owners. Also, their ownership is more distributed and diverse than that of Google, so they might find hard to redirect quickly big part of its finances into this. Additionally, their strategy binds them to the fashion followers – those who bind to the early adopters are potentially better positioned in the race, and the difference will increase with the speed of the progress.
Microsoft
They know well the value of the big data on every user and are rumored to create specialized tools in that area for intelligence agencies and biggest corporations. That puts them on the road of creating powerful intelligence aids, giving key advantage to assisted human minds. And they will not be averse to using these to their own advantage. The default setup of Windows 10, and very probably of all further Windows versions is to keep most confidential user data on Microsoft servers – this gives them data hold comparable to that of Google and Apple. The public statements of some key people in the company creates the impression that they have realized the importance of the powerful AI, and probably that of the Singularity. They also have probably the biggest software programming potential among all entities. Finally, if anyone in the IT world knows how to be close with the big players and to use their influence to its advantage, that is Microsoft.
The sale of Bing might hamper their access to keeping locally the info wealth of the Net. (But rumors say that they kept the ability to mirror a good selection of all quality knowledge there, and maybe the access to the Bing storage.) Their corporate culture is oriented to profits chase and smaller-scale power plays – this might keep them away from massive Singularity-targeting investing into AI and the internal cooperation needed to successfully engage such a huge task. Some of the early adopters and some IT personnel also tend to mistrust them, and this might silently hinder their efforts.
Yahoo
Currently they lag behind, but have good info search / access infrastructure and technologies. Add an ambitious boss with experience from other Singularity contenders, and some good financing might position them well in the race. If they go to sharing know-how and effort with NSA and/or the Pentagon, they might get that financing and some key technologies.
On the minus side, currently (end-2015) their chances are mostly hypothetical. The publicly available info does not show an AI development of a scale anywhere near that of the other company-type contenders. Also, they are under strong pressure from their owners to improve the short and mid-term profits, which can limit the investment into a long-term projects like the AI / Singularity race. Even if they succeed to turn to profit again, they might be too late for the race.
When it comes to the progress, the humankind can be viewed as a single intellect, and the links between the people are not less important than these between the neurons in a brain. Facebook has more info about the inter-human links than anyone else, and thus is the best positioned to solve some key tasks in the development of a powerful AI. Their connections, influence and money can buy them most, if not all of the remaining key pieces of the puzzle. So it is possible that they will appear to lag behind the race leaders for a long time, but can suddenly spurt ahead on the stage of powerful AI development. They also show some of the marks of silent involvement into mass-scale information collecting – for example, their project to supply to big parts of the world free Internet, which will pass through channels controlled by them.
Few things can hold Facebook back. The apparent lack of Singularity / AI vision, even if real, will not last for long. None can replace them at the social networks throne. Still, if they manage to underestimate enough the importance of strong AI development, the race might pull too far ahead of them to catch up.
IBM
Their research on prospective hardware gives unique advantage over the mostly software-oriented contenders. Also, they have good software engineers and a leading role in the big servers market. Thus, they are an important wild card to watch. Powerful AI needs powerful hardware to run on, and they are the company for it. Having big experience with using IP rights, they are also more than able to successfully hamper the competition. Finally, does the word “Watson” ring a bell?
Not having some key AI prerequisites themselves (eg. big data cache) appears to relegate them to the role of a kingmaker. However, kingmakers often can easily put on the throne themselves, and Microsoft taught this lesson to none other than IBM. Also, they have invested for two decades in some key AI technologies, and have the best achievements there. So it would be prudent to assume that IBM is one of the most serious contenders.
The security agencies
These currently include the security agencies of China and Russia, NSA, the Pentagon, and the agencies of Great Britain and probably Germany and France, separately or in a joint initiative. Other agencies might be also in the game, or to join it at a later stage.
China
It is smart enough to realize the powerful AI importance, and is able to concentrate into its development money and human numbers that dwarf anyone else – maybe everyone else together. Its technological prowess in the supercomputers area is indisputable. Add to that its huge military-based spying network, and it might be the most powerful contender of all.
The hurdles at this race are the typical ones for an authoritarian / totalitarian country. It is not hospitable to disruptive ventures and personal initiative, and these are key parts of the technological advancement. (However, military goals are an exception to a big degree.) Also, it is close to exhausting the possibilities for extensive growth and having to switch to intensive one instead, and that is often a problem for authoritarian countries. (Its “communist capitalism” semi-feudal system might solve this to a degree, but is still far from the productivity a truly free system can unfold with the time.) Not solving this problem satisfactorily might limit the finances and the human potential China can invest in the Singularity race. (But a period of economical struggles might turn it to the idea for the Singularity as a solution for both economical and political problems.)
Russia
It is also not averse to spying and never saved money on military goals. Has a lot of talent and its culture is welcoming to technological advancement. (Especially in the military area, where they keep abreast with the US in a surprising number of areas, despite the much smaller investments.) Add to this a very strong and strategically minded ruler, and they might be also a top contender.
In addition to the typical authoritarian country hurdles, Russia might become cash-strapped in the near future. Shales are abundant across the world and the shale oil / natural gas technologies are relatively cheap and easy to master. So the hydrocarbons price probably will not stabilize above $60 per oil barrel for at least a decade, and Russia exports mostly oil and natural gas and imports a lot of the goods it consumes – its future income might be severely limited. Still, much like China, the power there might see the powerful AI and the Singularity as the means to everything they might dream of.
NSA and the Pentagon
If Russia and China are in a “winner takes all” race, NSA and the Pentagon are bound to be in, too. Their financing might be relatively smaller, but they have access to the top technologies, a lot of the top talent and the top e-spying in existence. Their start might be later, but they are able to quickly catch up on need. And they might be able to convince the company-type contenders, all of these being American, to cooperate.
The downside is that if they create a powerful AI, it will be classified for a long time, and directed exclusively to military and intelligence goals during this period. (In some scenarios, all American AI achievements might be classified too, and take the same path.) This will effectively delay the advent on the Singularity, maybe for decades. Thus it will increase the risk that malicious and/or clandestine contenders might spy out, steal or develop first Singularity-capable AI technologies.
Great Britain
It will probably try to stick with the American efforts, building upon “Five Eyes” and other joint security efforts. However, if that turns impossible or unsatisfactory, it might decide to go alone. Has an excellent tradition in comparable projects and, through the British Commonwealth, access to plenty of talent. And its economics appears generally more resilient to crises than the average – that is, the financing of such an initiative might be smaller, but will be more reliable.
On the minus side, its resources are generally smaller than those of most state / intelligence agency players. Also, this will almost surely be a military / intelligence project, and will be classified and directed to the corresponding goals, with the effects described above.
France and Germany
Both are quickly becoming powerful AI and Singularity smart. Both are pissed by the American big IT companies and intelligence. Both will hate to see a powerful AI- or Singularity-enabled Russia or China. Germany has tons of money, organizational experience and talent, France – more talent and security experience. So a classified joint European Singularity-related AI venture is possible, if not even likely. Even if they decide to go separately, they will probably join efforts at a later point.
The downsides are similar to these of the described NSA, Pentagon or Great Britain projects.
Other contenders
The biggest retail companies
Most retail companies invest heavily in data mining and statistics, and have done it for more than a decade. These areas are closely related to some key AI technologies, and are often numbered among them. Verbose information on a person’s purchases has proved to supply a lot of insight into that person; such an info on many people gives a lot of info about the society. If augmented with a big cache of other personal data, it can be used to build powerful market research tools. The biggest retailers can easily buy such a cache, and often do.
The further development of market research tools naturally turns them into intellect-augmenting (IA) tools, which can substantially increase the profits. This tendency can easily cross the border with the assisted brain type AI development. A stronger Singularity awareness might drive it further, into the powerful AI development. (The retailers owners might be reluctant, but managers of big corporations have a lot of executive privilege. Approaching the Singularity and the creation of powerful AI increases the awareness of what these might bring, so the likelihood of such a development increases with the time. Also, the retailers’ financial power and strength of interest into increasing profits can create very lucrative market for powerful AI tools, able to finance their quick development.)
The data collection and resale companies
As mentioned above, big caches of personal data are a good base for developing powerful IA tools, and the personal data market constantly demands such tools. So, the data collection and resale companies try to develop and/or order them from IT companies. This market is no-nonsense one: even a small power advantage is quickly noticed and attracts the buyers. Such a relentless pressure for more power can elevate the products into the area of assisted brain type AI development, and then further into the powerful AI. This market, though usually hidden from the public view, contains enough money to finance that.
And it might be a dangerous development. The collection and trade of personal data is unethical and often not completely legal. This makes this market attractive to entities willing to violate the ethics and the law, and experienced in this. Such entities might not be the best custodians for a Singularity-level AI power.
The social networks companies
These companies hold caches of personal data, rivaled only by the biggest Internet search companies. (Most of the latter, being among the “big” AI contenders, are discussed above.) It is inevitable that they will be drawn into the personal data market, dealing and/or competing with the data collection and resale companies, and will eventually take the same road. And most of them have enough in-house tech skill to try and develop AI.
The problem here is that the social networks user “market” is very strongly aggregating one. Facebook has cornered most of it; the other players are niche ones, and don’t have nearly the same amount of personal data. That is, their base for developing powerful AI is narrower – using it for that will require bigger investments than a Facebook or a Google will need to do the same. Still however, any of the bigger among these companies might provide another race contender with key data and technical expertise in the personal networking area, and thus to unexpectedly strengthen that contender’s position.
Bruce's Blog 