0 Members and 1 Guest are viewing this topic.
Quote from: hamdani yusuf on 21/03/2021 07:36:55A commonly cited thought experiment to describe orthogonality thesis is a superintelligent machine whose terminal goal is to produce paper clips as many as possible. It's supposed to show that intelligence and terminal goal can be independent to each other.Nonetheless, the survival rate of conscious systems depend on the alignment between their terminal goal and their own survival. The survival rate is at lowest point when the terminal goal is diametrically opposed to their survival. It's highest when the terminal goal is perfectly aligned to their survival, which means that their own survival is set as their terminal goal. Everything else is in between those two extremes.In the case of paper clip maker superintelligent machine, the terminal goal is clearly not its own survival, but it's not diametrically opposed either, hence it lies in between. It means that it cannot be the most efficient system possible to survive, which means that it is in a disadvantaged position when it has to compete with other conscious systems with similar superintelligence, but less burdens unrelated to their survival.
A commonly cited thought experiment to describe orthogonality thesis is a superintelligent machine whose terminal goal is to produce paper clips as many as possible. It's supposed to show that intelligence and terminal goal can be independent to each other.
Though human brains may have adequate resources to simulate some parts of the universe, their existences depend on other organs forming the human individuals. Hence their expressions represent the individuals as a whole, not merely the brain as an organ.
I also mentioned about minimum requirements for a system to be called conscious. It must have parts serving the function of virtualization of objective reality , including its own representation in its virtual world, which we often call self awareness. To do that, it needs some sensing mechanisms, some memory to store the results and convert them into its internal model. It must also have actuation function, which gives it access to modify or manipulate its real world environment.
Quote from: hamdani yusuf on 27/03/2021 12:21:33Though human brains may have adequate resources to simulate some parts of the universe, their existences depend on other organs forming the human individuals. Hence their expressions represent the individuals as a whole, not merely the brain as an organ.The universal superorganism consciousness will execute most of its information processings in supercomputers, similar to currently existing e-government and IT infrastructures of large corporations. They will likely involved more in strategic thinking and long term decision makings, while shorter term decisions will be left to computing tools in lower hierarchies and edge computers.In the past, thinking process at all hierarchical levels in superorganisms from village to imperium were done by human brains through their rulers or representatives at each level.At some point in the future, human individuals, at least in current form, would be seen more as burdens rather than tools. That's why we would need to improve ourselves. There are many ways to do that, such as gene editing, epigenetics, nanotechnology, direct brain interface, exoskeleton, etc.
The Future of Humankind with Yuval Harari. What is the next stage of human evolution? How will we protect this fragile planet and humankind itself from our own destructive powers? Professor and author Yuval Harari envisions our future: a not-too-distant world in which we face a new set of challenges and possibilities. With his trademark blend of science, history, philosophy and every discipline in between, Harari investigates the projects, dreams and nightmares that will shape the twenty-first century.
Some dystopian stories cast fear by imagining that the supercomputer equipped with AGI would detach itself from humanity and becomes an independent conscious entity. It's like other cells in a human body become fearfull that the brain would detach itself from the body and becomes an independent conscious entity once it gets smarter.At a glance, it may sound absurd. But it's not completely impossible either. Once the brain gets access to modify the body at will, it is likely that it will be done.Humans are known to cut their hair and nails for a long time. Some of them get circumcised. Some get amputated due to accident or cancer. Some have replaced their hearts. Some get lasik to fix their vision. A few had half of their brains removed.
We are on the cusp of a major disruption in how we feed ourselves. This video is a quick summary of a report from RethinkX on where agriculture is headed over the next decade, and it's mind blowing!
Quote from: hamdani yusuf on 27/03/2021 12:21:33Though human brains may have adequate resources to simulate some parts of the universe, their existences depend on other organs forming the human individuals. Hence their expressions represent the individuals as a whole, not merely the brain as an organ.In ancient kingdoms and empires, the virtualizations happened in their documentation and administration systems, which can take forms of clay tablets or writings on paper. In modern organizations, they take place in computers. They are more flexible in the mechanisms, such as mechanical, vacuum tubes, electronic semiconductors, or optical computers.They virtualize taxes, budgetings, assets, plans, supply chains, etc.In biological systems, the virtualizations are usually formed by neural networks.
How can many stupid things combine to form smart things? How can proteins become living cells? How become lots of ants a colony? What is emergence?
In early human civilizations as superorganisms, external data storages came into existence as an alternative for human brains which are typically less reliable as long term memory. Painting on cave walls and clay tablets are some examples. More practical, and more capacity data storage evolved in the form of writings on paper, printing press, microfilms. Invention of computer requires better version of data storage more suited to digital information. They started with punched cards, then magnetic tapes/discs, optical discs, and solid state drives. With the advancement of telecommunication through Internet, cloud based data servers become more feasible. Currently, semiconductor-based memories seem to outperform biological neurons in many categories. The reason why they were not employed by natural biology is likely because they don't readily self duplicate, and the process to produce them is too long and complex in biological standard. As a superorganism, human civilization has found a better functionality of data storage in semiconductor based memories. It doesn't matter if they don't self duplicate since they can be produced through mass production in chip factories. This is another example of specialization process at work.
IBM has come up with an automatic debating system called Project Debater that researches a topic, presents an argument, listens to a human rebuttal and formulates its own rebuttal. But does it pass the Turing test? Or does the Turing test matter anymore?The Turing test was first introduced in 1950, often cited as year-one for AI research. It asks, “Can machines think?”. Today we’re more interested in machines that can intelligently make restaurant recommendations, drive our car along the tedious highway to and from work, or identify the surprising looking flower we just stumbled upon. These all fit the definition of AI as a machine that can perform a task normally requiring the intelligence of a human. Though as you’ll see below, Turing’s test wasn’t even for intelligence or even for thinking, but rather to determine a test subject’s sex.
Does it matter if any of today’s AIs can pass the Turing test? That’s most often not the goal. Most AIs end up as marketed products, even the ones that don’t start out that way. After all, eventually someone has to pay for the research. As long as they do the job then it doesn’t matter.IBM’s goal for Project Debater is to produce persuasive arguments and make well informed decisions free of personal bias, a useful tool to sell to businesses and governments. Tesla’s goal for its AI is to drive vehicles. Chatbots abound for handling specific phone and online requests. All of them do something normally requiring the intelligence of a human with varying degrees of success. The test that matters then is whether or not they do their tasks well enough for people to pay for them.Maybe asking if a machine can think, or even if it can pass for a human, isn’t really relevant. The ways we’re using them require only that they can complete their tasks. Sometimes this can require “human-like” behavior, but most often not. If we’re not using AI to trick people anyway, is the Turing test still relevant?
So far, I've seen that progress of continuous improvement of organized information system can be classified into two types, generalization and specialization. Perhaps it's comparable to bulking and cutting process in body building.Generalization works by expanding functionality of existing components of the system. This concept emphasizes on effectiveness over efficiency. Specialization works by removing unnecessary capability of components which are not related to their main function in a system. This concept emphasizes on increasing efficiency while maintaining effectiveness.
Researchers from Chalmers University of Technology have produced a structural battery that performs ten times better than all previous versions. It contains carbon fiber that serves simultaneously as an electrode, conductor, and load-bearing material. Their latest research breakthrough paves the way for essentially ’massless’ energy storage in vehicles and other technology.
Neuralink makes brain implants that it hopes can eventually be used to give people with quadriplegia the ability to control computers and other devices using only their minds. In the future, the company says, healthy people might do the same. Someday, this could conceivably eliminate the need for keyboards, speech-to-text, and thumb typing on phones.The company just took a big step toward that future. It implanted two of its devices into the brain of a nine-year-old macaque named Pager and then taught him to move a computer cursor--and to play Pong--using a joystick. (Pager likes to play because when he gets things right, he's rewarded with banana smoothie delivered through a metal tube.)As Pager played, the Neuralink devices recorded the signals in his brain that told his hand to move the joystick up, down, left, or right. The company's software learned to interpret those brain signals as movements, and then sent those movements directly to the computer, bypassing the joystick. Soon, Pager was able to move the cursor, and then play Pong, using just his brain. And he played really well. Despite the researchers speeding up the game to test his abilities, Pager only loses one point during the video, which is appropriately titled "Monkey MindPong."
On the other hand, we also have specialization step. For example, each neuron in human brain contains complete set of genes whose function are unrelated to information processing. Artificial neurons can be designed based on natural neurons with unnecessary parts removed.
For generations, the standard way to learn how to ride a bicycle was with training wheels or a tricycle. But in recent years, many parents have opted to train their kids with balance bikes, pedalless two-wheelers that enable children to develop the coordination needed for bicycling—a skill that is not as easily acquired with an extra set of wheels.Given the benefits of balance bikes, why did it take so long for them to replace training wheels? There are plenty of other examples in which overlooked solutions that involve subtraction turn out to be better alternatives. In some European cities, for example, urban planners have gotten rid of traffic lights and road signs to make streets safer—an idea that runs counter to conventional traffic design.
To determine why people tended to choose additive solutions, the team dug deeper by conducting a series of eight experiments with more than 1,500 individuals recruited either from a university campus or through Amazon Mechanical Turk, a crowdsourcing Web site. In one experiment, people were asked to stabilize the roof of a Lego structure held up by a single block that rested atop a cube-shaped base. The reward for completing the task was $1, and participants could add new blocks for 10 cents apiece or get rid of blocks for free. The researchers wrote that one group was provided a cue about potential subtractive solutions by being told, “Each piece that you add costs ten cents but removing pieces is free,” while another group was just told, “Each piece that you add costs ten cents.” Almost two thirds of people in the cued group ended up choosing to eliminate the single block rather than adding new ones, compared with 41 percent of those who had not received the prompt.