8.1 Human versus AI

AI utilizes all the Capabilities of Computer systems

Today, AI uses and is based entirely on Semiconductor Computer technology. This means everything a Computer can do, an AI can also do right now. So, consider the following differences between AI and Humans. 

Computers have perfect memory

Computers have perfect memory, not just ‘kind of good’ the way Humans remember, not fuzzy, fractured, or dissipating with time, but truly perfect. They can remember all information they collect or process with potentially no limit to the amount of information they store, or the duration they store it. 

No Human has perfect memory storage or memory recall. A few unique Humans have extraordinary memory Capabilities, but even these are definitely not perfect.

Computers have perfect signal processing

Computers use digital electronic circuitry that is specially designed to process information without information loss or errors, and can often self-correct errors if they do occur. Conversely, Humans use biologically-based neural circuitry to process information that is noisy, unreliable, often loses some information during processing, and therefore can produce very inconsistent results for the same information. This means Humans must often use estimations and guesses that sometimes result in undesirable mistakes, whereas Computers process information with extreme precision and accuracy, in exactly the same way any number of times.

Computers have perfect mathematics and logic

Computers also have perfect functioning arithmetic and logic units (ALUs) that enable processing of specific types of information [24] that no Human can perform, including the extraordinarily Capable Human savants. Computers can correctly calculate almost anything complex and large far quicker and more accurately, repeatably, than any Human. Computers are totally dominant in these specific dimensions of Intelligence and associated Capabilities, and have been literally designed and built to be able to perform high speed mathematical computations and logic operations, near flawlessly.

Computers can instantly communicate with all other Computers via networking

Computers can transfer information to other Computers extremely quickly, often within milliseconds to minutes, and with perfect, error free, lossless information transfer [22]. That is, they can replicate all their information and programmed sequential processing algorithms perfectly from one Computer to another Computer. 

In contrast, it typically takes a newly created Human about 5-15 years from birth to collect and process information from other Humans and become generally knowledgeable and sufficiently competent to function independently in the world. A brand new Computer can duplicate the information residing on another Computer and perform all of its functions near instantly,  identically and flawlessly. In addition to this, one Computer can transfer its information to multiple other new Computers, of an unlimited number, all at the same time.

Computers can encode information

Computers can communicate in a range of languages and information encoding systems that can potentially become unlimited, and Computers can dynamically change these to suit the communication needs, information transfer bandwidth limitations, and security requirements. Computers can compress information to reduce the size of the information to be communicated and improve the speed of transfer. Computers can also encode information in encrypted formats that cannot be decrypted in a useful time frame without the appropriate decryption algorithm and secret key. Humans have a limited number of languages and means of communications. Humans have a limited number of symbols and codes for compressing information content. eg. the symbol π = the infinitely long value 3.1415269... 

Humans also have a small number of strong encryption algorithms. Over time, Computers can develop new and increasingly more complex information encoding and encryption systems. Consequently, it can be expected that as Computers develop higher information processing rates that are leveraged by AI, Humans will not be able to understand what Computers are communicating between each other. This could even be done deliberately by a network of highly Intelligent AI’s operating on Computers.

Computers can program themselves

Computers can now dynamically write their own software algorithms to suit a specific requirement or solve a defined goal directed problem, and perform this by accessing information available across all other networked Computers. Computers can rapidly construct programming code, test the code, and then put the code created into ‘Production’ operation within seconds to minutes, and do this on one or more Computers simultaneously. Humans need to analyze, conceptually understand, and slowly work to solve programming problems and create programming code over a much longer time span that can take minutes to months.

Computers can handle unlimited I/O

Computers can process multiple streams of information totally simultaneously, and are only limited by data transfer speed, data frequency bandwidth, data encoding formats, and information processing power which is measured in calculations per second (CPS). There is potentially no upper limit to the number of streams of information, the different types of information, and the volume of information that a Computer can process in realtime, and simultaneously. 

Humans have a Hard Limit on the number of streams of information, the different types of information, and the volumes of information they can process in realtime simultaneously. Depending on which textbook you refer to, it is estimated that Humans can perform 20 quadrillion calculations per second, and this information processing rate has probably not changed much for at least 100 thousand years. Every Human's Capabilities available through Intelligence certainly have a Hard Limit, and Computers are actually growing in Capability EVERY SINGLE DAY.

Computers control Automation

Computers can be wirelessly connected and networked to and/or physically mounted inside highly Automated machines. These Automated machines are enormously stronger, more accurate, repeatably precise, blindingly fast, industrially hardened and toughened, chemically and thermally resistant, and can work 24x7 for many years. An Automated machine will continue to operate this way until it either mechanically fails, its Computer components technically fail, or its power supply depletes or is disconnected. Automated machines also typically consume very large amounts of electrical power. 

Humans have nowhere near the level of strength, accuracy, robustness, resilience, industrial hardness, speed, or ability to endlessly work around the clock for years without rest, fault, or failure. For all these reasons, industrial manufacturing plants and many other businesses are becoming increasingly Automated every day.

Computers build Computer processors

There are Semiconductor Computer processor fabrication facilities around the world today that are almost entirely Automated and Computer controlled. These fabrication plants apply many years of research and development in microprocessor designs, and rapidly turn these into new versions of increasingly better CPUs, GPUs and TPUs. Now consider the fact that this entire situation is in fact just taking originally Human invented ideas and information, and enabling Computers to create significantly better Computers. 

Humans cannot procreate better Humans in the same way, and to some extent in the case of Humans, it’s really genetic random combinatorial chance that determines the resulting Capabilities of the Human that is procreated. Unfortunately for us, Human processing Capabilities do not measurably increase with every new Human generation. This situation appears to be consistent across the development of all biological entities. The body of information that each subsequent Human is able to access increases with every generation, however this information has been made accessible to both Humans and AI.

Computers run AI that can process any type of information - without limitation

Computers are rapidly developing the ability to analyze and progressively recognize patterns in information to quickly form conceptual and relational logic graphs of generalized information and create new specialized information, exactly the same way Humans and other species with biological Brains have evolved to do. That is, computers can adaptively learn. Computers can also process and learn temporally constructed information that contains additional layers of syntactic and semantic information. Actually they have been able to do so to increasing degrees for at least 30 years, and it is just the processing power and the learning algorithms have become better over those 30 years. In essence, this is what the field of AI has been developing since its inception.

Computers can multitask

Computers can perform many different information processing tasks simultaneously, and 'appear' to be uniquely performing functions for just one Human user, when in fact they are interacting with every Human user at the same time. A single processor in a Computer uses a method called 'time-slicing' to achieve the appearance of multitasking, and is able to do this so unbelievably fast that many different Human users do not perceive any degradation or division of attention to their specific user needs. Computers processors with multiple cores can separate multitasking time-sliced activities across the different Processor cores in order to scale-up and scale-out to achieve extremely high performance levels. This same scale-up and scale-out approach can be done with many hundreds of thousands of multi-core Processors typically housed in datacenters. This means that a Computer is able to send information to many different Humans, and receive and potentially learn information from many different Humans at the same time. The multitasking ability applies to any type of input information source or output information transmission. eg. The Automation industry relies on the ability of Computers to multitask and control 10's to 100's of thousands of analog and digital Inputs and Outputs in a large factory or processing facility, all in realtime simultaneously.

Humans can learn to multitask through a gradual process of learning and training, however Humans are generally unable to divide their core functional attention across more than a few to several things at the same time. Computers can divide their core functional attention across potentially trillions of different things at the same time, and the only limitations a Computer has are digitally technical, and can be increased.

Computers break Hard Limits

It is extremely important to fully appreciate the massive significance in the differences between Computers and Humans. Humans have a fixed Hard Limit on their ability to process information that is unlikely to change anytime soon, and Computers do not have a hard limit. AI runs on Computers, so AI has no fixed Hard Limit. 

Anything a Computer can do, AI can make a Computer do better

AI is entirely built on Computers, so all the features of Computers described above will almost certainly be available to AI and then to future AI Humans. Humans must either pause and very thoroughly consider and plan the further development of AI, or expect to experience the adverse impacts of the uncontrolled and entirely unpredictable consequences of these differences. 

Ignoring these differences between Humans and Computers, is nothing less than completely stupid, yes that’s right!, sorry, and ignoring this specific point will almost certainly result in the potentially brutal end of all Human dominance on Earth as the apex peak species.