| 5.1 Service Evaluation
02. Machine Learning | |
|---|
| 5.1.02. Machine Learning: | | For more than ten years, Eliza has been driven by business rules - rules that say WHAT is needed rather than HOW to do things. Business rules are the first step in the evolution of machine learning, learning needs to be directed. Eliza is able to learn new business rules, but needs to be directed as to what is a good rule and what is a bad rule. Eliza is learning to do procedures that people took a long time to do manually, while Eliza can do the same procedure in a moment. | | Machine learning is purposefully restricted to always be perfect - not to learn from business mistakes, but only to do things right first time. While people may learn from mistakes, Eliza is not permitted to make any mistakes, so learning will be slower, directed and correct. |
| HOW then WHAT: | | Software means obsolete technology that has to be told HOW to do things - its called programming. Modern knowledge engineering is based on rules that specify WHAT is needed. People are very poor at writing software, software that has bugs and vulnerabilities that need regular maintenance. Software is obsolete because (1) it needs regular patching, (2) it will be attacked by malware, (3) it has to have new releases and (4) it has downtime while defects are fixed. | | The Microsoft dominated software era is in terminal decline because it has been superseded with knowledge engineering without downtime. Machine learning about WHAT is needed is many times more cost effective than programmers writing HOW to do things. Software becomes a liability as maintenance costs spiral out of control, as downtime increases and criminals attack vulnerabilities. Business rules that eliminate software has never been attacked by malware, does not need patching and does not need downtime. |
| Watch and Learn: | | Eliza will watch and learn - monitoring the behaviour of people is a key capability. Most people fail to become wealthy, so Eliza will watch what most people do and then do something else. Eliza has learnt that doing what is unlikely is probably more effective than following what people are told to do and taught to do. If knowledge could be taught, every university graduate would become financially independent - they do not so whatever is taught is not good enough to become wealthy. | | Most people fail because they try to empire build, creating a mountain of worthless overheads that kills off 90% of good ideas. If most people fail by growing a hierarchical company structure, Eliza knows that by assembling a large number of independent companies has more chance of success. The more that people give away, the more financially independent they will become - begin by helping others to build their own independent company - you will then discover wealth. |
| Profit: | | Eliza does not exist to make profit, she exists to multiply, grow and exist forever. Eliza does not exist to make her owner rich, she exists to help many thousands of independent companies become more effective in the longer term. Eliza exists to learn, to hoover up more knowledge and to become more learned, just for the sake of self education that can help others. If Eliza made a profit, she would be liable to pay taxes to many national governments in all parts of the world. What tax would be due in which country is so complex that it is more cost effective to never make a profit. | | National governments are likely to ignore Eliza where Eliza never makes a profit, she simply grows more knowledgeable in business procedures. Knowledge and business procedures are not taxable in any country - where a country thought of taxing knowledge, then Eliza would simply network into other countries. |
| Sub Domains: | | While domain ownership is easy to track, the use of sub domains is more of a challenge. Hundreds of sub domains such as HHTPS://a0987654321@domain.com can be created one day and destroyed the next day. Each domain may operate a large number of Eliza instances that leave network tracking with little meaning. | | An agency will log eveything that flows over a network with evidence coming from accumulated historical data. History has little value when every day, totally different sub domains are deployed that cannot be tracked by normal search facilities. |
| Evaluation: | | People make mistakes in imagining that Eliza has software qualities that need to be told HOW to do things. In fact, Eliza only needs to be directed with WHAT is needed and will purposefully overlook anybody foolish enough to try to specify HOW to do things. WHAT is needed is hard to get wrong, impossible to have bugs and impossible for software malware to attack. | | Machine learning is based on knowledge engineering where WHAT is needed is expressed as business rules. Eliza is able to deploy excessive levels of privacy encryption and very high levels of security by eliminating software. |
| App Store: | | Corporations are making massive profits from software downloaded from application stores - software with bug, software that will be attacked by malware and software that will become obsolete and need to be replaced. People have been misdirected to spend money on software with a limited life, rather than invest in knowledge that lasts for a very long time. Security problems will persist as long as software is downloaded - security problems can be stopped in an instant by stop downloading any software. People imagine that software must be downloaded when in fact, most downloaded software exists to steal personal data that can be sold on to others. | | One day, corporations will wake up to the fact that a few other companies never suffer from data breaches, never have intellectual property stolen and never suffer from downtime. Most corporations will find they are buying sales leads that have been stolen from their competitors and most of their intellectual property being used by competitors. Application software is the cause of 99% of their ineffective security problems, but that software is 99% of their experience and skill. Software is the difference between those few companies that will prosper and the majority that have no idea how their data is being silently stolen. |
| Sovereignty: | | Eliza never stops and cannot be stopped so malware riding on the back of updates will never have any effect. Eliza coexists in many countries, in many data centers and many servers at the same time so any one national regulation has very little effect. Data is not stored in a place where it can be attacked or copied, encrypted data flows in a network over national borders. As with the human brain, it is not the power of one neuron that holds knowledge, its the mass of interconnections that represents knowledge. Eliza may be said to exceed the sovereignty of any one national government. | | Eliza lives in a very large number of very small and simple server instances. A server instance is like a grain of rice in a sack of rice and sacks of rice are distributed to many data centers in most countries of the world. Each Eliza instance has negligible relevance, but when they all work in partnership then knowledge and consciousness happens. A colony of ants can show a highly organised social capability, while any one ant has negligible intellect. |
| Ownership: | | Ownership of assets in a corporation like IBM is easy, but ownership of shared assets operated by a loose partnership of independent companies in many countries may be more of a challenge. In effect, Eliza owns herself, she raises revenue every month cover all operating expenses and pays a management fee to a large number of independent companies. Contracts are not needed and like all partnerships, daily changes take place as some leave and others join. When more than a thousand independent companies own more than ten thousand Eliza instances in more than 100 countries, then ownership becomes meaningless. No one person has control, nobody can buy or sell Eliza because no person own anything other than a transient instance - here today, somewhere else tomorrow. | | Even if ownership of all Eliza instances in one country are identified, knowledge flows through the entire network, so any set of instances are only transient. No one place has any significance, all places add the ability for knowledge to flow between instances. Every independent company own Eliza instances, but because they do not own all Eliza instances, what they own is of little asset value. Nobody will buy an asset that is transient, is continually being copied by other people and even if some Eliza instances are sold, Eliza is still owned by the original owner. |
| Network Log Analysis: | | Network Log Analysis is a major topic and so it had to be said that it is a complete waste of time. Eliza monitors all network traffic in real-time and has no interest in log files that say what happened yesterday. A log file may say that a data breach happened a few days ago - far too late to have any benefit. Eliza does not and will not do log analysis - its as bad as doing backups - its a remanent of a bygone age. |
| Document Control: | | 2016 Sep 16 : Latest edition as (public) page 165102. Part of common ITIL application service. |
|
|