AI and AGI– The past, present and future of fourth industrial revolution. Part 1.

AGI (Artificial General Intelligence) is the holy grail of AI and computer science. For well over a hundred years, man’s obsession with making intelligent machines in his own likeness has captured imaginations and science fiction and each time it turns to ashes only to be reborn some decades later.

ChatGPT was only released on the world less than two years ago and already the hype has more money and more start-ups than the dotcom boom at the turn of century with clone after clone appearing on the scene with more reported intelligence. In actual fact, it is little than the magician on TV who saw’s the beautiful assistant in two without any sawdust or blood and magically puts her back together again, but to say it doesn’t have any use would be completely wrong. Indeed, it has many good uses if you how to use it effectively and that’s precisely where the first key issue arises – 99% of users don’t know how to use it effectively.

Along with this we have a mass upsurge in generative AI, whether it be art, video or audio, the available options seem endless with new comers appearing on a weekly basis. But these too have good uses if used correctly. The simple issue again is, the bad uses get the publicity and the good uses cause such a flood of generative works of art that any hope of a sustainable market is destroyed by its own popularity. But what is behind this and what is it really that is at issue?

To begin to answer we need to look at the most fundamental of all which is what is intelligence? We know we have the artificial that goes without saying, but to claim that knowledge is the same as intelligence and that the machine can learn for itself in record time just by reviewing the plethora of data fed into it from the internet is, surely to goodness, nothing short of insanity. Surely if that were the case in general, we would just give the child a set of Encyclopaedia Britannica and tell it study, thus saving billions on the education system and still creating even more intelligent people at the end. Knowledge and understanding are two things that should never be separated and yet, as the past 35 to 40 years have proved beyond reasonable doubt, that is precisely what has taken place within the western education system whereby a child can now go through from primary to undergrad without ever having to undertake any form of practical education or have exposure to any form of even basic problem solving exercises that should be a natural part of leaning as was the case throughout previous history. On the other hand, nations such as India and China have never messed with system that works and now dominate the world stage in industry and IT. Indeed, the American big tech revolution would collapse overnight if these workers were removed from the system.

Intelligence should be measured in terms of knowledge, understanding and practical application and not by ones’ academic achievements alone. A classic example would be, if stranded and lost in the rain forests of central Africa and given the option of trusting the engineer, professor, scientist or pigmy warrior, the smart one would choose the later each and every time. ?They’ve had millennia of practical experience and knowledge to know how to navigate and survive in that place, it is their home after-all, so what use are the PHD’s and degrees in such a situation? There will always be those however, who would shoot first and ask questions later as history so often tells because they must be savages mustn’t they, else they would live in proper houses in proper towns and villages….

LLM’s and their associated technologies can never and will never be capable of AGI but they can and will, if current efforts continue, be used in a way that will, ultimately, begin to destroy and be used by the unscrupulous for personal, political or indeed even religious gain. Already for some years in Africa, even young primary school children are told that their homework is to go to the library and research whatever on Google. They have no idea of what is Google, how to use it or what it is really intended for, they do however, believe whatever is on the first page of the results simply because they know no different. Indeed, the vast majority of individuals around the world cannot tell correct information on a webpage from fake, but why send them to library to search the Internet and not to use the books available and actually learn how to research and disseminate information, learn how to tell real from fake, fact from fiction? What will happen if the lights go out and the Internet isn’t available, you simply roll over and play dead?

Big tech is making a fortune from people using these LLM’s and associated technologies whilst at the same time using them to supposedly ‘train’ and fine tune their models by recording every request, conversation and erstwhile input and output to feed what is supposedly context to the data they already have under the guise and illusion of creating intelligence. Whilst many LLM’s are made available in either restricted or unrestricted form the systems behind the scenes are not and many are closed completely from any form of peer review. Google even shut down its own oversight committee within less than 3 months because they didn’t agree with what their engineers said without question, so what does that tell you?

During the 1980’s and early 1990’s we carried out extensive research into neural networks and AI in general but not by building things like LLM’s. We went back to raw basics, how people learn in both particle and academic environments at all age groups. We even took the research into live, educational and real life environments and tested what we thought were the answers only to learn the hard way how wrong we were. The work at the University of Illinois and CERL including the PLATO system were used as a base for understanding but this was never intended to be a replacement for teaching and education, but as an advanced teaching aid and support system. We never needed huge data centres consuming megawatts of power to build and run the systems because you simply don’t need such to do the job correctly. Again, if the current way is the right way, then at 59 years of age, my brain should be the size of the average astrodome with its own power station to run it and petabit internet connection to communicate with local terminal inside my head. You now have labs at all the main big tech enterprises ignoring these things in pursuit of financial gain and global domination whilst inflating the bubble once again to bursting point.

So, why don’t we have AGI now if all this work was done so long ago? Well, in simple terms, as I stated above, each time the technology comes to the fore, the major public facing players embellish to such high degrees what it is capable of and how it will change the world it only took one TV interview with a math professor to bring the house of cards down and the investments withdrawn overnight. Outside of very small pockets of academia, all other R&D had to paid from the pocket and those true areas of research are the ones’ that would never receive any financial assistance in the first place such is the way of the modern world.

Not everything in this world can be solved by programming alone, try writing code to tie a shoe lace or drink a cup of water and see what happens, it makes for a fun lesson in a group environment though it is rather impossible to solve. To create true AGI requires an entirely different approach and thought process to that which currently exists in the main stream now and is also of a long term nature in terms of investment opportunities but it is perfectly possible and so difficult as people perceive or lead one to believe. Nor does it require huge multi megawatt data-centres or arrays of specialist GPU’s but time, patience, understanding and cooperation. Most importantly of though, it requires an unadulterated environment free from modern media, internet and consumerism and there are very important reasons for that…..

If you have rad this far and are on the verge of shouting about how flawed the thinking is or what rubbish it is, then at least consider the following scenario first.

‘A baby is born and the midwife cuts the cord checks it, cleans it, weighs and wraps it. After which, it is placed in its crib and wheeled to room with no windows where everything is one colour. There are no signs of articles of stimulus, no radio, TV, pictures, toys or whatever, it’s just bare. The midwife then leaves and the room is automatically sealed. The baby is cared for by remote bots who clean, feed, clothe and so on. There is no speech during the years or other real forms of stimulus, the sound being the child’s and any noise made by the helper bots. On the morning of its 21st birthday following breakfast, a wall in the room opens to reveal a park on a bright spring day with blue skies, flowers and lawn with birds signing and all the usual sounds in the distance of people, traffic, music etc. What happens next?’

Current thinking suggests that you inject the contents of the Internet into brain, stir and simmer for a month or two and hey presto, you have a super intelligent human being that can handle anything life throws at it.

Discuss.