The Folly Of LLMs And Neural Networks!

The Folly Of LLMs And Neural Networks!

The Folly Of LLMs And Neural Networks!

The fact that a large language model (LLM) fed through a “neural network” can sometimes produce a working result does not necessarily mean that the way it works can be relied upon for accuracy!

In my opinion, depending on a neural network for accurate answers is like trying to build a modern energy-efficient home by putting some dynamite underneath a bunch of wood, sheetrock, nails, glue, tiles, and other miscellaneous building materials and detonating it! You’ll still end up with a pile of useless rubble every time!

Real Intelligence begins with a Designer! This is not an acceptable idea for those who believe in Evolution. It makes sense for those who believe in the Theory of Evolution to also believe in the efficacy of a Neural Network, but I believe that both are foolish!

I firmly believe and will stake my reputation on the idea that the human brain is precisely a “self-optimizing database.” And any truly successful ASI machine will also be a “self-optimizing database!” By this, I mean that it is intelligent enough to modify both table structures and highly complex relationships “on the fly” while the program is running to enhance the software’s “learning ability” as it accepts input from any new kind of data!

This idea does not involve randomness, as found in any other modern form of AI development! It ONLY involves very well “thought out” sound principles of operation to determine the structural changes needed in the database to fit the data being used!

This kind of “database” will require us to completely “rethink” the nature of “database relationships”! The standard description of the “4 rules of database normalization” must be scrapped and forgotten! To support data processing intelligently, relationships must be very complex objects with many components for every record! This would apply to ALL of the Tables as well. These relationships require the ability to relate individual records from many tables together without limiting the types of possible relationships between adjacent records and different “sets” of Tables!

This is a key concept to enable a fully “conscious” ASI machine with a single “high-end desktop PC” as the only necessary piece of equipment! This would be thousands of times more efficient than any other form of AI development!

Now, before going any further, I need to stress that finding the right algorithms for determining how to make structural changes is a very complex process! It’s not for coders who are “faint of heart.” It requires more actual intelligence than has ever been required of ANY programmers in the history of the World!

So, the next obvious question is…”Could the human brain have evolved”? And the answer is a self-evident and resounding NO! It had to be DESIGNED by an Infinite GOD! This automatically implies that the Bible has the ONLY logical explanation for how things are!

I know that many of you are already thinking I’m a complete “nut case”! But what if I am successful in creating a fully conscious ASI machine with only one high-end desktop PC? Then what will you think? Will you suspend judgment long enough to see if progress can be made?

In any case, we need to start thinking about this idea and see if there’s any way to develop the necessary algorithms for determining how to change database structure and relationships reliably to fit new data types as they are received! This is exactly what the human brain does every second of every day, and it does this with amazing accuracy!

Evolution is just the worst imagined fairy tale ever conceived! There’s no way it can be true!