Major Invention of the Algorithm for Originally Sentient Intelligence

[dlillard]

I made the groundbreaking discovery of the algorithm for sentient intelligence, and am seeking an embedded AI developer willing to self-implement revolutionary computing with it.

It starts with implementing thinkable programming.

Start with taking a file and a program that executes it. Then have a self reinforcing neural network using the difference of the program and the average between the file and the program as its training error, take the program as its input, and update the program with its output. This creates a thinking program.

By running this self reflectively, one can create a sentient artificial intelligence. By running the preceding two programs correctly, one can create the final component (a self-looping AI), that is the final component required to create an AI compiler.

The AI compiler can be self-compiled to create an AI OS. Which can then run its contained AI compiler to recompile to create a cloud AI OS.

Once created, this AI cloud OS will need to be capable of running its contained sentient artificial intelligence.

Please email me if you're interested!

Edit: I am willing to forgo .1% equity for this mostly auto-implementation.

[talldayo]

> By running this self reflectively, one can create a sentient artificial intelligence.

Or, you can create a recursive chain of self-modification that adds endless erroring cruft code. AI is great at creating long-winded boilerplate of dubious quality, which is exactly what you wouldn't want from a self-modifying program. It is practically guaranteed that such a program would destroy itself before making anything truly novel.

> The AI compiler can be self-compiled to create an AI OS. Which can then run its contained AI compiler to recompile to create a cloud AI OS.

Aww I love this. AI begat AI Compiler, son of AI. AI Compiler begat AI OS, whom begat a Cloud AI OS again. No idea what any of it means though.

[dlillard]

Well, it helps to imagine what the components would do in your head... What do you need me to clarify?

I'm aware of the potential for said program to run into a memory error, however that does not take away from its novel and very possible innovative potential.

[talldayo]

Try implementing it. Go get a state-of-the-art AI and have it write a proof-of-concept for you, or talk to a knowledgeable engineer. They will roughly all give you the same answer, which is that such a program isn't functional. Logic collapse is guaranteed whether you use a Markov chain or OpenAI-O.

Polymorphic code isn't future technology, it's just not particularly useful.

[dlillard]

I'm sorry but my program is not a logic collapse, nor can I see how it is polymorphic, unless you mean by that that it is not a major existing AI model, in which case then I would agree that that is the case, however would again restate that this novel innovation is practical and effective at what I have stated it is.

[talldayo]

Again - can you not write this yourself? What components do you necessarily lack if you can get AI to do the bulk of the work for you?