Every indexed document

by recognizing patterns among the nodes of the tree

driving transforms a program into a huge “decision tree”

corresponding to a rapidly-executable but unacceptably humongous version

it creates a new node in the tree. This part

and since this time the concept has been avidly

and goal-directed program creation requires another

susceptible to analysis and improvement by the

so the system can see what is going on inside them. Pragmatically

and we suspect that such code will be comprehensible to AGI systems

or even the mathematical algorithms and structures underlying the code

especially complex AI code. The hard problem isn’t understanding the formal syntax of

and opt to either improve its source code or create an entirely new

it seems extremely likely that an AGI will become unsatisfied with the way

it will be better at programming than humans

then the odds are high that given its ability to carry out

thornier than e.g. the creation of narrow-AI chess players that play better chess

even with respect to the environments in which humans evolved.

compared to current real-world AI systems

one AGI approach that we believe to ultimately be capable

objective way of assessing the percentage