The Observant Watchmaker
Intelligence has not been well defined lets start with some definitions.
A strategy specifies goals, their desirability and, at what likelihoods to take what actions on what (set of) conditions.
Strategizing is Deviseing a set of strategies.
Devising strategies consists of sub processes of creating and assessing conditions for actions , weight of goals, estimates of costs of actions, estimates of effectiveness of actions, comparing strategies, combining known strategies covering contingencies and evaluating strategies.
Strategies can be devise by processes such as evolution.
A gene encoded on a strand or DNA is a strategy that executes on some trigger conditions. Evolution often creates a programmed response to some action. Evolution uses mating and, survival to evaluate strategies. Evolution uses vertical gene transfer and recombines DNA to combine strategies. We are finding new strategies evolution uses for coming up with and modifying strategies.
Strategies are devised and combined by computer programmers to make computer programs and perform intelligent new task. Different methods of code are combined to come up with new strategy. Predetermined set of instruction under determined conditions (sets of states input) and terminate on goals or error. Performance and adaptability, efficiency and other properties determine survival of a program.
Genes are Programs are strategies. Adaptability determines survival. Humans and evolution create methods for coming up with and improving strategies.
Devising strategies is equivalent to intelligence.
Trying to derive intelligence with models just based on biological evolution is primitive and brute force. I am sure that this breeding program will lead to a wonderful screen saver. Methodically improving strategy scientifically and throw dissection of existing strategy allows iterative improvements to strategy.
Past blunders in estimating intelligence potential of self modifying code, and estimates on effects of creating a optimal strategies for some given problem like chess show a deep misunderstanding of intelligence. The effect of such endeavors mostly lead to a few new strategies for solving some types of problems.
As a measure for ETA of the singularity the effect of Moore's law is a useless metric compared to the measure of smarter formalized strategies for smarter creation and selection of strategies. Processes for creation and iterative improvement of sub-processes devising strategies is what is actually needed to make strong AI work.