What Next?

Where do we intend to go from here? Both BRUTUS.1 and MONA-LISA will live on, both to be gradually improved. In the case of BRUTUS.1, we will work sedulously as good engineers, clinging to a workday faith that some more robust system can capture all of $\cal I$. (When the workday ends, we will permit our suspicion that $\cal I$ is productive to have free rein.) In connection with MONA-LISA, we plan to

1.

Attempt to substitute artificial agents (treated as ``percept to action" functions [Russell & Norvig, 1995]) for the role of the human in the image elicitation process.

2.

Transfer the images elicitation approach to the domain of stories (so BRUTUS.1 can get some help!).

3.

Explore more carefully the complexity implications of our image elicitation work.³²