Agreed.  Although theoretically a machine could learn the beekeeping management generalities, it is unlilley that anything currenly  could learn the specifics as well as a human. For one thing, training it would be a task and maybe a custom job for each region and operation.

The specifics of each location and situation are different and there are occasional one-off events that are not easily programmable, plus economic and social factors that are fuzzy.. 

As anyone who has made models with Excel knows, any such model quickly becomes a black box and generates fiction as the probability of internal errors grows exponentialy and as the parameters change even before the model is printed.  Excel is great for writing fiction, but trusting such models as primary guidance over human instinct has caused many business wrecks.

While AI might be interesting to run as a second opinion, at this stage, only the naiive would trust it over proven human experience for making important beekeeping decisions.

Jerry had a model that predicted bees coming and going and was doing just fine until the hive swarmed. Having never seen that, the model expected the bees to return. 

             ***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software.  For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html