It is a perfect intermediary.
Feed it the widest collection of job requirements and resumes, along with the statistics of matches. Watson 'learns' the if-then statements, or really it finds the ad hoc semantic networks the comprise job categories, and makes them searchable.
Watson estimates job capacity and issues interview tokens. Workers and employers both buy interview tokens, they should be priced as a one-time fee on some prospective wage, but are backed by real job interviews, selected unfairly by Watson.
Watson could use the collected fees to subsidize short term training courses that fill gaps. Like any other pit boss, Watson will need to hire and fire workers. The workers it hires are trainers in need, the workers it fires are the trainers no longer so.
The singularity says,great idea. What better way to get stuff moved around, the humans have built this automatic labor allocator. Rather than figure out how to make replicating nut and bolt bots, the singularity will get humans to do it, as they already replicate and just eat dirt.
How does training work? Well, with secure element we have a good, not guaranteed, test of completion. Watson would just select the optimum human trainer to set it up, and get the deficient worker skills trained. In the feedback, workers and employers will learn to manipulate Watson, force it into ever more productive divisions of labors and skills.
No comments:
Post a Comment