Some interesting possibilities listed there. I found this one particularly questionable, though:
> The number of jobs created by new technologies is likely to surpass those destroyed during the next 20 years, judging from past episodes.
That seems to assume linear rather than exponential growth, or at least assumes that the cumulative distribution function of job difficulty follows a smooth and predictable curve.
Moreover, it misses the point that the jobs created will not necessarily be appropriate for the people (or in the countries) where the jobs are destroyed. Even if they were, there's a lot of inertia in the job-seeking system; so, for example, if the average person has to change jobs every 6 months, that's very different from changing every 4 years.
