In Ontario (at least pre-Ford) you can get full-time benefits as an hourly worker, as long as you work 40h a week or more (or maybe 35?). It's been awhile though so I'm not super confident on the details here.
Yeah, this is why, like so many other things involving regulation, there should be a gradual proportional phase-out, not a hard line in the sand. So rather than paying two people to do 20 hours each and pay for no benefits, if you did that, you'd be paying for half benefits for each, at which point you may as well just have one person doing 40 hours a week who's fully employed and getting full benefits like you should've been doing all along.
The same is true (at least in my experience) in America. Over 40 hrs/week means the employer has to pay out extra benefits. This is also why employers do not want you to work over 39 hours if you're not salaried.
In the US, over 40 hours simply means 1.5x your regular payrate after the 40th hour, if you're "non exempt", i.e. hourly. Some employers might offer something more.
(I've done work in QA and have played at building Jeopardy style QA models)
Watson (Jeopardy Watson, not the IBM branding exercise Watson is now) has much weaker text understanding models, but has much much better optimisations for the incremental style of data release that you see in Jeopardy (ie, you get more and more data the longer you listen). IBM did a lot of work optimising when to answer as well as trying to get the correct answer.
The closest analogy that is regularly studied in modern QA research is "Quizbowl"-style datasets, but these tend to be much smaller than the SQUAD datasets that most modern neural network QA systems are built against.
