4 ms·
> their ability to handle a lot of variable data and inputs and weigh them dynamically I dunno. Computers are pretty good at this one too, on different kinds o
by marcosdumay 7y ago
> their ability to handle a lot of variable data and inputs and weigh them dynamically
I dunno. Computers are pretty good at this one too, on different kinds of variance and weights.
The largest problem I see is that everyone is discussing "either or" on this thread, while the clear winner is both.
- duxup 7y agoI think ultimately if you're working with humans the question is if you want the schedule to solve a sort of math problem for you ... or if you want folks to be happy. They are very much not the same thing and the dynamic nature of keeping people happy and filling a schedule is extremely difficult to account for with software.
- bumby 7y agoI was reading some old literature a few months ago on this related to police staffing where union issues had to be taken into account. It seems adding social penalties is a way to do this algorithmically
- tonyarkles 7y agoYou’re giving me flashbacks to some software I worked on a while ago that had to use rules from three different collective agreements from nursing unions. Turns out writing business logic for rules that aren’t internally consistent is a challenge, and unlike most LOB projects I’ve worked on, the procedure can not be amended until the next round of bargaining.
- dwaltrip 7y agoThey are far worse at it than humans. This is the difference between narrow AI and general AI. Humans can do millions of different complex tasks, and can adapt when the situation changes. Computers are terrible at updating their models automatically based on new information, as they don't have understanding of the bigger context. Worse still, software today has no way of analyzing and adjusting the assumptions it is operating off of.
- bumby 7y agoPeople are (currently) better than computers at updating their models but I don’t think we’re particularly great at it. Look at how many business run on the “this is how we’ve always done it” model
- mntmoss 7y agoComputers solve the legible problems. But scheduling is one where the problems lack in hard specification, and plans involve several "layers" of prioritization, so what's called for is approximate results that avoid failure modes, which is something a skilled human does pretty well. I do think that software speeds up some parts of the process, but like so many automated processes, it's computer assistance, not computer intelligence.