Congrats on becoming a Dad too 
Thanks man!
I know it was loosely discussed here a while back but is there a way to get users to do most of the work to reduce the workload a bit? i.e I pick a vid and give you the timestamps of the 6 clips and you punch this info into your code/program? I'm sure it's not that simple but was wondering if this might make things quicker in the sense that you could quickly do a whole week or month's worth of games in a shorter time frame than if you're doing it all yourself?
Yeah I kinda touched on this in my post, it's still quite a bit of effort. For example the weekend around Jan 16th I had to queue at least a few weeks worth of videos because I was going down for surgery on the 18th which ended up knocking me on my ass for 2 weeks.
Most of the videos I queued were already "pre-processed" in the sense that they were suggestions from folks and ready to chop up.
That indeed helped out a lot but it still took me a total of almost 5 hours between a friday night and saturday morning to get ~15 or so up. It worked out this time because my wife and son were out of town but I would rather not have to do that again on a weekend.
I know it sounds a bit selfish as I have built this thing that people love that I'm now discontinuing. Trust me it was an extremely difficult decision to announce the pause.
I was especially let down with myself when the automation didn't pan out as that was my list glimmer of hope that this could go on forever. Well I mean the automation in a way did work, but the quality totally sucked and I think my curation of the clips played a bigger part than I realized in what made the game fun.
And unfortunately it's a bit out of my skill set to come up with some machine learning model that could spit out "good clips", which I'm sure is somehow possible in today's age but I don't know where I would start. The general problem of automating "subjective curation" is basically the cornerstone of the difficulty I'm facing.
Poll
Topic: Show Slap: Skateboardle (Read 44636 times)