Fliq.ai – Movie Intelligence Open Beta
fliq.ai
fliq.ai
I think AI can be awesome insofar as Hollywood 2.0 is concerned (to use a YC term). I'm working on figuring out what scenes two actors are in (using facial recognition), splitting a movie by scenes based on frequency of cuts, and an app that lets you see spoilers as you watch a movie on your TV!
Either way, awesome job! Fliq seems like a very interesting value prop, especially for industry professionals. Out of curiosity, how do you seed your data? I've had some trouble figuring out what the best way would be to do that. I'm using a hacky way of grabbing it from themoviedb.org, but it's not a particularly tenable solution.
Also, I notice you built it using Angular. Any particular reasons you guys didn't opt for React? It seems like a more straightforward way to build a dashboard-style app like Fliq has.
(Apologies for the barrage of questions, I'm very excited about this space.)
We've started working on some advanced language parsing out of subtitle dialogue, somewhat similar to what you're doing we're trying to make a link back to actors interacting (in our case via character names). I'd be curious to hear how you've developed your scene ontology for categorization/splitting purposes.
Agreed about potential for AI to inform better content decisions, or at the very least decent analytics would push things forward quite a bit. As far as your questions:
1. Data collection and mastering is like 80% of the work as you probably know, themoviedb.org is great and we use it too. Myapifilms.com, allmovie.com are solid sources. We also heavily leverage Wikipedia, google and waybackmachine.org.
2. Angular - personal comfort level + compatibility with the rest of our stack (certain charting libraries and themes we use). We're looking closely at React now and there's a good chance that the next generation of FLIQ moves in this direction.
Is signing up even truly integral to the app in some way, be it functionally or for monetization purposes? Either way, a registration wall during beta for a highly visual product doesn't seem like a great idea.
I know the counter-argument to this is probably that this limits the number of beta users and makes it less overwhelming to manage early on, but I side with the school of thought that favors exposure above all else, especially during early stages.
We do try to communicate the value prop in the blogs, the video and (yes) the screenshots. At least enough to get people to clear the relatively small hurdle of a SSO. If it's a matter of doing that more effectively I'd love to hear your thoughts there =)
In a better world perhaps this would be true...
That said, some of the feedback we've gotten so far from more casual users also likes the future movies and buzz score, it's an easy way to see what's coming up this year and how excited people are about it.
With a solid decision making tool, movie execs could finally have a financial reason to accept original scripts that are not yet another sequel and reboot.
Sequels are like a trap perfectly designed for movie execs, there is a sort of guaranteed baseline of success - but you're also guaranteed diminishing returns. Currently, they are squeezing a struggling market and sacrificing long-term viability for short-term revenue boosts that mask the structural problems. It's sort of still churning out $10B/year as a market, but that's hiding a lot of weakness and our suspicion is that things will change drastically in the next 5 years.
Hopefully, if we're all thinking about this stuff a little smarter, things change for the better.
Don't.
Scrolljack.
I feel terrible especially since you took the time to put together the feedback, and truly we would love to hear it. If you don't mind, please feel free to reach out to us via twitter DM/email with any thoughts you'd still be willing to share.
ComScore/Rentrak is the first place that comes to mind that has almost complete box office info, but there are others.