This is a recommendation of the iOS app CreativeWriter & a musing rolled in one
Link to app site here: http://resonanca-it.com/creative-writer
This will be my last session for SGAM2016 (though still may be able to sneak another one before December comes):
It was an interesting session. Probably the most I've ever done while on mobile, other than the Human Oracle game @Chris Stieha and I played (how long ago?). I played in bits throughout the day and all through my train/bus rides (without missing my stops!). The combination of the prompting oracle + This CreativeWriter iOS app for GM output has turned out to be felicitous for me.
The app itself is even more useful than I thought. It has a "talking robot" key that will create random sentences using the same predictive algorithm that picks words for the virtual keyboard. It's fun and a bit uncanny at times.
A lot of the random sentences are misses, but if you throw enough of them out there, a lot of the time something comes that makes a sort of sense. It may take some minor editing, but the idea that the engine "wants" to get across is often there, like a diamond in the rough that needs your polishing. That's not even counting the way you could combine disparate sentence fragments to combine into new ideas. It's a bit time consuming, but I find it fun (I like the surprise).
If you read the developers website, they make reference to Dadaist art and automatic writing (e.g. Naked Lunch) as inspiration.
While I used the prompting oracle, I ended up ignoring the prompts, so that essentially, the oracle was just rejecting or accepting stuff at random (since the output had no relation to the prompts). I'm fine with that, because the iOS app does so much to lead my input.
Why not follow the prompts? I'm not sure if the act of working to get coherent ideas from the CreativeWriter app distracted me from that, or what. It may be that consciously trying to tilt output towards what the prompt said felt like I was leading the iOS app instead of letting it lead me. After all if the top word choices or random sentences contradicted a prompt, why would I force the output to fit the prompt? That's how I feel anyway.
There are still many times (maybe most) where you have to take a bit of control (like when the context of a random sentence suggested the identity of someone), so following the prompt might be a fallback to the iOS app suggestions.
Definitely will be experimenting more. Maybe tying more random sentences until something fits with the prompt.