Item — HN Reader

My prediction: At this level of AI-generated writing, keeping the works short will attract (allow?) small audiences ... but only to watch the human talent in the production struggle and improvise over truly awful writing.

If a human was responsible for that writing, they wouldn't have much of a career.

Actors: 1. AI: 0.

2 comments

Agreed. Joeboy's description of this screenplay as "word salad" was perfect: https://news.ycombinator.com/item?id=11876333

I like your scoreboard too -- that pretty much sums it up.

Although I do wonder if actors/directors could use this as a practice tool. Challenge: turn word salad into a meaningful scene. It almost seems like it could be an exercise in a theatre class.

You wouldn't want to do this for the full 10 minute screenplay (it's a little painful even with these talented actors). Maybe generate a 2-3 minute scene or generate the whole screenplay and you get to pick a scene. An optional crutch -- the actors get to do their own 2-3 minute scene before and/or after to give it real context and meaning. That could be interesting.

I have a feeling that something like this kind of exercise is probably already done in training (any actors on HN?). Although I bet this algorithm is better than humans at coming up with difficult challenging incoherent word salad gibberish.

Doing scenes with "word salad" is something improv actors do as a practice tool, as well as scenes where each actor only has a single word or simple phrase they can use as dialog. The latter is often part of a performance for an improv troupe.

Good point. I have seen the single word / phrase prompt for improv. It would be a truly impressive improv group that could make this coherent with an on the spot improv performance (but that wouldn't really be improv). The screenplay seems more difficult than a single word prompt because with the single word there is so much that you get to make up on your own. I feel like the ability to take the word salad and convey emotion and meaning through body language and delivery/emphasis of each word is a whole different skill set for an actor. words from the generate scene only without any additional improv/screenwriting is definitely the most challenging (and that's what they did here for a full 10 minute screenplay!)

I don't mean a single word prompt, I mean you can only use the single word or phrase as dialog and have to perform a scene (also randomly provided) with other actors. Bob gets "roses," Jill gets "fire," and a third person has unrestricted dialog... now do a scene where Jill is a salesperson getting ready for pitch to a big client.

The word salad you're describing would be no more difficult than having a scene where actors have to speak gibberish (faux Klingon or something)... which coincidentally isn't unlike another improv exercise where one person has to convey a message given to them by the director to other actors but must talk gibberish or just use a single vowel sound.

In defence of word salad-generating screenwriting neural networks, arguably the finest spoken lines in any Sci Fi movie - the "tears in the rain" speech in Blade Runner - also owed more to the actor's ad-libbing than the original screenplay

> If a human was responsible for that writing, they wouldn't have much of a career.

I don't think that is fair to the computer. A computer doesn't need a career, and a human has a lot more to draw from than the limited number of screenplays used as a corpus. People have purchased paintings knowingly produced by a chimpanzee [1] over Warhol and Renoir, I don't think an AI should be written off so quickly.

[1]: https://en.wikipedia.org/wiki/Congo_%28chimpanzee%29