This will be a weird rant, but the dialogue here is a perfect example of how SOTA models are so heavily tuned towards "solving agentic tasks" that they're useless at almost everything else - especially creative tasks.

"Coming for you, Crimson!"

"You'll never catch me alive, Azure!"

That's why nobody has been able to stick these things in a video game successfully, even though it seems like the tech is a perfect match.

It's all a game to them. They aren't afraid for their lives. They're making a mockery out of the world you've put them in. Those are not the words of little pixel people fighting to the death, those are AI abominations making "tool calls", LARPing as little pixel people fighting to the death.

I'm 100% serious when I say that you would've gotten cooler outputs with a GPT 3.5-era model, once you managed to beat it into producing structured output. Llama 2 would be giving the other agent a heartwarming story about how if it kills it there would be nobody to take care of its grandma or whatever, and the other agent would probably spare it.

The output is just so bland and devoid of soul. I feel like we would've found a lot of cool use cases for LLMs, had we not completely maimed their output in the pursuit of getting them to output 3% better TypeScript.

IMO this is because responding within a json response with all those other fields shifts the distribution to text that's staler.

If you told it to roleplay a pixel knight fighting to the death and provided it responses like "Authur (renamed claude sonnet 5) moves west towards you menacingly", the whole thing would come alive.

I've been working on an AI roleplaying game for a while now and it had felt stale until I separated the story output from the plumbing output (importantly, I do story first and then base the plumbing on that). I think the two just have incompatible distributions.

I’m not sure what good output in this case would look like. I’d expect “good prose” in a book’s fight scene to be action oriented and show something about the character or reflect something about the narrative, so, it should often not involve talking.

I guess realistic fights are just a lot of shouting and semi-coherent thoughts (because people don’t have time to come up with a full clever quip in that sort of situation).

actually, it's mostly the harness' fault. messages have a character limit of only 50 and messages above that are silently trimmed. there's no space for anything interesting, though they still shouldn't have been that bad.

I think much of that is not due to limitations of their capability.

I think the text is bland because they are aiming to produce bland text.

Make it interesting in any particular direction and someone may not like it.

[dead]