July 28, 2025
More great voice AI example code from @uberboffin. Play the game Guess Who interactively with an AI partner.
We had a thread here last week about "selective refusal." Selective refusal means prompting or training an LLM to sometimes not respond to input. Or, more accurately, to generate a <no response> response. Your application then handles the <no response> response however makes sense for your conversation flow.
Sam's code defines a function call that facilitates selective refusal.
Sam also gives the LLM structured data about the game characters, and a very clear prompt defining the game rules. Based on my prompt engineering experience, I'm guessing it took a fair amount of iteration to get this prompt right.
If you're interested in multi-turn LLM instruction following, it's definitely worth looking at Sam's code.
These days, some of the most advanced voice LLM use cases I'm seeing are games.
The code[1]
The AI character "Humphrey" is powered by:
🎙️ @Speechmatics ASR + diarization
🔗 @pipecat_ai for WebRTC + function calls
🧠 @OpenAI for smarts
🗣️ @elevenlabsio for that smooth, smooth British voice
@Speechmatics @pipecat_ai @OpenAI @elevenlabsio Here's @uberboffin's post: https://t.co/vuEjVAIAYG
Sam Sykes@uberboffin🎲 Guess Who? — but make it real life 🤖🧍♂️🧍♀️
In our latest hack, I took the classic board game Guess Who? and threw it into the real world — powered by:
🎙️ Speechmatics ASR + diarization
🔗 Pipecat for WebRTC + function calls
🧠 ChatGPT for smarts
🗣️ Humphrey, our eloquent
